7-Mar-1997 13:27:55-GMT,1913;000000000001 Received: from axp14.ams.org (MATH.AMS.ORG [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with ESMTP id GAA27222 for ; Fri, 7 Mar 1997 06:27:54 -0700 (MST) Received: from pillar.elsevier.co.uk (GATE1.AMS.ORG) by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with ESMTP id <01IG7OTM0QOG0008H6@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Fri, 7 Mar 1997 07:40:56 EST Received: from snowdon.elsevier.co.uk (snowdon.elsevier.co.uk [193.131.197.164]) by pillar.elsevier.co.uk (8.8.3/8.8.3) with ESMTP id MAA07573 for ; Fri, 07 Mar 1997 12:38:47 +0000 (GMT) Received: from cadair.elsevier.co.uk by snowdon.elsevier.co.uk with SMTP (PP) ; Fri, 07 Mar 1997 12:40:55 +0000 Received: from lochnagarn.elsevier.co.uk (lochnagarn.elsevier.co.uk [193.131.216.1]) by cadair.elsevier.co.uk (8.8.3/8.8.3) with ESMTP id MAA26835 for ; Fri, 07 Mar 1997 12:40:47 +0000 (GMT) Received: (from srahtz@localhost) by lochnagarn.elsevier.co.uk (8.8.3/8.8.3) id MAA27242; Fri, 07 Mar 1997 12:40:45 +0000 (GMT) Date: Fri, 07 Mar 1997 12:40:45 +0000 (GMT) From: Sebastian Rahtz Subject: pdftex To: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199703071240.MAA27242@lochnagarn.elsevier.co.uk> TUG has set up a mailing list to discuss Han The Thanh's TeX variant which produces PDF instead of DVI output. If you want to join, send subscribe pdftex to majordomo@mail.tug.org I am the list moderator (cos i am interested, not because i am in any way involved in the project) pdftex is on CTAN in systems/tex2pdf, for those who want to dive into it in its current state. But Han The Thanh is not at all committed to the current state of the changes or the syntax of the new primitives, so anyone on this list looking at porting it should be careful :-} Sebastian 16-Mar-1997 12:58:59-GMT,6450;000000000001 Received: from math.ams.org (MATH.AMS.ORG [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with SMTP id FAA11855 for ; Sun, 16 Mar 1997 05:58:56 -0700 (MST) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 16 Mar 1997 12:58:54 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IGK97XUS9S001IM2@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Sun, 16 Mar 1997 07:34:42 EST Received: from kralle.zdv.Uni-Mainz.DE ([134.93.8.158]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Sun, 16 Mar 1997 12:34:30 +0000 (UT) Received: from frank.zdv.uni-mainz.de (Ufrank@localhost) by kralle.zdv.Uni-Mainz.DE (8.8.5/8.8.5) with UUCP id NAA31886; Sun, 16 Mar 1997 13:31:28 +0100 (MET) Received: (from latex3@localhost) by frank.zdv.uni-mainz.de (8.6.9/8.6.9) id NAA06900; Sun, 16 Mar 1997 13:23:22 +0100 Date: Sun, 16 Mar 1997 13:23:22 +0100 From: Frank Mittelbach Subject: Re: (La)TeX virus In-reply-to: <9703150154.AA10806@hemlock.pa.dec.com> To: lamport@pa.dec.com Cc: "Dr. Yvo Desmedt" , Keith.McMillan@ameritech.com, tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199703161223.NAA06900@frank.zdv.uni-mainz.de> References: <199703141743.LAA03559@blatz.cs.uwm.edu> <9703150154.AA10806@hemlock.pa.dec.com> X-Authentication-warning: kralle.zdv.Uni-Mainz.DE: Ufrank set sender to frank.zdv.uni-mainz.de!latex3 using -f > Dear Mr McMilland and Professor Desmedt, > > But co-authors e-mail (La)TeX documents! > > I'm not too worried about co-authors inserting viruses in the versions > they send to one another. (Yes, I know that this is a way to spread > an an infection.) not having seen the original paper message i can only guess as to what you think is the security issue involving TeX (not LaTeX by the way---as LaTeX can't implement any higher level security with respect to what TeX offers or does not offer) so my assumption is that you are worried about the possibility of TeX opening write channels to any file on the system that allows being written to, e.g., to .profile on a unix system (or read channels to files like .netrc and then copying this information to some other file accessible to the intruder) this is certainly a possibility and TeX does not offer right now a way to identify such actions, however, although I agree it would be better if there is a way to determine which files are written to or prevent writing to files in other directories (using a settable switch) i don't consider this a very serious security problem. TeX respects permission bits and so files that are read-only will not be modified --- as it is advisable to keep .profile and similar files in read-only mode anyway and that takes care of this problem already. In comparison to standard software distribution these days (with a configure/make/install type of setup) where it is often nearly impossible to figure out the exact actions carried out easily, the danger when using TeX seems to me remote; and in fact i never heard that anybody was seriously getting into problems > >However, any changes would have to be made in TeX itself, and such a > >change would be up to Don Knuth. > > He replied and said he forwarded the letter to "people who should be > able to take appropriate actions", I presume you know who that is. > > No, I don't. Perhaps Frank does. those are the people who implement TeX for various platforms. In fact those people have discussed the above issues some time back and if i remember correctly the final recommendation was to just inform users that TeX can write such files > Alternately, you could have a security parameter that allows to write > to any, same directory, or same file. The last is more secure. i don't know what is meant by "same file". Most TeX macro package rely on the possibility to produce scratch files to write to and read from so that it is impossible to restrict a TeX run just to produce a single file. > That parameter would have to be set when TeX was made, or at least > when the Format file was made. Such parameters tend to gravitate > towards the value that causes the least inconvenience--namely, the one > that is least secure. People who are paranoid to create their own > more secure versions are probably not going to risk running random TeX > input anyway. So, I see no reason to add a parameter. I don't think that is true; it would be possible, for example, to implement TeX to have a runtime command-line switch that would for example display writing to OS files via \write channels. > NOTE: we plan to send this to CERT and to a conference. As you see, > we are not in a rush. How long would it take before a corrective > action can be taken? > > Not knowing who is to do the correcting, I have no idea. Perhaps other than the above i don't see that there will or should be any corrective action to be taken. Anyway, all this is up to the TeX implementors --- it has nothing to do with macro package implementation on top of TeX, ie talking to Leslie or me is addressing the wrong people. As Don Knuth is not changing anything in the base source of TeX any such addition would need to come from implementors on the various platforms (some of which are commerical some of which are freeware or shareware). that in turn means that it is very unlikely to get a single corrective action implemented. > Frank does. Anyway, it's not clear whether it is better to publicize > this weakness so people will be careful about accepting TeX sources > from strangers, or to keep it quiet to avoid inspiring knaves. well, i think i would need to see a source that is doing any harm and is not looking suspicious in the first place. Clearly if one accepts arbitrary TeX code from strangers then too bad---that is like getting shell scripts or executables and just executing them so i think it might be a good idea to tell people that TeX code can do harm (as most stuff can --- such as ps files ....) so that it might be wise to look over a file first and perhaps keep important files read-only or not execute unknown TeX stuff from stranger on an important account. regards frank cc tex-implementors 16-Mar-1997 15:35:08-GMT,5072;000000000001 Received: from math.ams.org (MATH.AMS.ORG [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with SMTP id IAA14745 for ; Sun, 16 Mar 1997 08:35:07 -0700 (MST) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 16 Mar 1997 15:35:05 UT Received: from AXP14.AMS.ORG by AXP14.AMS.ORG (PMDF V5.1-8 #16534) id <01IGKCG9HDQO001L43@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Sun, 16 Mar 1997 10:23:38 EST Date: Sun, 16 Mar 1997 10:23:29 -0500 (EST) From: bbeeton Subject: potential virus spreadable by (la)tex To: TeX-implementors@MATH.AMS.ORG Cc: keith.mcmillan@ameritech.com, desmedt@cs.uwm.edu, bnb@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <858525809.195705.BNB@MATH.AMS.ORG> MIME-version: 1.0 Content-type: TEXT/PLAIN; CHARSET=US-ASCII Mail-system-version: attached are verbatim copies of a letter to don knuth regarding a potential virus using (la)tex, and don's response. apologies for the delay in distributing this. frank mittelbach has already addressed this topic in a message to tex-implementors, on the basis presumably of a message forwarded from leslie lamport. on account of the nature of the letters, please refrain from placing this message into any public archive. thanks. -- bb -------------------- University of Wisconsin, Milwaukee College of Engineering and Applied Science Department of Electrical Engineering and Computer Science PO Box 784 Milwaukee, WI 53201 February 24, 1997 Dr. Donald Knuth Department of Computer Science Stanford University Margaret Jacks Hall, Bldg. 460 Stanford, CA 94305 Dear Dr. Knuth, As you may be aware, a virus has recently been discovered that inhabits macro files used by the Microsoft Word program. Unfortunately, we need to inform you that in the course of completing an MS degree at the University of Wisconsin-Milwaukee in May 1994, I, Keith McMillan, created just such a virus using LaTeX, relying mostly on the facilities of TeX, and GNU Emacs. This virus, while not damaging (it carries no payload) is virulent. This virus only existed in a controlled environment, and no copies exist in the wild. In brief, the virus inhabits LaTeX document files, inserting a small routine into the .emacs file to locate other files to infect. It reads the target file into the .aux file generated by LaTeX, adds itself (in the form of macros) into the file, and then copies the newly infected file back over the original file. We believe that while this virus never existed outside a controlled environment, just such a virus is within the reach of any reasonably fluent TeX user, and precautions should be taken to prevent the misuse of TeX in such a manner. We further believe that restricting TeX to the types of files it can open for writing, for instance restricting it to only opening files of the form FILE.*, where FILE is the top level (i.e. user) input being read, could make an attack such as the one Keith developed ineffective. (This was suggested by Yair Frankel.) Further, it may be enough in the short term to inform the TeX user of what files are being opened and closed on the standard output channel, instead of only logging it to the log file (which is typically not read, unless something goes wrong). We must admit we have a degree of trepidation regarding your reaction to the news. We can't help but feel as though we are misusing a facility you provided in good faith, and betraying a legendary figure in the Computer Science field. On the other hand, it is probably better that these problems are uncovered and documented in the research community, rather than in the wild. Keith would be happy to provide you with a copy of the thesis. Yours Sincerely, Keith McMillan Yvo Desmedt Keith.McMillan@Ameritech.COM desmedt@cs.uwm.edu 3413 N. Cramer Street Professor Milwaukee, WI 53211 cc: Leslie Lamport -------------------- Stanford University Stanford, California 94305-9045 Donald E. Knuth Professor Emeritus of The Art of Computer Programming Computer Science Department -- Gates 4B Telephone [415] 723-4367 March 4, 1997 Professor Yvo Desmedt Mr. Keith McMillan Department of Electrical Engineering and Computer Science University of Wisconsin, Milwaukee P O Box 784 Milwaukee, WI 53201 Dear Professer Desmedt and Mr. McMillan, Thanks for writing about the potential virus. We had alerted implementors to this possibility many years ago, and in particular I thought it was impossible to write ".emacs" from within TeX. But evidently people have not followed the guidelines we suggested. I am forwarding copies of your letter to the people who should be able to take appropriate precautions. Sincerely, Donald E. Knuth Professor DEK/pw bcc: Barbara Beeton Barbara, Please show a copy of their letter to everybody that you think should see it. 17-Mar-1997 1:31:15-GMT,2076;000000000001 Received: from math.ams.org (MATH.AMS.ORG [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with SMTP id SAA26126 for ; Sun, 16 Mar 1997 18:31:10 -0700 (MST) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 17 Mar 1997 01:31:08 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IGKZVZOLV4001FHU@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Sun, 16 Mar 1997 20:18:32 EST Received: from Xenon.Stanford.EDU ([171.64.64.24]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Mon, 17 Mar 1997 01:18:21 +0000 (UT) Received: (from rokicki@localhost) by Xenon.Stanford.EDU (8.8.4/8.8.4) id RAA22901; Sun, 16 Mar 1997 17:20:39 -0800 (PST) Date: Sun, 16 Mar 1997 17:20:39 -0800 (PST) From: "Tomas G. Rokicki" Subject: Re: potential virus spreadable by (la)tex To: BNB@MATH.AMS.ORG, TeX-implementors@MATH.AMS.ORG Cc: bnb@MATH.AMS.ORG, desmedt@cs.uwm.edu, keith.mcmillan@ameritech.com Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199703170120.RAA22901@Xenon.Stanford.EDU> Since this was brought up in the context of TeX, I feel it only fair to warn that the same thing is very much more easily accomplished in dvips (through the use of specials) and/or .dvipsrc files. In addition, dvips has numerous places where automatic arrays are used with gets or some other equivalent procedure, allowing sendmail-style attacks from within dvi files. So I guess it is time for me to make dvips virus-proof, probably not a trivial undertaking. In the interim, I'd like to suggest that we not make use of any features of dvips disabled by -DSECURE (primarily, the execution of commands by backtick-specials). Also, any suggestions on precisely what the security holes are and recommendations on how they can be patched would be greatly appreciated. The world has become a very different place in the last ten years, and it is about time I caught up with modern reality . . . -tom 17-Mar-1997 13:41:10-GMT,2575;000000000001 Received: from math.ams.org (MATH.AMS.ORG [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with SMTP id GAA10934 for ; Mon, 17 Mar 1997 06:41:07 -0700 (MST) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 17 Mar 1997 13:41:06 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IGLOHO0AAO001P98@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Mon, 17 Mar 1997 08:03:18 EST Received: from helios.edvz.univie.ac.at ([131.130.1.2]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Mon, 17 Mar 1997 13:02:15 +0000 (UT) Received: from AWIUNI11.EDVZ.UNIVIE.AC.AT by AWIUNI11.EDVZ.UniVie.AC.AT (IBM VM SMTP V2R2) with BSMTP id 0945; Mon, 17 Mar 1997 13:56:18 +0100 (MEZ) Received: from AWIUNI11.EDVZ.UNIVIE.AC.AT (NJE origin A8131DAL@AWIUNI11) by AWIUNI11.EDVZ.UNIVIE.AC.AT (LMail V1.2a/1.8a) with BSMTP id 4523; Mon, 17 Mar 1997 13:56:17 +0100 Date: Mon, 17 Mar 1997 13:48:36 +0100 (MEZ) From: Peter Schmitt Subject: Re: potential virus spreadable by (la)tex In-reply-to: "17 Mar 1997 12:02:47 +0100 from" <"vieth"@thphy.uni-duesseldorf.de> To: Ulrik Vieth , TeX-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <01IGLOIFG84I001P98@AXP14.AMS.ORG> >This way writing to files like ".rhosts" or ".emacs" should be >prevented. The only remaining danger might come from a TeX virus that >first writes a suitable "texmf.cnf" file to the current directory and >than tricks the user into running TeX again with the new settings from >the local config file taking precedence over the global settings. >Anyway, this should still be safer than not having any checks at all. > >Ulrik Vieth. > Just for completeness: There are even simpler methods to cause (limited) damage: \openout1 \jobname \closeout1 will overwrite the inputfile (and does not appear in the .log file) Or you can write a batch file (maybe named tex.bat or latex.bat) which might be invoked accidently and deletes *.tex (this is for DOS). Peter Schmitt a8131dal@awiuni11.edvz.univie.ac.at ----------------------------------------------------------------------------- Institute of Mathematics Strudlhofgasse 4 University of Vienna A-1090 Wien Austria 17-Mar-1997 17:22:43-GMT,4384;000000000001 Received: from math.ams.org (MATH.AMS.ORG [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with SMTP id KAA16217 for ; Mon, 17 Mar 1997 10:22:41 -0700 (MST) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 17 Mar 1997 17:22:40 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IGLWCN2MYO001JF4@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Mon, 17 Mar 1997 11:47:26 EST Received: from pppl.gov ([192.55.106.85]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Mon, 17 Mar 1997 16:47:15 +0000 (UT) Received: from carl.pppl.gov (karney@carl.pppl.gov [198.35.4.72]) by pppl.gov (8.8.5/8.8.5) with ESMTP id LAA03511; Mon, 17 Mar 1997 11:47:07 -0500 (EST) Received: (from karney@localhost) by carl.pppl.gov (8.8.5/8.8.5) id LAA27638; Mon, 17 Mar 1997 11:47:04 -0500 (EST) Date: Mon, 17 Mar 1997 11:47:04 -0500 (EST) From: Charles Karney Subject: Re: potential virus spreadable by (la)tex In-reply-to: <858525809.195705.BNB@MATH.AMS.ORG> (message from bbeeton on Sun, 16 Mar 1997 10:23:29 -0500 (EST)) To: BNB@MATH.AMS.ORG Cc: TeX-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Reply-to: karney@princeton.edu Message-id: <199703171647.LAA27638@carl.pppl.gov> MIME-version: 1.0 Content-type: text/plain; charset=US-ASCII It is important to alert the TeX user community to some elementary precautions when running TeX. Following the practice with other security alerts, this should be done fairly soon (within a week?), once some useful advice can be given. I suggest something along the lines of: ================================================================ (1) TeX has the ability to write files with (pretty much) arbitrary names in (pretty much) arbitrary directories. On multi-user systems (Unix, VMS) this is subject to the standard system protections on files. On some systems and with some implementations of TeX, further restrictions are placed on the files that can be written (e.g., some implementations for Unix prevent TeX from writing to "dot" files, .cshrc, .rhosts, etc.) (2) For this reason, you should not run TeX on a file from an untrusted source, nor should you use TeX macros from an untrusted source. (3) Similarly, on multi-user systems, you should avoid running TeX with special privileges (i.e., root on Unix systems). (Creation of "format" files should be done as a regular user.) (4) If possible, use an implementation of TeX which restricts which files can be written and which reports which files are written to in the log file. ================================================================ Left unsaid is whether the macros at CTAN are safe. The core distributions for TeX and LaTeX can almost certainly safe. But what about the other umpteen thousand macro files? Some consideration also needs to be given to other related programs (dvips, etc.). I hope too that no-one is running TeX, by default, in a mode which allows the execution of arbritrary commands (via the -shell-escape mechanism in some Unix implementations). I have posted a patch (posted to the tex-k mailing list) for Berry's implementation of TeX which restricts writes to a directory at or below the current directory or TEXMFOUTPUT. I don't think that this impacts normal use of TeX and it provides a way to run TeX safely on untrusted files via TEMPDIR=/tmp/tex.$$ umask 077 mkdir $TEMPDIR cd $TEMPDIR tex $1 Another possibility is to forbid ANY directory specification on output files and to write to TEXMFOUTPUT, if it's defined, or to the current directory, otherwise. (This is a reversal of the normal rules for TEXMFOUTPUT, which is normally only tried after the current directory.) Then, to run TeX safely you could do TEMPDIR=/tmp/tex.$$ umask 077 mkdir $TEMPDIR env TEXMFOUTPUT=$TEMPDIR tex $1 This scheme is attractive, since it is unnecessary to change TEXINPUTS to find multiple input files. Restricting TeX to write only files of the form \jobname.* is a non-starter. This would break lots of things. -- Charles Karney Plasma Physics Laboratory E-mail: Karney@Princeton.EDU Princeton University Phone: +1 609 243 2607 Princeton, NJ 08543-0451 FAX: +1 609 243 3438 17-Mar-1997 22:45:44-GMT,2228;000000000001 Received: from math.ams.org (MATH.AMS.ORG [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with SMTP id PAA24298 for ; Mon, 17 Mar 1997 15:45:42 -0700 (MST) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 17 Mar 1997 22:45:40 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IGM8ENS0MO001WIK@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Mon, 17 Mar 1997 17:32:39 EST Received: from mailrelay.tiac.net ([199.0.65.237]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Mon, 17 Mar 1997 22:32:29 +0000 (UT) Received: from yandy.tiac.net (p6.ts25.metro.MA.tiac.com [206.119.196.167]) by mailrelay.tiac.net (8.8.5/) with SMTP id RAA26217; Mon, 17 Mar 1997 17:32:30 -0500 (EST) Date: Mon, 17 Mar 1997 17:33:00 -0500 From: Y&Y Inc Subject: Re: potential virus spreadable by (la)tex X-Sender: yandy@pop.tiac.net To: BNB@MATH.AMS.ORG, TeX-implementors@MATH.AMS.ORG Cc: bnb@MATH.AMS.ORG, desmedt@cs.uwm.edu, keith.mcmillan@ameritech.com, rokicki@CS.Stanford.EDU Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <3.0.32.19970317173256.007dc660@pop.tiac.net> MIME-version: 1.0 X-Mailer: Windows Eudora Pro Version 3.0 (32) Content-type: text/plain; charset="us-ascii" Hi: DVIWindo has hyper-text support, which can be used to jump to other places in a DVI document or other DVI documents, but also has other capabilites, including launching an application. This should be disabled when used as a web browser. The ability to launch other applications is prettey handy. D.P. Story has a whole calculus course set up that way with Mathematica examples linked to via hyper-text buttons - he even does a quiz via batch files called this way! But obviously this capability can be misused. These hyper-text links get translated by DVIPSONE to pdfmarks for Acrobat Disitller. Which then has the same potential problem, since it will launch an application when the button is pushed. as far as I know there isn't much in Acrobat to protect one from mischief there. Sigh. Another thing to waste time on! Regards, Berthold. 19-Mar-1997 20:43:51-GMT,5408;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with SMTP id NAA24945 for ; Wed, 19 Mar 1997 13:43:46 -0700 (MST) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 19 Mar 1997 20:43:45 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IGOW9F9OHC001XRO@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Wed, 19 Mar 1997 15:17:16 EST Received: from pppl.gov ([192.55.106.85]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Wed, 19 Mar 1997 20:17:02 +0000 (UT) Received: from carl.pppl.gov (karney@carl.pppl.gov [198.35.4.72]) by pppl.gov (8.8.5/8.8.5) with ESMTP id PAA03952; Wed, 19 Mar 1997 15:16:59 -0500 (EST) Received: (from karney@localhost) by carl.pppl.gov (8.8.5/8.8.5) id PAA13579; Wed, 19 Mar 1997 15:16:57 -0500 (EST) Date: Wed, 19 Mar 1997 15:16:57 -0500 (EST) From: Charles Karney Subject: Re: potential virus spreadable by (la)tex In-reply-to: <33303230.2DF6@teleport.com> (message from Arthur Ogawa on Wed, 19 Mar 1997 10:39:16 -0800) To: Ogawa@teleport.com Cc: BNB@MATH.AMS.ORG, TeX-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Reply-to: karney@princeton.edu Message-id: <199703192016.PAA13579@carl.pppl.gov> MIME-version: 1.0 Content-type: text/plain; charset=US-ASCII > Date: Wed, 19 Mar 1997 10:39:16 -0800 > From: Arthur Ogawa > > Charles Karney wrote: > > > I have posted a patch (posted to the tex-k mailing list) for Berry's > > implementation of TeX which restricts writes to a directory at or below the > > current directory or TEXMFOUTPUT. > > I favor this scheme. Why not at least have the default compile-time > option for UNIX TeX cleave to this standard? > > OTOH, if someone wished to create a TeX executable that was dangerously > unrestricted, then they would be free to do so via re-compilation. Such > implementations should probably announce that they are "dangerous". With Karl Berry's implementation (texk 7.x), recompilation is not necessary. Running env openout_any=1 latex ... circumvents all the checks on output file names. This is right sort of solution, 99.9% of TeX users will use TeX in the safe default mode. The 0.1% of users who don't want the checks can do so easily and hopefully they tend to be more aware of potential security issues. > Note that the virus described by McMillan and Desmedt requires the > participation of both TeX and of Emacs; this seems to be a common > thread. I submit that in running TeX, mf, or even dvips alone it would > be pretty much impossible to cause much damage: you have to invoke a > powerful application like Emacs or a shell to cause damage (beyond > overwriting e.g. \jobname.tex). Well the more juicy file for a badly configured TeX's to write (or overwrite) is ~/.rhosts. And every cracker's favorite entry for this file is "+ +", allowing anyone on the Internet login access to the users account. A cracker is 80% of the way to gaining root access on many systems once he can log in as a regular user. [A fair fraction (> 50%) of CERT notices are about program defects that allow ordinary users to gain root access.] This then is a MAJOR security problem (but it's not a virus!). Of course, texk 7.x closes this hole by disallowing writing to dot files. (And some systems also restrict the use of rshd or monitor .rhosts for suspicious entries.) > My concrete proposal is for webc to by default prevent the app from > writing to a dot file or a file above "./". And for apps capable of > doing so to announce themselves as "potentially violent". Any thoughts? Another potentially dangerous non-dot file that a cracker might want to write to is ~/.ssh/authorized_keys This lists the public keys for people accessing the system via ssh which is a secure replacement for rsh. I don't think that texk 7.x can write to this file either because it tacks ".tex" onto file names which don't contain a ".". (The check for dot files is NOT triggered by the .ssh above, since since check is done only on file name and not on the directory name.) But to be safe, I would suggest the following: (1) no initial dot is allowed in the file name (2) no initial dot is allowed in any component of the path name (3) directory must be at or below ./ or $TEXMFOUTPUT. A couple of final thoughts for the perverse: (1) Running latex on a malicious TeX file could create article.cls or article.sty in the current directory. This would then be used the next time latex is run, since TEXINPUTS nearly always has the current directory first. Other than having TeX/LaTeX misbehave, I'm not sure what impact this would have. I suppose a hijacked letter.cls could insert an inappropriate greeting or closing in a letter which a user who carefully read his letter in an editor might miss. (2) The default Unix setup allows PK files to be created on the fly. Presumably this process could be subverted and a copy of cmr10.300pk inserted with various letters permuted. It's difficult to see how this would go unnoticed for very long. -- Charles Karney Plasma Physics Laboratory E-mail: Karney@Princeton.EDU Princeton University Phone: +1 609 243 2607 Princeton, NJ 08543-0451 FAX: +1 609 243 3438 19-Mar-1997 21:27:29-GMT,2673;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with SMTP id OAA25970 for ; Wed, 19 Mar 1997 14:27:28 -0700 (MST) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 19 Mar 1997 21:27:26 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IGOYI3T0GG002714@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Wed, 19 Mar 1997 16:21:37 EST Received: from kim.teleport.com ([192.108.254.26]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Wed, 19 Mar 1997 21:21:22 +0000 (UT) Received: from 205.138.246.20 (ppp1024.inreach.com [205.138.245.24]) by kim.teleport.com (8.8.5/8.7.3) with SMTP id NAA27749; Wed, 19 Mar 1997 13:21:06 -0800 (PST) Date: Wed, 19 Mar 1997 13:22:49 -0800 From: Arthur Ogawa Subject: Re: potential virus spreadable by (la)tex To: karney@princeton.edu Cc: BNB@MATH.AMS.ORG, TeX-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Reply-to: Ogawa@teleport.com Message-id: <3330585F.EB0@teleport.com> Organization: TeX Consultants MIME-version: 1.0 X-Mailer: Mozilla 3.0Gold (Macintosh; I; PPC) Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit References: <199703192016.PAA13579@carl.pppl.gov> Charles Karney wrote: > > With Karl Berry's implementation (texk 7.x), recompilation is not > necessary. Running > > env openout_any=1 latex ... > > circumvents all the checks on output file names. This is right sort of > solution... Can we be assured that the switch to openout_any=1 is something a TeX document would never be able to accomplish? > But to be safe, I would suggest the following: > (1) no initial dot is allowed in the file name > (2) no initial dot is allowed in any component of the path name > (3) directory must be at or below ./ or $TEXMFOUTPUT. You have my vote. > A couple of final thoughts for the perverse: > (1) Running latex on a malicious TeX file could create article.cls... > (2) The default Unix setup allows PK files to be created on the fly. > Presumably this process could be subverted... How very creative! Basically, TeX must be able to read in files that it has written out in an earlier run, so there is little way I can see to prevent such schemes. Their effect is solely on files, albeit important, in "./", however. -- Arthur Ogawa/TeX Consultants voice: +1 209 561-4585 Fax: +1 209 561-4584 emailto://ogawa@teleport.com http://www.teleport.com/~ogawa ftp://ftp.teleport.com/users/ogawa PGP key: finger -l ogawa@teleport.com 19-Mar-1997 22:54:40-GMT,3740;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with SMTP id PAA28188 for ; Wed, 19 Mar 1997 15:54:38 -0700 (MST) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 19 Mar 1997 22:54:37 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IGP0XPFF9C001T4K@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Wed, 19 Mar 1997 17:31:27 EST Received: from pppl.gov ([192.55.106.85]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Wed, 19 Mar 1997 22:31:09 +0000 (UT) Received: from carl.pppl.gov (karney@carl.pppl.gov [198.35.4.72]) by pppl.gov (8.8.5/8.8.5) with ESMTP id RAA13000; Wed, 19 Mar 1997 17:31:05 -0500 (EST) Received: (from karney@localhost) by carl.pppl.gov (8.8.5/8.8.5) id RAA17355; Wed, 19 Mar 1997 17:31:01 -0500 (EST) Date: Wed, 19 Mar 1997 17:31:01 -0500 (EST) From: Charles Karney Subject: Re: potential virus spreadable by (la)tex In-reply-to: <3330585F.EB0@teleport.com> (message from Arthur Ogawa on Wed, 19 Mar 1997 13:22:49 -0800) To: Ogawa@teleport.com Cc: BNB@MATH.AMS.ORG, TeX-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Reply-to: karney@princeton.edu Message-id: <199703192231.RAA17355@carl.pppl.gov> MIME-version: 1.0 Content-type: text/plain; charset=US-ASCII > Date: Wed, 19 Mar 1997 13:22:49 -0800 > From: Arthur Ogawa > > Running > > env openout_any=1 latex ... > > circumvents all the checks on output file names. > Can we be assured that the switch to openout_any=1 is something a TeX > document would never be able to accomplish? The document can't set openout_any, unless tex/latex is invoked with latex -shell-escape (and even then it would be indirect, since an inferior process can't change a parent's environment). This flag allows the execution of arbritary commands and so you should be careful using this. Again 99.9% of TeX users won't know about this. And the 0.1% who use it should know they are playing with fire if they run latex like this on a file from an untrusted source. > > (1) Running latex on a malicious TeX file could create article.cls... > How very creative! Basically, TeX must be able to read in files that it > has written out in an earlier run, so there is little way I can see to > prevent such schemes. Their effect is solely on files, albeit important, > in "./", however. I suppose TEXINPUTS could have "." at the end (like the usual advice on the Unix PATH nowadays). However, users sometimes want their macro files to override the system ones. In this case, they could use something like TEXINPUTS=~/tex:/usr/local/share/texmf/tex//:. (Compare PATH=$HOME/bin:/usr/bin:/usr/X11R6/bin:/usr/local/bin:. ) This would need to be changed in texmf.cnf. I won't be doing this, since (a) it will break things for some people and (b) it won't be completely effective because plenty of users on our systems are setting TEXINPUTS directly themselves. Incidentally, I am logging the names of the output files written by TeX. My guess is that no-one is writing output files to other than the current directory. If my guess is confirmed after a couple of weeks, then I'll modify the openoutok routine to allow only output files in "." or $TEXMFOUTPUT (and NOT in their subdirectories). This will make it somewhat easier for users to see if strange files have been written by TeX. -- Charles Karney Plasma Physics Laboratory E-mail: Karney@Princeton.EDU Princeton University Phone: +1 609 243 2607 Princeton, NJ 08543-0451 FAX: +1 609 243 3438 24-Mar-1997 19:59:30-GMT,2878;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.4/8.8.4) with SMTP id MAA16787 for ; Mon, 24 Mar 1997 12:59:28 -0700 (MST) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 24 Mar 1997 19:59:27 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IGVU4O4WVK000MNN@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Mon, 24 Mar 1997 14:31:40 EST Received: from pppl.gov ([192.55.106.85]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Mon, 24 Mar 1997 19:31:29 +0000 (UT) Received: from carl.pppl.gov (karney@carl.pppl.gov [198.35.4.72]) by pppl.gov (8.8.5/8.8.5) with ESMTP id OAA26703; Mon, 24 Mar 1997 14:31:26 -0500 (EST) Received: (from karney@localhost) by carl.pppl.gov (8.8.5/8.8.5) id OAA03089; Mon, 24 Mar 1997 14:31:24 -0500 (EST) Date: Mon, 24 Mar 1997 14:31:24 -0500 (EST) From: Charles Karney Subject: Re: potential virus spreadable by (la)tex In-reply-to: <858525809.195705.BNB@MATH.AMS.ORG> (message from bbeeton on Sun, 16 Mar 1997 10:23:29 -0500 (EST)) To: BNB@MATH.AMS.ORG Cc: TeX-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Reply-to: karney@princeton.edu Message-id: <199703241931.OAA03089@carl.pppl.gov> MIME-version: 1.0 Content-type: text/plain; charset=US-ASCII Here are two other ideas for restricting what file names TeX can write to. I don't think either is really worth pursuing, but perhaps others would like to comment (1) restrict the extension of the filenames. Thus allow latex only to write to *.aux, *.lot, *.loc, etc. files. kpathsea would allow you easily to specify separate lists of extensions of each program. This would hopefully limit the scope of any damage to the product of TeX itself. I view this as a rather intrusive change since either macro writers will have to conform their file names to a "standard" or else the user/system administrator will have to keep on updating the list of allowable extensions. (2) restrict the allowed characters in pathnames. With Unix any non-null character is allowed in pathname. Possibly the presence of characters special to the shell would break poorly written system scripts. A potential attack along these lines is to insert a newline into a filename thereby causing purge scripts like this one find /tmp -atime +10 -print | xargs rm -f to delete the wrong files (with the GNU utilities this should read 'find ... -print0 | xargs -0 rm -f'). If any restriction along these lines were implemented, it should allow accented characters. -- Charles Karney Plasma Physics Laboratory E-mail: Karney@Princeton.EDU Princeton University Phone: +1 609 243 2607 Princeton, NJ 08543-0451 FAX: +1 609 243 3438 12-Jun-1997 15:27:37-GMT,1625;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id JAA20951 for ; Thu, 12 Jun 1997 09:27:35 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 12 Jun 1997 15:27:35 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IJZE5260CG0002TF@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Thu, 12 Jun 1997 11:02:47 EST Received: from alpha1.rhbnc.ac.uk ([134.219.201.113]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Thu, 12 Jun 1997 15:02:28 +0000 (UT) Date: Thu, 12 Jun 1997 16:03:23 +0100 From: Philip Taylor (RHBNC) Subject: Tex the program, module 103: print_scaled. To: tex-implementors@MATH.AMS.ORG Cc: CHAA006@vms.rhbnc.ac.uk Errors-to: tex-implementors-request@MATH.AMS.ORG Reply-to: P.Taylor@vms.rhbnc.ac.uk Message-id: <970612160323.5a18@vms.rhbnc.ac.uk> According to my copy of Vol.~B, the 8th line of print_scaled reads: repeat if delta > unity then s <-- s + '100000 - (delta div 2); { } this reading is confirmed by "tex82.bug" as found at CTAN, where I find repeat if delta>unity then s:=s+@'100000-(delta div 2); {round the last digit} However, in TeX 3.14159, I find instead: repeat if delta>unity then s:=s+@'100000-50000; {round the last digit} Since delta is a variable, the expression (delta div 2) cannot possibly be flattened to 50000; could anyone explain how this difference has come about, and for what reason? Philip Taylor, RHBNC. 16-Jun-1997 15:45:09-GMT,1974;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id JAA15598 for ; Mon, 16 Jun 1997 09:44:58 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 16 Jun 1997 15:44:57 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IK502G6ASW000KLB@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Mon, 16 Jun 1997 11:24:23 EST Received: from mailrelay.tiac.net ([199.0.65.237]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Mon, 16 Jun 1997 15:24:10 +0000 (UT) Received: from p2.ts7.bedfo.MA.tiac.com (p2.ts7.bedfo.MA.tiac.com [207.60.9.227]) by mailrelay.tiac.net (8.8.5/) with SMTP id LAA05520; Mon, 16 Jun 1997 11:21:46 -0400 (EDT) Date: Mon, 16 Jun 1997 11:24:20 -0400 From: support@YandY.com (Y&Y Inc) Subject: Re: Tex the program, module 103: print_scaled. X-Sender: yandy@pop.tiac.net (Unverified) To: tex-implementors@MATH.AMS.ORG Cc: CHAA006@vms.rhbnc.ac.uk Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199706161521.LAA05520@mailrelay.tiac.net> MIME-version: 1.0 X-Mailer: Windows Eudora Version 2.0.3 Content-type: text/plain; charset="us-ascii" Hi: re: `CM' font metrics Is there some good reason why the EC font metrics don't match the CM font metrics for the characters that are common to both font sets (except where intentional changes were made)? The EC advance widths etc. are consistently about 1/4000 th smaller than the corresponding CM advance widths. Seems like an unnecccesary inconsistency to me. For example, it makes for potential inaccuracies when switching between EC and `fake EC' obtained via VF from CM. The CM font advance widths themselves are on average 2 scaled points too large, but that is a miniscule discrepancy compared to the above. Regards, Berthold Horn Berthold K. P. Horn Concord, MA 01742 16-Jun-1997 19:12:19-GMT,2784;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id NAA20961 for ; Mon, 16 Jun 1997 13:12:16 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 16 Jun 1997 19:12:15 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IK57BWSTXC000MA1@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Mon, 16 Jun 1997 14:52:49 EST Received: from elessar.ics.muni.cz ([147.251.4.10]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Mon, 16 Jun 1997 18:52:37 +0000 (UT) Received: from daeron (daeron.fi.muni.cz [147.251.48.91]) by elessar.ics.muni.cz (8.8.5/8.8.5) with SMTP id UAA22302; Mon, 16 Jun 1997 20:52:39 +0200 (MET DST) Received: by daeron id AA11491 (5.67b8/IDA-1.4.4); Mon, 16 Jun 1997 20:52:35 +0200 Date: Mon, 16 Jun 1997 20:52:35 +0200 (MET DST) From: Petr Sojka Subject: Re: Tex the program, module 103: print_scaled. In-reply-to: <970612160323.5a18@vms.rhbnc.ac.uk> To: P.Taylor@vms.rhbnc.ac.uk Cc: tex-implementors@MATH.AMS.ORG, thanh@informatics.muni.cz (Han The Thanh) Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199706161852.AA11491@daeron> Organization: Masaryk University, Brno, The Czech Republic, www.fi.muni.cz MIME-version: 1.0 X-Mailer: ELM [version 2.4ME+ PL25 (25)] Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7bit Postal-Address: Faculty of Informatics, Botanicka 68a, 60200 Brno Telephone: +420-5-41512352 (my room), +420-5-41512329 (secretary), Fax: +420-5-41212568, 41213219 "RHBNC wrote:" : According to my copy of Vol.~B, the 8th line of print_scaled reads: : : repeat if delta > unity then s <-- s + '100000 - (delta div 2); { } : : this reading is confirmed by "tex82.bug" as found at CTAN, where I find : : repeat if delta>unity then s:=s+@'100000-(delta div 2); {round the last digit} : : However, in TeX 3.14159, I find instead: : : repeat if delta>unity then s:=s+@'100000-50000; {round the last digit} : : Since delta is a variable, the expression (delta div 2) cannot possibly : be flattened to 50000; could anyone explain how this difference has come : about, and for what reason? The reason is optimization. Even as this leads to less readable code, DEK don't relies on clever compilers and optimizes by hand. It can be proven given the cycle invariant that the condition is fulfilled after exactly five iterations and thus delta=10*10*10*10*10*10 and delta div 2 = 100000 div 2 = 50000. This example showes that getting rid of these optimizations woudn't be trivial in future possible reimplementation of TeX. Petr : Philip Taylor, RHBNC. 17-Jun-1997 7:36:09-GMT,2158;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id BAA12237 for ; Tue, 17 Jun 1997 01:36:06 -0600 (MDT) From: KNAPPEN@MZDMZA.ZDV.UNI-MAINZ.DE Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 17 Jun 1997 07:36:01 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IK5X80Z2GW000P4O@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Tue, 17 Jun 1997 03:14:07 EST Received: from dzdmzc.zdv.Uni-Mainz.DE ([134.93.8.34]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Tue, 17 Jun 1997 07:13:57 +0000 (UT) Received: from MZDMZA.ZDV.UNI-MAINZ.DE by MZDMZA.ZDV.UNI-MAINZ.DE (PMDF V5.0-4 #22141) id <01IK69MSYUCGH63WP9@MZDMZA.ZDV.UNI-MAINZ.DE>; Tue, 17 Jun 1997 09:15:10 +0100 Date: Tue, 17 Jun 1997 09:15:10 +0100 Subject: Re: Tex the program, module 103: print_scaled. To: support@YandY.com Cc: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <01IK69MSYZ1UH63WP9@MZDMZA.ZDV.UNI-MAINZ.DE> X-VMS-To: IN%"support@YandY.com" X-VMS-Cc: IN%"tex-implementors@MATH.AMS.ORG" MIME-version: 1.0 Content-type: TEXT/PLAIN; CHARSET=US-ASCII Content-transfer-encoding: 7BIT Berthold Horn wrote: >>>The EC advance widths etc. are consistently about 1/4000 th smaller than the corresponding CM advance widths. Seems like an unnecccesary inconsistency to me. For example, it makes for potential inaccuracies when switching between EC and `fake EC' obtained via VF from CM. Berthold, which fonts did you compare? Only one of them (e.g. cmr10 vs. ecrm1000) or a larger set considering several design sizes? As fas as I know, the macros governing the advance width are the same for the cm fonts and the ec fonts. A potential source of those small difference is the interpolation mechanism used by the ec fonts. It introduces small rounding errors which also affect the starting values given in the parameter files. I found this phenomenon being relevant while designing the ec slitex typewriter fonts. --J"org Knappen 17-Jun-1997 8:35:52-GMT,4080;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id CAA13510 for ; Tue, 17 Jun 1997 02:35:50 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 17 Jun 1997 08:35:49 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IK5ZEVVQMO000PED@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Tue, 17 Jun 1997 04:17:00 EST Received: from beach.frankfurt.netsurf.de ([194.64.181.2]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Tue, 17 Jun 1997 08:16:45 +0000 (UT) Received: from puma.npc.de (deck-75.frankfurt.netsurf.de [194.64.181.107]) by beach.frankfurt.netsurf.de (8.6.12/8.6.12) with ESMTP id KAA00686; Tue, 17 Jun 1997 10:14:47 +0200 Received: (from schrod@localhost) by puma.npc.de (8.6.12/8.6.12) id JAA12919; Tue, 17 Jun 1997 09:27:19 +0200 Received: (from schrod@localhost) by puma.npc.de (8.6.12/8.6.12) id JAA12917; Tue, 17 Jun 1997 09:27:18 +0200 Date: Tue, 17 Jun 1997 09:27:18 +0200 From: Joachim Schrod Subject: Re: Tex the program, module 103: print_scaled. In-reply-to: <199706161852.AA11491@daeron> To: Petr Sojka Cc: P.Taylor@vms.rhbnc.ac.uk, tex-implementors@MATH.AMS.ORG, thanh@informatics.muni.cz (Han The Thanh) Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199706170727.JAA12917@puma.npc.de> MIME-version: 1.0 (generated by tm-edit 7.95) Content-type: text/plain; charset=US-ASCII References: <970612160323.5a18@vms.rhbnc.ac.uk> <199706161852.AA11491@daeron> >>>>> "PS" == Petr Sojka writes: PS> "RHBNC wrote:" PS> : According to my copy of Vol.~B, the 8th line of print_scaled reads: PS> : PS> : repeat if delta > unity then s <-- s + '100000 - (delta div 2); { } PS> : PS> : this reading is confirmed by "tex82.bug" as found at CTAN, where I find PS> : PS> : repeat if delta>unity then s:=s+@'100000-(delta div 2); {round the last digit} PS> : PS> : However, in TeX 3.14159, I find instead: PS> : PS> : repeat if delta>unity then s:=s+@'100000-50000; {round the last digit} PS> : PS> : Since delta is a variable, the expression (delta div 2) cannot possibly PS> : be flattened to 50000; could anyone explain how this difference has come PS> : about, and for what reason? PS> The reason is optimization. Even as this leads to less readable PS> code, DEK don't relies on clever compilers and optimizes by hand. PS> It can be proven given the cycle invariant that PS> the condition is fulfilled after exactly five iterations PS> and thus delta=10*10*10*10*10*10 and delta div 2 = 100000 div 2 = 50000. PS> This example showes that getting rid of these optimizations PS> woudn't be trivial in future possible reimplementation of TeX. Why? Exactly this example, invariant removal, is handled by most good compilers automatically. It's a good point _for_ writing readable, maintainable code, where profis like Phil are not baffled but that is still fast. Honestly, I worry more about _human_ productivity than _machine_ productivity, the latter gets faster all the time, the former doesn't increase. Having seen what a good optimizing compiler does to a source code, and having read DEK's code (that _hinders_ optimization with all his global variables and parameter/result passing by side effects) I'd bet that a clean implementation in C with following optimization is _faster_ than his code compiled on a good Pascal compiler. Note that I didn't say that it's a sensible thing to recode TeX with exactly the same algorithms in C. I'd happily spend a few machine cycles (maybe even a few more ;-) for a more modern system, both in terms of functionality and implementation. Cheers, Joachim -- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= Joachim Schrod Email: jschrod@acm.org Net & Publication Consultance GmbH Tel.: +49-6074-861530 Roedermark, Germany Fax: +49-6074-861531 17-Jun-1997 17:24:33-GMT,2494;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id LAA24372 for ; Tue, 17 Jun 1997 11:24:31 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 17 Jun 1997 17:24:29 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01IK6HX6X9SW000SSU@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Tue, 17 Jun 1997 13:07:06 EST Received: from mailrelay.tiac.net ([199.0.65.237]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Tue, 17 Jun 1997 17:06:55 +0000 (UT) Received: from denali (p14.ts3.bedfo.MA.tiac.com [207.60.9.111]) by mailrelay.tiac.net (8.8.5/) with SMTP id NAA28932; Tue, 17 Jun 1997 13:07:17 -0400 (EDT) Date: Tue, 17 Jun 1997 13:06:57 -0400 From: Y&Y Inc Subject: Re: Tex the program, module 103: print_scaled. In-reply-to: <01IK69MSYZ1UH63WP9@MZDMZA.ZDV.UNI-MAINZ.DE> X-Sender: yandy@pop.tiac.net (Unverified) To: KNAPPEN@MZDMZA.ZDV.UNI-MAINZ.DE Cc: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <3.0.1.32.19970617130657.009d4680@pop.tiac.net> MIME-version: 1.0 X-Mailer: Windows Eudora Pro Version 3.0.1 (32) Content-type: text/plain; charset="us-ascii" Hi: At 09:15 AM 97/06/17 +0100, you wrote: >>>>The EC advance widths etc. are consistently about 1/4000 th smaller than >the corresponding CM advance widths. Seems like an unnecccesary >inconsistency to me. For example, it makes for potential inaccuracies >when switching between EC and `fake EC' obtained via VF from CM. >Berthold, which fonts did you compare? Only one of them (e.g. cmr10 vs. >ecrm1000) or a larger set considering several design sizes? >As fas as I know, the macros governing the advance width are the same for >the cm fonts and the ec fonts. A potential source of those small difference >is the interpolation mechanism used by the ec fonts. It introduces small >rounding errors which also affect the starting values given in the >parameter files. I found this phenomenon being relevant while designing the >ec slitex typewriter fonts. The only one I compared in detail is ecrm1000 versus cmr10. The reason I compared this one in detail is that I found the discrepancy in some other fonts that I was looking at like, ecbx1000 etc. All of the ones I was looking at were at 10pt design size. Regards, Berthold. 17-Jul-1997 11:03:22-GMT,2040;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id FAA08943 for ; Thu, 17 Jul 1997 05:03:21 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 17 Jul 1997 11:03:16 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #16534) with SMTP id <01ILC11PD1Z40000UN@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Thu, 17 Jul 1997 06:36:49 EST Received: from xerxes.thphy.uni-duesseldorf.de ([134.99.64.123]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Thu, 17 Jul 1997 10:36:40 +0000 (UT) Received: from macbeth.uni-duesseldorf.de (macbeth.thphy.uni-duesseldorf.de) by thphy.uni-duesseldorf.de (4.1/SMI-4.1) id AA23239; Thu, 17 Jul 1997 12:29:13 +0200 Received: by macbeth.uni-duesseldorf.de (SMI-8.6/SMI-SVR4) id MAA29417; Thu, 17 Jul 1997 12:29:02 +0200 Date: Thu, 17 Jul 1997 12:29:02 +0200 From: Ulrik Vieth Subject: TeX History Alert To: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199707171029.MAA29417@macbeth.uni-duesseldorf.de> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1 Content-transfer-encoding: 8bit While I was once again reading the Errorlog of TeX I noticed that another important day in TeX's history almost went unnoticed. On July 15, 1982 (excactly 15 years and 2 days ago), Knuth began debugging TeX 82, presumably version -100.0. Time to celebrate! Cheers, Ulrik. P.S. Whoever picked the date for EuroTeX'98 (March 29--31, 1998) apparently also made a good choice. If you look into the Errorlog for March 29, 1978, you'll find that it says: 29 Mar 1978 [...] # After these corrections, the test routine worked\thinspace\dots\ I feel that \TeX\ is now pretty well debugged (except perhaps for error recovery)---it's time to celebrate! Time for a "20 years of TeX 78" party at EuroTeX'98 next year? 17-Jul-1997 18:56:12-GMT,2061;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id MAA19555 for ; Thu, 17 Jul 1997 12:56:10 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 17 Jul 1997 18:56:05 UT Received: from AXP14.AMS.ORG by AXP14.AMS.ORG (PMDF V5.1-8 #16534) id <01ILCHRIBIDC00018I@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Thu, 17 Jul 1997 14:40:10 EST Date: Thu, 17 Jul 1997 14:40:01 -0400 (EDT) From: bbeeton Subject: re-sending -- TeX History Alert To: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <869164801.118779.BNB@MATH.AMS.ORG> MIME-version: 1.0 Content-type: TEXT/PLAIN; CHARSET=US-ASCII Mail-system-version: sorry if this is a duplicate, however, last night there seem to have been some system-wide gremlins at work, and i was blessed with a flock of failed addresses for usually reliable sites. so at the risk of annoying some of you, i am re-sending ulrik's message. -- bb -------------------- Date: Thu, 17 Jul 1997 12:29:02 +0200 From: Ulrik Vieth To: tex-implementors@MATH.AMS.ORG Subject: TeX History Alert While I was once again reading the Errorlog of TeX I noticed that another important day in TeX's history almost went unnoticed. On July 15, 1982 (excactly 15 years and 2 days ago), Knuth began debugging TeX 82, presumably version -100.0. Time to celebrate! Cheers, Ulrik. P.S. Whoever picked the date for EuroTeX'98 (March 29--31, 1998) apparently also made a good choice. If you look into the Errorlog for March 29, 1978, you'll find that it says: 29 Mar 1978 [...] # After these corrections, the test routine worked\thinspace\dots\ I feel that \TeX\ is now pretty well debugged (except perhaps for error recovery)---it's time to celebrate! Time for a "20 years of TeX 78" party at EuroTeX'98 next year? 3-Sep-1997 6:43:54-GMT,6463;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id AAA11711 for ; Wed, 3 Sep 1997 00:43:48 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 3 Sep 1997 06:43:47 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IN6TRY045S000GRA@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Wed, 3 Sep 1997 02:12:07 EST Received: from yarkon.eng.tau.ac.il ([132.66.48.2]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Wed, 03 Sep 1997 06:11:50 +0000 (UT) Date: Wed, 03 Sep 1997 09:11:10 +0300 (IDT) From: Michael Slavutin Subject: Roman Numerals In-reply-to: <199708091724.AA18859@eng.tau.ac.il> To: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Reply-to: slavutin@eng.tau.ac.il Message-id: MIME-version: 1.0 Content-type: MULTIPART/MIXED; BOUNDARY="1921388315-362165037-873267070=:3852" This message is in MIME format. The first part should be readable text, while the remaining parts are likely unreadable without MIME-aware tools. Send mail to mime@docserver.cac.washington.edu for more info. --1921388315-362165037-873267070=:3852 Content-Type: TEXT/PLAIN; charset=US-ASCII Dear TeX community! The current algorithm of converting the numbers into roman ones that is used in TeX and is described in TEX.WEB (\paragraph 69) gives the following results: for gives instead of 1990 mcmxc mxm 1995 mcmxcv mvm 1999 mcmxcix mim Really, I don't know what is right, but I want to propose another algorithm that gives the results on the right column of the above table. It is written in PASCAL and the program called ROMAN.PAS is attached. Sincerely yours, Michael Slavutin, slavutin@eng.tau.ac.il --1921388315-362165037-873267070=:3852 Content-Type: TEXT/PLAIN; charset=US-ASCII; name="roman.pas" Content-Transfer-Encoding: BASE64 Content-ID: Content-Description: e1RIRSBGT0xMT1dJTkcgUFJPR1JBTSBDT01QVVRFUyBUSEUgUk9NQU4gTlVN RVJBTCBPRiBBIEdJVkVODQpQT1NJVElWRSBJTlRFR0VSfQ0KDQpQUk9HUkFN IFJPTUFOOw0KICB7RklSU1QgT0YgQUxMIElUIERFQ0xBUkVTIFRIRSBCQVNJ QyBTWU1CT0xTIE9GIFJPTUFOIENPVU5USU5HIEFTICJzeW1ib2wiDQogICBB UlJBWSBPRiBDSEFSQUNURVJTIEFORCBUSEVJUiBOVU1FUklDQUwgRVFVSVZB TEVOVFMgQVMgImNvbnN0YW50Ig0KICAgQVJSQVkgT0YgSU5URUdFUlN9DQog IENPTlNUDQogICAgY29uc3RhbnQgOiBBUlJBWVsxLi43XU9GIElOVEVHRVIg PSAoMSw1LDEwLDUwLDEwMCw1MDAsMTAwMCk7DQogICAgc3ltYm9sICAgOiBB UlJBWVsxLi43XU9GIENIQVIgPSAnaXZ4bGNkbSc7DQogIHtUSEUgVkFSSUFC TEVTIElOIFVTRSBBUkU6DQogICAgICAgICAgICAgICAgICAgICAgICAgICAg biBJUyBUSEUgSU5URUdFUiBOVU1CRVIgVE8gQkUgQ09OVkVSVEVEDQogICAg ICAgICAgICAgICAgICAgICAgICAgICAgayBJUyBUSEUgSU5ERVggSU5UTyBU SEUgUk9NQU4gQVJSQVlTDQogICAgICAgICAgICAgICAgICAgICAgICAgICAg aiBJUyBBTk9USEVSIElOREVYIElOVE8gVEhFIFJPTUFOIEFSUkFZU30NCiAg VkFSIG46aW50ZWdlcjsNCiAgICAgIGs6aW50ZWdlcjsNCiAgICAgIGo6aW50 ZWdlcjsNCiAgTEFCRUwgMTAsIDIwOw0KQkVHSU4NCiAge1JFQUQgbiBBTkQg U1RBUlQgQlVJTERJTkcgVEhFIFJPTUFOIEVRVUlWQUxFTlQgRlJPTSBUSEUg QkVHSU5OSU5HDQogICAoMTAwMCl9DQogIFJlYWQobik7DQogIGs6PTc7DQog IHtXRSBVU0UgVEhFIEZBQ1QgVEhBVCBUSEUgUk9NQU4gQVJSQVkgSVMgQlVJ TEQgRlJPTSBUSEUgTlVNQkVSUyBUSEFUDQogICBBUkUgT0YgVEhFIEtJTkQg MTBeTCBBTkQgNSoxMF5MIFdIRVJFIEwgSVMgQSBOT04tTkVHQVRJVkUgSU5U RUdFUi4NCiAgIFRIRSBPREQgRU5UUklFUyBBUkUgT0YgVEhFIFRZUEUgMTBe TCwgVEhVUyBUSEVZIENBTiBCRSBSRVBFQVRFRCBJTg0KICAgVEhFIFJPTUFO IE5VTUVSQUx9DQogIDEwOldISUxFICgobj49Y29uc3RhbnRba10pIEFORCBP ZGQoaykpIERPDQogICAgICAgQkVHSU4NCiAgICAgICAgIFdyaXRlKHN5bWJv bFtrXSk7DQogICAgICAgICBuOj1uLWNvbnN0YW50W2tdOw0KICAgICAgIEVO RDsNCiAgICAge05FWFQgV0UgSEFWRSBUTyBDSEVDSyBJRiBuIElTIE5PVCBF WEhBVVNURUQgQUxSRUFEWSAoT1IgSUYgSVQgV0FTDQogICAgICBORUdBVElW RSBJTiBUSEUgRklSU1QgVElNRS4gSUYgSVQgSVMsIFRIRSBDT01QVVRBVElP TiBJUyBURVJNSU5BVEVEfQ0KICAgICBJRiBuPD0wDQogICAgICAgVEhFTiBH T1RPIDIwOw0KICAgICB7VEhFTiBXRSBDSEVDSyBJRiBUSEUgUkVNQUlOREVS IE9GIG4gRklUUyBJTlRPIE9ORSBPRiBUSEUgSU5URVJWQUxTDQogICAgICBP RiBUWVBFIChjb25zdGFudFtrXS1jb25zdGFudFtqXSkuLmNvbnN0YW50W2td IChGT1IgRVhBTVBMRSwgSUYNCiAgICAgIG49OTIxLCBUSEUgRklSQ1QgVEhJ TkcgVE8gRElTUExBWSBJUyAiY20iLiBBRlRFUiBFVkVSWSBTVUNDRVNGVUxM DQogICAgICBDSEVDSyBuIElTIFVQREFURUQgQU5EIFNPIElTIGsgQUNDT1JE SU5HIFRPIFRIRSBSRU1BSU5ERVIgT0Ygbn0NCiAgICAgRk9SIGo6PTEgVE8g ay0xIERPDQogICAgICAgQkVHSU4NCiAgICAgICAgIElGIChuPj0oY29uc3Rh bnRba10tY29uc3RhbnRbal0pKQ0KICAgICAgICAgICBUSEVOIElGIChqPD5r LTEpDQogICAgICAgICAgICAgICAgICB7VEhFIENBU0VTIFNVQ0ggQVMgbj05 MjAgQVJFIFRSRUFURUQgSEVSRSBTSU5DRSBUSElTIElTDQogICAgICAgICAg ICAgICAgICAgImNtLi4uIn0NCiAgICAgICAgICAgICAgICAgIFRIRU4gQkVH SU4NCiAgICAgICAgICAgICAgICAgICAgICAgICBXcml0ZShzeW1ib2xbal0p Ow0KICAgICAgICAgICAgICAgICAgICAgICAgIFdyaXRlKHN5bWJvbFtrXSk7 DQogICAgICAgICAgICAgICAgICAgICAgICAgbjo9bitjb25zdGFudFtqXS1j b25zdGFudFtrXTsNCiAgICAgICAgICAgICAgICAgICAgICAgICBrOj1qLTE7 DQogICAgICAgICAgICAgICAgICAgICAgICAgR09UTyAxMDsNCiAgICAgICAg ICAgICAgICAgICAgICAgRU5EDQogICAgICAgICAgICAgICAgICBFTFNFIElG IE9kZChrKQ0KICAgICAgICAgICAgICAgICAgICAgICAgIHtBTkQgVEhFIENB U0VTIFNVQ0ggQVMgbj01MjAgSEVSRSBTSU5DRSBUSElTIElTDQogICAgICAg ICAgICAgICAgICAgICAgICAgICJkLi4uIn0NCiAgICAgICAgICAgICAgICAg ICAgICAgICBUSEVOIEJFR0lODQogICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgIFdyaXRlKHN5bWJvbFtqXSk7DQogICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgIG46PW4tY29uc3RhbnRbal07DQogICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgIGs6PWotMTsNCiAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgICAgR09UTyAxMDsNCiAgICAgICAgICAgICAgICAgICAg ICAgICAgICAgIEVORA0KICAgICAgICAgICAgICAgICAgICAgICAgIHtIT1dF VkVSIElGIG4gSVMgU09NRVRISU5HIExJS0UgNDgxLCBUSEUgU0NSSVBUDQog ICAgICAgICAgICAgICAgICAgICAgICAgICJzeW1ib2xbal0iInN5bWJvbFtr XSIgSVMgUEVSRkVDVExZIFJJR0hUIFNJTkNFDQogICAgICAgICAgICAgICAg ICAgICAgICAgIFRISVMgSVMgImNkLi4uIn0NCiAgICAgICAgICAgICAgICAg ICAgICAgICBFTFNFIEJFR0lODQogICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgIFdyaXRlKHN5bWJvbFtqXSk7DQogICAgICAgICAgICAgICAgICAg ICAgICAgICAgICAgIFdyaXRlKHN5bWJvbFtrXSk7DQogICAgICAgICAgICAg ICAgICAgICAgICAgICAgICAgIG46PW4rY29uc3RhbnRbal0tY29uc3RhbnRb a107DQogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIGs6PWotMTsN CiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgR09UTyAxMDsNCiAg ICAgICAgICAgICAgICAgICAgICAgICAgICAgIEVORDsNCiAgICAgICBFTkQ7 DQogICAgICAgazo9ay0xOw0KICAgICAgIEdPVE8gMTA7DQogICAgICAgMjA6 V3JpdGVsbignJyk7DQpFTkQu --1921388315-362165037-873267070=:3852-- 3-Sep-1997 12:30:17-GMT,1940;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id GAA17770 for ; Wed, 3 Sep 1997 06:30:09 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 3 Sep 1997 12:30:08 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IN769J3GSG000HNZ@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Wed, 3 Sep 1997 08:08:59 EST Received: from heaton.cl.cam.ac.uk ([128.232.32.11]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Wed, 03 Sep 1997 12:08:49 +0000 (UT) Received: from ouse.cl.cam.ac.uk [128.232.33.87] (rf) by heaton.cl.cam.ac.uk with esmtp (Exim 1.70 #1) id 0x6EEg-00045S-00; Wed, 03 Sep 1997 13:08:38 +0100 Date: Wed, 03 Sep 1997 13:08:36 +0100 From: Robin Fairbairns Subject: Re: Roman Numerals In-reply-to: "Your message of Wed, 03 Sep 1997 09:11:10 +0300." To: slavutin@eng.tau.ac.il Cc: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: > The current algorithm of converting the numbers into roman ones > that is used in TeX and is described in TEX.WEB (\paragraph 69) > gives the following results: > for gives instead of > 1990 mcmxc mxm > 1995 mcmxcv mvm > 1999 mcmxcix mim > Really, I don't know what is right, Knuth's original is what's most commonly used. > but I want to propose another algorithm > that gives the results on the right column of the above table. It is > written in > PASCAL and the program called ROMAN.PAS is attached. Since this would give such strange-looking results, I would prefer that it not be used (quite apart from its potential effect on compatibility). Robin Fairbairns Cambridge 3-Sep-1997 12:44:01-GMT,5654;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id GAA18009 for ; Wed, 3 Sep 1997 06:43:53 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 3 Sep 1997 12:43:39 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IN76WE2374000HS2@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Wed, 3 Sep 1997 08:27:24 EST Received: from mail2.panix.com ([198.7.0.33]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Wed, 03 Sep 1997 12:27:15 +0000 (UT) Received: from [134.74.144.6] ([134.74.144.6]) by mail2.panix.com (8.8.5/8.7.1/PanixM1.0) with SMTP id IAA16365; Wed, 03 Sep 1997 08:26:42 -0400 (EDT) Date: Wed, 03 Sep 1997 08:26:16 -0400 From: Mike Vulis Subject: Re: Roman Numerals X-Sender: mv@panix.com To: slavutin@eng.tau.ac.il Cc: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <2.2.32.19970903122616.008d15a8@panix.com> MIME-version: 1.0 X-Mailer: Windows Eudora Pro Version 2.2 (32) Content-type: text/plain; charset="us-ascii" I believe that in Roman numerals, one can "subtract" a digit only from a digit of the next higher order. Thus, mxm is incorrect (subtracting 10 from 1000 is jumping 2 orders). Granted, your idea would lead to more compact and readable notation, but it would not be "Roman Numerals" any longer. At 09:11 AM 9/3/97 +0300, you wrote: > > Dear TeX community! > > The current algorithm of converting the numbers into roman ones > that is used in TeX and is described in TEX.WEB (\paragraph 69) > gives the following results: > for gives instead of > 1990 mcmxc mxm > 1995 mcmxcv mvm > 1999 mcmxcix mim > Really, I don't know what is right, but I want to propose another > algorithm > that gives the results on the right column of the above table. It is > written in > PASCAL and the program called ROMAN.PAS is attached. > > Sincerely yours, > Michael Slavutin, > slavutin@eng.tau.ac.il >Content-Type: TEXT/PLAIN; charset=US-ASCII; name="roman.pas" >Content-ID: >Content-Description: > >{THE FOLLOWING PROGRAM COMPUTES THE ROMAN NUMERAL OF A GIVEN >POSITIVE INTEGER} > >PROGRAM ROMAN; > {FIRST OF ALL IT DECLARES THE BASIC SYMBOLS OF ROMAN COUNTING AS "symbol" > ARRAY OF CHARACTERS AND THEIR NUMERICAL EQUIVALENTS AS "constant" > ARRAY OF INTEGERS} > CONST > constant : ARRAY[1..7]OF INTEGER = (1,5,10,50,100,500,1000); > symbol : ARRAY[1..7]OF CHAR = 'ivxlcdm'; > {THE VARIABLES IN USE ARE: > n IS THE INTEGER NUMBER TO BE CONVERTED > k IS THE INDEX INTO THE ROMAN ARRAYS > j IS ANOTHER INDEX INTO THE ROMAN ARRAYS} > VAR n:integer; > k:integer; > j:integer; > LABEL 10, 20; >BEGIN > {READ n AND START BUILDING THE ROMAN EQUIVALENT FROM THE BEGINNING > (1000)} > Read(n); > k:=7; > {WE USE THE FACT THAT THE ROMAN ARRAY IS BUILD FROM THE NUMBERS THAT > ARE OF THE KIND 10^L AND 5*10^L WHERE L IS A NON-NEGATIVE INTEGER. > THE ODD ENTRIES ARE OF THE TYPE 10^L, THUS THEY CAN BE REPEATED IN > THE ROMAN NUMERAL} > 10:WHILE ((n>=constant[k]) AND Odd(k)) DO > BEGIN > Write(symbol[k]); > n:=n-constant[k]; > END; > {NEXT WE HAVE TO CHECK IF n IS NOT EXHAUSTED ALREADY (OR IF IT WAS > NEGATIVE IN THE FIRST TIME. IF IT IS, THE COMPUTATION IS TERMINATED} > IF n<=0 > THEN GOTO 20; > {THEN WE CHECK IF THE REMAINDER OF n FITS INTO ONE OF THE INTERVALS > OF TYPE (constant[k]-constant[j])..constant[k] (FOR EXAMPLE, IF > n=921, THE FIRCT THING TO DISPLAY IS "cm". AFTER EVERY SUCCESFULL > CHECK n IS UPDATED AND SO IS k ACCORDING TO THE REMAINDER OF n} > FOR j:=1 TO k-1 DO > BEGIN > IF (n>=(constant[k]-constant[j])) > THEN IF (j<>k-1) > {THE CASES SUCH AS n=920 ARE TREATED HERE SINCE THIS IS > "cm..."} > THEN BEGIN > Write(symbol[j]); > Write(symbol[k]); > n:=n+constant[j]-constant[k]; > k:=j-1; > GOTO 10; > END > ELSE IF Odd(k) > {AND THE CASES SUCH AS n=520 HERE SINCE THIS IS > "d..."} > THEN BEGIN > Write(symbol[j]); > n:=n-constant[j]; > k:=j-1; > GOTO 10; > END > {HOWEVER IF n IS SOMETHING LIKE 481, THE SCRIPT > "symbol[j]""symbol[k]" IS PERFECTLY RIGHT SINCE > THIS IS "cd..."} > ELSE BEGIN > Write(symbol[j]); > Write(symbol[k]); > n:=n+constant[j]-constant[k]; > k:=j-1; > GOTO 10; > END; > END; > k:=k-1; > GOTO 10; > 20:Writeln(''); >EN ============= Michael Vulis ============= MicroPress, Inc. 68-30 Harrow Street, Forest Hills NY 11375 USA Email: support@micropress-inc.com, sales@micropress-inc.com WEB: http://www.micropress-inc.com TEL: 718 575 1816 FAX: 718 575 8038 ============= 4-Sep-1997 18:46:48-GMT,6488;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id MAA26500 for ; Thu, 4 Sep 1997 12:46:42 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 4 Sep 1997 18:46:41 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IN8XPP0EW0000NBV@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Thu, 4 Sep 1997 14:26:18 EST Received: from clavin.cs.tamu.edu ([128.194.130.106]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Thu, 04 Sep 1997 18:26:08 +0000 (UT) Received: from dilbert.cs.tamu.edu (205@dilbert [128.194.133.100]) by cs.tamu.edu (8.8.7/8.8.7) with ESMTP id NAA29016; Thu, 04 Sep 1997 13:24:42 -0500 (CDT) Received: (from bart@localhost) by dilbert.cs.tamu.edu (8.8.7/8.8.7) id NAA08533; Thu, 04 Sep 1997 13:23:49 -0500 (CDT) Date: Thu, 04 Sep 1997 13:23:49 -0500 (CDT) From: Bart Childs Subject: Re: Roman Numerals To: tex-implementors@MATH.AMS.ORG Cc: support@micropress-inc.com Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199709041823.NAA08533@dilbert.cs.tamu.edu> Don taught a special topics class at Stanford some years ago about the TeX system from a somewhat Software Engineering viewpoint. In one of the classes this topic was discussed at some length. As I remember the result was that Don was aware of several different possibilities and he chose the one that is reflected by the code. It is not totally unlike `center in the US' and `centre in the UK.' I am sure there are hundreds of better comparisons. I think the attitude of most for a possible change would be either no! or why? Regards, Bart Childs > From support@micropress-inc.com Wed Sep 3 07:38:08 1997 Date: Wed, 03 Sep 1997 08:26:16 -0400 From: Mike Vulis Subject: Re: Roman Numerals X-Sender: mv@panix.com To: slavutin@eng.tau.ac.il Cc: tex-implementors@math.ams.org MIME-version: 1.0 I believe that in Roman numerals, one can "subtract" a digit only from a digit of the next higher order. Thus, mxm is incorrect (subtracting 10 from 1000 is jumping 2 orders). Granted, your idea would lead to more compact and readable notation, but it would not be "Roman Numerals" any longer. At 09:11 AM 9/3/97 +0300, you wrote: > > Dear TeX community! > > The current algorithm of converting the numbers into roman ones > that is used in TeX and is described in TEX.WEB (\paragraph 69) > gives the following results: > for gives instead of > 1990 mcmxc mxm > 1995 mcmxcv mvm > 1999 mcmxcix mim > Really, I don't know what is right, but I want to propose another > algorithm > that gives the results on the right column of the above table. It is > written in > PASCAL and the program called ROMAN.PAS is attached. > > Sincerely yours, > Michael Slavutin, > slavutin@eng.tau.ac.il >Content-Type: TEXT/PLAIN; charset=US-ASCII; name="roman.pas" >Content-ID: >Content-Description: > >{THE FOLLOWING PROGRAM COMPUTES THE ROMAN NUMERAL OF A GIVEN >POSITIVE INTEGER} > >PROGRAM ROMAN; > {FIRST OF ALL IT DECLARES THE BASIC SYMBOLS OF ROMAN COUNTING AS "symbol" > ARRAY OF CHARACTERS AND THEIR NUMERICAL EQUIVALENTS AS "constant" > ARRAY OF INTEGERS} > CONST > constant : ARRAY[1..7]OF INTEGER = (1,5,10,50,100,500,1000); > symbol : ARRAY[1..7]OF CHAR = 'ivxlcdm'; > {THE VARIABLES IN USE ARE: > n IS THE INTEGER NUMBER TO BE CONVERTED > k IS THE INDEX INTO THE ROMAN ARRAYS > j IS ANOTHER INDEX INTO THE ROMAN ARRAYS} > VAR n:integer; > k:integer; > j:integer; > LABEL 10, 20; >BEGIN > {READ n AND START BUILDING THE ROMAN EQUIVALENT FROM THE BEGINNING > (1000)} > Read(n); > k:=7; > {WE USE THE FACT THAT THE ROMAN ARRAY IS BUILD FROM THE NUMBERS THAT > ARE OF THE KIND 10^L AND 5*10^L WHERE L IS A NON-NEGATIVE INTEGER. > THE ODD ENTRIES ARE OF THE TYPE 10^L, THUS THEY CAN BE REPEATED IN > THE ROMAN NUMERAL} > 10:WHILE ((n>=constant[k]) AND Odd(k)) DO > BEGIN > Write(symbol[k]); > n:=n-constant[k]; > END; > {NEXT WE HAVE TO CHECK IF n IS NOT EXHAUSTED ALREADY (OR IF IT WAS > NEGATIVE IN THE FIRST TIME. IF IT IS, THE COMPUTATION IS TERMINATED} > IF n<=0 > THEN GOTO 20; > {THEN WE CHECK IF THE REMAINDER OF n FITS INTO ONE OF THE INTERVALS > OF TYPE (constant[k]-constant[j])..constant[k] (FOR EXAMPLE, IF > n=921, THE FIRCT THING TO DISPLAY IS "cm". AFTER EVERY SUCCESFULL > CHECK n IS UPDATED AND SO IS k ACCORDING TO THE REMAINDER OF n} > FOR j:=1 TO k-1 DO > BEGIN > IF (n>=(constant[k]-constant[j])) > THEN IF (j<>k-1) > {THE CASES SUCH AS n=920 ARE TREATED HERE SINCE THIS IS > "cm..."} > THEN BEGIN > Write(symbol[j]); > Write(symbol[k]); > n:=n+constant[j]-constant[k]; > k:=j-1; > GOTO 10; > END > ELSE IF Odd(k) > {AND THE CASES SUCH AS n=520 HERE SINCE THIS IS > "d..."} > THEN BEGIN > Write(symbol[j]); > n:=n-constant[j]; > k:=j-1; > GOTO 10; > END > {HOWEVER IF n IS SOMETHING LIKE 481, THE SCRIPT > "symbol[j]""symbol[k]" IS PERFECTLY RIGHT SINCE > THIS IS "cd..."} > ELSE BEGIN > Write(symbol[j]); > Write(symbol[k]); > n:=n+constant[j]-constant[k]; > k:=j-1; > GOTO 10; > END; > END; > k:=k-1; > GOTO 10; > 20:Writeln(''); >EN ============= Michael Vulis ============= MicroPress, Inc. 68-30 Harrow Street, Forest Hills NY 11375 USA Email: support@micropress-inc.com, sales@micropress-inc.com WEB: http://www.micropress-inc.com TEL: 718 575 1816 FAX: 718 575 8038 ============= 12-Sep-1997 7:06:31-GMT,5918;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id BAA28060 for ; Fri, 12 Sep 1997 01:06:29 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 12 Sep 1997 07:06:27 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01INJFHBZIM80003L5@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Fri, 12 Sep 1997 02:42:56 EST Received: from newton.feld.cvut.cz ([147.32.244.10]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Fri, 12 Sep 1997 06:42:42 +0000 (UT) Received: from localhost (olsak@localhost) by newton.feld.cvut.cz (8.8.5/8.8.5) with SMTP id IAA06805 for ; Fri, 12 Sep 1997 08:41:20 +0200 Date: Fri, 12 Sep 1997 08:41:20 +0200 (MET DST) From: Petr Olsak Subject: The encTeX extension of TeX X-Sender: olsak@newton.feld.cvut.cz To: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII X-Authentication-warning: newton.feld.cvut.cz: olsak owned process doing -bs Dear TeX implementors, I deduce from the name of this e-mail list that the implementors of TeX for various operating systems from TeX WEB source are subscribed to this e-mail address. I don't know who is subscribed here explicit but I suppose that there is the right address for my offer here. I have made a little extension of TeX. All changes are done via WEB tex.ch change file, thus they are system independent. The new primitives are introduced. They give the possibility to read and change the xord/xchr reencoding vectors in TeX input preprocessor and to set the "printability" of characters for text output to log, terminal and to \write files. The "printability" means that the character outputs by its xchr[internal] code and not by ^^notation. This problem is solved by TCP tables in emTeX, for example. This solution is system dependent, but my solution is system independent. The encoding tables for xchr/xord vectors are inputted via \input and via appropriate macros during iniTeX in my extension. The settings are stored in the format file. The access to xchr/xord vectors are unconditionally needed in TeX implementation in our country. The Czech language has its alphabet with accented letters which are encoded in eight different encoding "standards" in computers. Four of this encodings are used commonly on DOS and PCs. I think, the similar problems are in another languages too. There were some solution for Czech language so far: The emTeX is widely used in DOS systems. Thus the TCP tables are used here. The web2c TeX is used in UNIX like systems. The ISO8859-2 encoding is defined for our language on these systems. I managed so called "CS-fonts" in Czech TeX implementation (named CSTeX) with a special encodnig (different from T1). This encoding matches with ISO8859-2 on our alphabet. The web2c TeX on UNIX is used with these fonts, thus no reencoding is needed. But this solution is not so universal. Now, the popularity of web2c TeX is adavanced because the 32 bit Operating systems go to the commonly purchased PC's. The problem arises: These systems do not use the ISO8859-2 for our language. The creation of the new font encoding for TeX fonts is bad idea. The TeX fonts are implemented independently on the operating system and differencies between operating systems and TeX font encoding must be solved via xchr/xord. It was the primary Don Knuth's idea, when EBDIC or ASCII seven bit encoding were used and TeX CM fonts were in ASCII only. The TeX implementation, where only identity mapping on codes > 128 are possible, is not usable for our language. The Skarvada's patch for web2c does some reencoding, but the xord/xchr setting is dependent on an environment variable setting and is independent on the format file. I mean, that it is the fragile solution and I prefer, the xord/xchr reencoding is stored in format file. You can look into my package (named encTeX) which introduces three new primitives to xchr/xord/printability setting. This package includes the patch file simply applicable to tex.ch in web2c 7.0 TeX implementation. The changes are independent on web2c and they can be implemented into various others tex.ch files manually. The README.eng file includes the short introduction to the new primitives and my arguments, why the xord/xchr accessing is very important in the TeX program. The pacakge is accesible via anonymous ftp on ftp://math.feld.cvut.cz/pub/olsak/enctex/ The TRIP test reports only two differencies: 1. The banner is changed, 2. The number of multiletter control sequences are geater by three. If you take my work as interesting for your implementation of TeX, please, contact me. We can discuss the details about implementation, documentation, license and so on. The compromisses are possible. For instance, the encTeX extension will be accessible only if the command line option -enc is specified on iniTeX state and this option is stored in the format file. This feauture is not implemented in my package bacause the command line parser is system dependent. If you include the encTeX into your TeX implementation, you can add this feature. It will be very good news for hunderts Czech users of TeX (and for users from other countries may by too), if the xord/xchr accessing will be implemented into widely used web2c TeX, MikTeX and others packages. The second package seems very interesting for Microsoft 32 bit platforms, but without reencoding possibility it is not usable for these users. Petr Olsak Departament of Mathematics, FEL CVUT Prague, Czech Republics http://math.feld.cvut.cz/olsak 16-Sep-1997 2:15:17-GMT,1646;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id UAA07227 for ; Mon, 15 Sep 1997 20:15:15 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 16 Sep 1997 02:15:14 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01INOQL16AZK000PK1@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Mon, 15 Sep 1997 21:54:22 EST Received: from zothommog.evcom.net ([208.136.203.8]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Tue, 16 Sep 1997 01:54:13 +0000 (UT) Received: (from kinch@localhost) by zothommog.evcom.net (8.8.5/8.6.12) id VAA14698 for tex-implementors@math.ams.org; Mon, 15 Sep 1997 21:54:32 -0400 Date: Mon, 15 Sep 1997 21:54:31 -0400 (EDT) From: "Richard J. Kinch" Subject: Meaningless conditional in tex.web? To: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199709160154.VAA14698@zothommog.evcom.net> Content-type: text The following conditional appears in tex.web, sections 174 and 176: if (font(p); Thu, 18 Sep 1997 09:43:36 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 18 Sep 1997 15:43:30 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01INSBFILJW000138H@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Thu, 18 Sep 1997 11:24:01 EST Received: from fsnif.neuroinformatik.ruhr-uni-bochum.de ([134.147.176.16]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Thu, 18 Sep 1997 15:23:50 +0000 (UT) Received: from maloche.neuroinformatik.ruhr-uni-bochum.de (dak@maloche.neuroinformatik.ruhr-uni-bochum.de [134.147.176.143]) by fsnif.neuroinformatik.ruhr-uni-bochum.de (8.8.7/8.8.7) with SMTP id RAA23200; Thu, 18 Sep 1997 17:23:48 +0200 (MET DST) Received: by maloche.neuroinformatik.ruhr-uni-bochum.de (SMI-8.6/SMI-SVR4) id RAA09518; Thu, 18 Sep 1997 17:23:43 +0200 Date: Thu, 18 Sep 1997 17:23:43 +0200 From: David Kastrup Subject: Re: Meaningless conditional in tex.web? In-reply-to: <199709160154.VAA14698@zothommog.evcom.net> (kinch@holonet.net) To: kinch@holonet.net Cc: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199709181523.RAA09518@maloche.neuroinformatik.ruhr-uni-bochum.de> MIME-version: 1.0 (generated by tm-edit 7.93) Content-type: text/plain; charset=ISO-8859-1 X-MIME-Autoconverted: from 8bit to quoted-printable by fsnif.neuroinformatik.ruhr-uni-bochum.de id RAA23200 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by csc-sun.math.utah.edu id JAA20155 Date: Mon, 15 Sep 1997 21:54:31 -0400 (EDT) From: "Richard J. Kinch" The following conditional appears in tex.web, sections 174 and 176: if (font(p) Received: (from beebe@localhost) by plot79.math.utah.edu (8.8.5/8.8.5) id JAA02087; Thu, 18 Sep 1997 09:54:06 -0600 (MDT) Date: Thu, 18 Sep 1997 09:54:06 -0600 (MDT) To: David Kastrup Cc: beebe@math.utah.edu, kinch@holonet.net, tex-implementors@MATH.AMS.ORG X-US-Mail: "Center for Scientific Computing, University of Utah, 155 S 1400 E RM 233, Salt Lake City, UT 84112-0090, USA" X-Telephone: +1 801 581 5254 X-FAX: +1 801 581 4148 X-URL: http://www.math.utah.edu/~beebe Subject: Re: Meaningless conditional in tex.web? In-Reply-To: Your message of Thu, 18 Sep 1997 17:23:43 +0200 Message-ID: >> David Kastrup writes: >> The definition of quarterword is a system dependency (if I'm not >> mistaken), and could well be -128..127 on systems other than the >> canonical Pascal/H system for which tex.web is written. That is correct; see section 110, p. 47, of ``TeX: The Program'', where Don explicitly allowed for both signed and unsigned ranges, via the constants min_quarterword and max_quarterword. The former has 44 references in the index, and the latter, 11. ======================================================================== Nelson H. F. Beebe Tel: +1 801 581 5254 Center for Scientific Computing FAX: +1 801 581 4148 University of Utah Internet e-mail: beebe@math.utah.edu Department of Mathematics, 105 JWB URL: http://www.math.utah.edu/~beebe 155 S 1400 E RM 233 Salt Lake City, UT 84112-0090, USA ======================================================================== 18-Sep-1997 17:10:05-GMT,1760;000000000001 Received: from baygate.bayarea.net (root@baygate.bayarea.net [204.71.212.2]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with ESMTP id LAA22359 for ; Thu, 18 Sep 1997 11:10:00 -0600 (MDT) Received: from shell2.bayarea.net (shell2.bayarea.net [205.219.84.7]) by baygate.bayarea.net (8.8.5/8.8.5) with ESMTP id KAA12132; Thu, 18 Sep 1997 10:12:31 -0700 (PDT) Received: (from quixote@localhost) by shell2.bayarea.net (8.8.5/8.8.5) id KAA11138; Thu, 18 Sep 1997 10:10:55 -0700 (PDT) From: Quixote Digital Typography Message-Id: <199709181710.KAA11138@shell2.bayarea.net> Subject: Re: Meaningless conditional in tex.web? To: beebe@math.utah.edu (Nelson H. F. Beebe) Date: Thu, 18 Sep 1997 10:10:54 -0700 (PDT) Cc: tex-implementors@math.ams.org In-Reply-To: from "Nelson H. F. Beebe" at Sep 18, 97 09:54:06 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit > >> David Kastrup writes: > >> The definition of quarterword is a system dependency (if I'm not > >> mistaken), and could well be -128..127 on systems other than the > >> canonical Pascal/H system for which tex.web is written. > That is correct; see section 110, p. 47, of ``TeX: The Program'', > where Don explicitly allowed for both signed and unsigned > ranges, via the constants min_quarterword and max_quarterword. > The former has 44 references in the index, and the latter, 11. Yes, but the constant in question must fall within the quarterword rang and is one of the consistency checks that occurs elsewhere in tex.web I was unable to figure out why that test was there. -dh 18-Sep-1997 17:25:15-GMT,2212;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id LAA22752 for ; Thu, 18 Sep 1997 11:25:14 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 18 Sep 1997 17:25:14 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01INSF4N2WI8000YKC@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Thu, 18 Sep 1997 13:09:49 EST Received: from baygate.bayarea.net ([204.71.212.2]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Thu, 18 Sep 1997 17:09:36 +0000 (UT) Received: from shell2.bayarea.net (shell2.bayarea.net [205.219.84.7]) by baygate.bayarea.net (8.8.5/8.8.5) with ESMTP id KAA12132; Thu, 18 Sep 1997 10:12:31 -0700 (PDT) Received: (from quixote@localhost) by shell2.bayarea.net (8.8.5/8.8.5) id KAA11138; Thu, 18 Sep 1997 10:10:55 -0700 (PDT) Date: Thu, 18 Sep 1997 10:10:54 -0700 (PDT) From: Quixote Digital Typography Subject: Re: Meaningless conditional in tex.web? In-reply-to: To: beebe@math.utah.edu (Nelson H. F. Beebe) Cc: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199709181710.KAA11138@shell2.bayarea.net> MIME-version: 1.0 X-Mailer: ELM [version 2.4 PL24] Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7bit > >> David Kastrup writes: > >> The definition of quarterword is a system dependency (if I'm not > >> mistaken), and could well be -128..127 on systems other than the > >> canonical Pascal/H system for which tex.web is written. > That is correct; see section 110, p. 47, of ``TeX: The Program'', > where Don explicitly allowed for both signed and unsigned > ranges, via the constants min_quarterword and max_quarterword. > The former has 44 references in the index, and the latter, 11. Yes, but the constant in question must fall within the quarterword rang and is one of the consistency checks that occurs elsewhere in tex.web I was unable to figure out why that test was there. -dh 19-Sep-1997 8:32:09-GMT,1786;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id CAA10643 for ; Fri, 19 Sep 1997 02:32:07 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 19 Sep 1997 08:32:06 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01INTAK8JA4W000XNR@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Fri, 19 Sep 1997 04:09:34 EST Received: from hermes.ucd.ie ([137.43.1.49]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Fri, 19 Sep 1997 08:09:24 +0000 (UT) Received: from maths.ucd.ie by hermes.ucd.ie (PMDF V5.1-8 #16181) with SMTP id <0EGQ00G9IXBLQR@hermes.ucd.ie> for tex-implementors@MATH.AMS.ORG; Fri, 19 Sep 1997 09:09:21 +0100 (BST) Received: by maths.ucd.ie (16.8/0.0) id AA05344; Fri, 19 Sep 1997 09:03:15 +0100 Date: Fri, 19 Sep 1997 09:03:14 +0100 (BST) From: wgs@maths.ucd.ie (Wayne G. Sullivan) Subject: Re: Meaningless conditional in tex.web? To: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <9709190803.AA05344@maths.ucd.ie> Mailer: Elm [revision: 70.30] The two procedures, short_display and print_font_and_char, in which the check "(font(p)font_max)" belong to TeX's diagnostic routines. If some fundamental error has arisen, e.g., a format file has become corrupted, then font(p) might be beyond the range in the equivalence table which corresponds to fonts. Note that the test "p>mem_end" in print_font_and_char should never be satisfied when TeX is running correctly. Incidently, font_max in texk7.0/web2c/tex.ch is 2000, and the checks for font(p); Sun, 21 Sep 1997 01:39:50 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 21 Sep 1997 07:39:49 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01INW0S83GI8001HF5@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Sun, 21 Sep 1997 03:02:08 EST Received: from shum.cc.huji.ac.il ([132.64.1.3]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Sun, 21 Sep 1997 07:01:54 +0000 (UT) Received: from localhost (rama@localhost) by shum.cc.huji.ac.il (8.8.6/8.6.10) with SMTP id JAA05333 for ; Sun, 21 Sep 1997 09:01:51 +0200 (GMT+0200) Date: Sun, 21 Sep 1997 09:01:51 +0200 (GMT+0200) From: rama porrat Subject: Right-to-left TeX and LaTeX. In-reply-to: <199709181523.RAA09518@maloche.neuroinformatik.ruhr-uni-bochum.de> To: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII X-Authentication-warning: shum.cc.huji.ac.il: rama owned process doing -bs Dear TeX implementor! Please allow me once again to draw you attention to right-to-left tex's (tex--xet's) poor condition, and ask for your advice re latex. TeX for unix: We work with an edited version of tex.web, having an accompnying suitable change file. It has been prepared by Uri Feldman of Israel. It works ok, only can't be upgraded to the next tex's version. I understand that a beta version of teTeX, which includes etex (and hence, tex--xet) is available. This will, I hope, replace our non official version. TeX for PC: There exists Peter Breitenlohner's tex--xet, included in etex. It works well, but it is a 16 bit program. So even in DOS we have difficulties when the input tex file is not small enough. LaTeX for all systems: We don't have a hebrew.sty for latex2e - we still work with latex209. A long time ago I have started to write a hebrew.sty suitable for latex2e, but I have stopped when I realized that this means editing numerous *cls and *sty files, editing which will have to be redone twice a year. Also, I do not have the knowledge of latex which is required in order to do that thing well. However, again I realize it has to be done now for the sake of latex users which need documents in right-to-left typesetting, and I see that an official right-to-left latex will not be available in the near future. I assume that if I am going to do all that editing, I have to look for all occurences or ...right.../..left... and supply the ...left.../...right.... equivalence. Is that right? Will that be sufficient? Can you give me some guidance in this work? I have already created a hebrew.sty with Babel's parameters (it might have to be reedited, as Babel had probably changed since I have done it, about 2 years ago), but this is just the beginning. Thanks! Rama. rama@cc.huji.ac.il The Hebrew University of Jerusalem. 22-Sep-1997 13:01:03-GMT,2076;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id HAA00365 for ; Mon, 22 Sep 1997 07:01:02 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 22 Sep 1997 13:01:01 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01INXQU91K6O001BSZ@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Mon, 22 Sep 1997 08:38:58 EST Received: from pillar.elsevier.co.uk ([193.131.222.35]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Mon, 22 Sep 1997 12:38:47 +0000 (UT) Received: from snowdon.elsevier.co.uk (snowdon.elsevier.co.uk [193.131.197.164]) by pillar.elsevier.co.uk (8.8.5/8.8.5) with ESMTP id NAA00174 for ; Mon, 22 Sep 1997 13:38:10 +0100 (BST) Received: from SRAHTZ (actually host srahtz.elsevier.co.uk) by snowdon.elsevier.co.uk with SMTP (PP); Mon, 22 Sep 1997 13:38:06 +0100 Date: Mon, 22 Sep 1997 13:24:22 +0100 From: s.rahtz@elsevier.co.uk (Sebastian Rahtz) Subject: Re: Right-to-left TeX and LaTeX. In-reply-to: To: rama@cc.huji.ac.il Cc: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <9249-Mon22Sep1997132422+0100-s.rahtz@elsevier.co.uk> MIME-version: 1.0 X-Mailer: VM 6.33 under Emacs 19.34.4 Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit References: <199709181523.RAA09518@maloche.neuroinformatik.ruhr-uni-bochum.de> I dont understand the comment about the `poor conditon' of tex--xet. Do you mean the lack of TeX implementations that include it? or that its internally inadequate? as for LaTeX, thats a matter to take up with the LaTeX people, not tex implementors, isn't it? obviously, i challenge the assumption that you have to update your .cls file twice a year - it simply isnt true! sebastian 24-Sep-1997 14:01:11-GMT,2899;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id DAA29341 for ; Wed, 24 Sep 1997 03:02:01 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 24 Sep 1997 09:02:00 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IO0ALLET1S001U6R@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Wed, 24 Sep 1997 04:26:17 EST Received: from shum.cc.huji.ac.il ([132.64.1.3]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Wed, 24 Sep 1997 08:26:04 +0000 (UT) Received: from localhost (rama@localhost) by shum.cc.huji.ac.il (8.8.6/8.6.10) with SMTP id KAA24416; Wed, 24 Sep 1997 10:24:41 +0200 (GMT+0200) Date: Wed, 24 Sep 1997 10:24:41 +0200 (GMT+0200) From: rama porrat Subject: Re: Right-to-left TeX and LaTeX. In-reply-to: <9249-Mon22Sep1997132422+0100-s.rahtz@elsevier.co.uk> To: Sebastian Rahtz Cc: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Reply-to: rama porrat Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII X-Authentication-warning: shum.cc.huji.ac.il: rama owned process doing -bs > I dont understand the comment about the `poor conditon' of tex--xet. > Do you mean the lack of TeX implementations that include it? or that > its internally inadequate? I'm sorry I used the wrong words. The program's condition is ok, and certainly internally inadequate. It is the user's condition which is not ok. If he works with tex--xet on his pc, he cannot include Gnuplot's graphs, because there will probably not be enough memory for more then one or two graphs. (As I have mentioned in my previous mail, I hope e-tex for unix will replace the current frozen tex--xet version used.) If the user uses Hebrew latex with tex--xet, he cannot write an article without having to ask for correction of numerous small problems. And he cannot use latex2e. This, I think, means that a user is not ok. The user does not have a 100% reliable program if he wants to write a Hebrew latex document. > as for LaTeX, thats a matter to take up with the LaTeX people, not tex > implementors, isn't it? Yes. I thought all implementors, including the persons in the latex3 project, are subscribed to this list. > obviously, i challenge the assumption that you > have to update your .cls file twice a year - it simply isnt true! I guess you mean that not many changes are maded every half year. But nevertheless, if one wants to be careful, one has to go over all the files twice a year and see if things have changed somewhere. Otherwise, why is there a new version of latex2e twice a year? Bye -- Rama. 24-Sep-1997 14:01:20-GMT,2989;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id DAA01999 for ; Wed, 24 Sep 1997 03:43:31 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 24 Sep 1997 09:43:31 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IO0C3QUV9S001ZXI@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Wed, 24 Sep 1997 05:09:05 EST Received: from pillar.elsevier.co.uk ([193.131.222.35]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Wed, 24 Sep 1997 09:08:56 +0000 (UT) Received: from snowdon.elsevier.co.uk (snowdon.elsevier.co.uk [193.131.197.164]) by pillar.elsevier.co.uk (8.8.5/8.8.5) with ESMTP id KAA14986 for ; Wed, 24 Sep 1997 10:08:05 +0100 (BST) Received: from SRAHTZ (actually host srahtz.elsevier.co.uk) by snowdon.elsevier.co.uk with SMTP (PP); Wed, 24 Sep 1997 10:08:06 +0100 Date: Wed, 24 Sep 1997 10:04:40 +0100 From: s.rahtz@elsevier.co.uk (Sebastian Rahtz) Subject: Re: Right-to-left TeX and LaTeX. In-reply-to: To: rama@cc.huji.ac.il Cc: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <8157-Wed24Sep1997100440+0100-s.rahtz@elsevier.co.uk> MIME-version: 1.0 X-Mailer: VM 6.33 under Emacs 19.34.4 Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit References: <9249-Mon22Sep1997132422+0100-s.rahtz@elsevier.co.uk> rama porrat writes: > If he works with tex--xet on his pc, he cannot include Gnuplot's > graphs, because there will probably not be enough memory for more > then one or two graphs. good gracious, you are descending to a *very* specific level here. How ever do I know how you include gnuplot graphs??????? thats your problem..... > The user does not have a 100% reliable program if he wants > to write a Hebrew latex document. it sounds like you want to use eTeX, and build a LaTeX 209 format file until someone fixes hebrew.sty for 2e. > Yes. I thought all implementors, including the persons in the latex3 > project, are subscribed to this list. no, just people who implement TeX or are interested in the subject. perhaps the word `implement' is confusing - this list is simply those people who have the problem of compiling a running TeX from tex.web > But nevertheless, if one wants to be careful, one has to go over all > the files twice a year and see if things have changed somewhere. > Otherwise, why is there a new version of latex2e twice a year? if you choose to use undocumented LaTeX internals in your style file, yes, you will have to check it twice a year. Moral - dont use undocumented internals... The LaTeX people dont change the meaning of the `official' commands Sebastian 24-Sep-1997 17:03:25-GMT,4375;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id LAA05721 for ; Wed, 24 Sep 1997 11:03:23 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 24 Sep 1997 17:03:19 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IO0KJNAEMO001ZD8@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Wed, 24 Sep 1997 09:11:02 EST Received: from fsnif.neuroinformatik.ruhr-uni-bochum.de ([134.147.176.16]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Wed, 24 Sep 1997 13:10:48 +0000 (UT) Received: from maloche.neuroinformatik.ruhr-uni-bochum.de (dak@maloche.neuroinformatik.ruhr-uni-bochum.de [134.147.176.143]) by fsnif.neuroinformatik.ruhr-uni-bochum.de (8.8.7/8.8.7) with SMTP id PAA26745; Wed, 24 Sep 1997 15:08:53 +0200 (MET DST) Received: by maloche.neuroinformatik.ruhr-uni-bochum.de (SMI-8.6/SMI-SVR4) id PAA03700; Wed, 24 Sep 1997 15:08:48 +0200 Date: Wed, 24 Sep 1997 15:08:48 +0200 From: David Kastrup Subject: Re: Right-to-left TeX and LaTeX. In-reply-to: (message from rama porrat on Wed, 24 Sep 1997 10:24:41 +0200 (GMT+0200)) To: rama@cc.huji.ac.il Cc: s.rahtz@elsevier.co.uk, tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <199709241308.PAA03700@maloche.neuroinformatik.ruhr-uni-bochum.de> MIME-version: 1.0 (generated by tm-edit 7.93) Content-type: text/plain; charset=ISO-8859-1 X-MIME-Autoconverted: from 8bit to quoted-printable by fsnif.neuroinformatik.ruhr-uni-bochum.de id PAA26745 Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by csc-sun.math.utah.edu id LAA05721 Date: Wed, 24 Sep 1997 10:24:41 +0200 (GMT+0200) From: rama porrat > I dont understand the comment about the `poor conditon' of tex--xet. > Do you mean the lack of TeX implementations that include it? or that > its internally inadequate? I'm sorry I used the wrong words. The program's condition is ok, and certainly internally inadequate. ^^^^^^^^^^ You sure that's what you wanted to write? It is the user's condition which is not ok. If he works with tex--xet on his pc, he cannot include Gnuplot's graphs, because there will probably not be enough memory for more then one or two graphs. (As I have mentioned in my previous mail, I hope e-tex for unix will replace the current frozen tex--xet version used.) I guess you misunderstand what TeX--XeT is: it is a set of changefiles used when compiling a version of TeX. This set includes absolutely no information about the internal sizes of the resulting binary. So it seems that you have a version of TeX using TeX--XeT which has been compiled with small internal sizes. That's not the problem of the TeX--XeT patches, but of the person who prepared your binaries. > as for LaTeX, thats a matter to take up with the LaTeX people, not tex > implementors, isn't it? Yes. I thought all implementors, including the persons in the latex3 project, are subscribed to this list. It might have helped checking the list's charter before posting to it. The TeX implementor list is only concerned with TeX binaries, mainly with their bugs. The LaTeX people have their own forums. > obviously, i challenge the assumption that you > have to update your .cls file twice a year - it simply isnt true! I guess you mean that not many changes are maded every half year. But nevertheless, if one wants to be careful, one has to go over all the files twice a year and see if things have changed somewhere. Otherwise, why is there a new version of latex2e twice a year? Internals change and bugs are removed. The interfaces should mostly remain. What you might be thinking of is doing changes immediately in the LaTeX classes, which is neither prudent nor allowed according to the licence. David Kastrup Phone: +49-234-700-5570 Email: dak@neuroinformatik.ruhr-uni-bochum.de Fax: +49-234-709-4209 Institut für Neuroinformatik, Universitätsstr. 150, 44780 Bochum, Germany 24-Sep-1997 19:27:47-GMT,2707;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id NAA09253 for ; Wed, 24 Sep 1997 13:27:45 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 24 Sep 1997 19:27:40 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IO0X78V2680020A5@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Wed, 24 Sep 1997 15:13:35 EST Received: from alpha1.rhbnc.ac.uk ([134.219.201.113]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Wed, 24 Sep 1997 19:13:26 +0000 (UT) Date: Wed, 24 Sep 1997 20:13:25 +0100 From: Philip Taylor (RHBNC) Subject: TeX--XeT To: TEX-IMPLEMENTORS@MATH.AMS.ORG Cc: CHAA006@vms.rhbnc.ac.uk Errors-to: tex-implementors-request@MATH.AMS.ORG Reply-to: P.Taylor@vms.rhbnc.ac.uk Message-id: <970924201325.10a60@vms.rhbnc.ac.uk> >> I guess you misunderstand what TeX--XeT is: it is a set of >> changefiles used when compiling a version of TeX. This set includes >> absolutely no information about the internal sizes of the resulting >> binary. So it seems that you have a version of TeX using TeX--XeT >> which has been compiled with small internal sizes. That's not the >> problem of the TeX--XeT patches, but of the person who prepared your >> binaries. But in Rama's case, the two are the same: Peter Breitenlohner; Peter is both the implementor of TeX--XeT and of PubliC (e-)TeX. Note that there is nothing _wrong_ with Peter's implementation of either: PubliC (e-)TeX is quite intentionally a Small Program which will run on a Small PC... Sebastian advises me that the Web2C/Win32 system _does_ include e-TeX, and if Rama can get that to run on his system, it should be quite large enough to deal with his documents (we hope!). >> It might have helped checking the list's charter before posting to >> it. The TeX implementor list is only concerned with TeX binaries, >> mainly with their bugs. The LaTeX people have their own forums. It might help if TeX-Implementors were not so dismissive of real users as the three lines above suggest: OK, perhaps this is the wrong forum in which to raise LaTeX-specific issues, but if we (TeX-Implementors) dismiss LaTeX queries in such a cavalier fashion, we can hardly complain when Rama and other TeX users with genuine problems decide to migrate to non-TeX-based systems on which they get more helpful responses. ** Phil (who wants to see TeX continue into the 21st Century, and who would therefore like to encourage an ethos in which users such as Rama get help rather than flak...). 25-Sep-1997 6:50:15-GMT,1781;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id AAA23319 for ; Thu, 25 Sep 1997 00:50:13 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 25 Sep 1997 06:50:12 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IO1KO7PZG0001ZFR@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Thu, 25 Sep 1997 02:25:32 EST Received: from shum.cc.huji.ac.il ([132.64.1.3]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Thu, 25 Sep 1997 06:25:17 +0000 (UT) Received: from localhost (rama@localhost) by shum.cc.huji.ac.il (8.8.6/8.6.10) with SMTP id IAA29243; Thu, 25 Sep 1997 08:23:50 +0200 (GMT+0200) Date: Thu, 25 Sep 1997 08:23:50 +0200 (GMT+0200) From: rama porrat Subject: Sorry .... In-reply-to: <199709241308.PAA03700@maloche.neuroinformatik.ruhr-uni-bochum.de> To: David Kastrup Cc: s.rahtz@elsevier.co.uk, tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII X-Authentication-warning: shum.cc.huji.ac.il: rama owned process doing -bs Sorry! Of course I meant > I'm sorry I used the wrong words. The program's condition is ok, > and certainly internally MOST adequate. ~~~~~~~~ Thanks to David Kastrup for pointing this to me, and I really think the latex3 group is doing a great job. All I am saying is, it would be nice if latex could be officially made to typeset from right-to-left. Rama. 25-Sep-1997 7:12:35-GMT,2392;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id BAA23714 for ; Thu, 25 Sep 1997 01:12:33 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 25 Sep 1997 07:12:32 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IO1LE8OYG0001L8U@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Thu, 25 Sep 1997 02:46:40 EST Received: from shum.cc.huji.ac.il ([132.64.1.3]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Thu, 25 Sep 1997 06:46:16 +0000 (UT) Received: from localhost (rama@localhost) by shum.cc.huji.ac.il (8.8.6/8.6.10) with SMTP id IAA29429; Thu, 25 Sep 1997 08:46:07 +0200 (GMT+0200) Date: Thu, 25 Sep 1997 08:46:07 +0200 (GMT+0200) From: rama porrat Subject: Re: Right-to-left TeX and LaTeX. In-reply-to: <199709241308.PAA03700@maloche.neuroinformatik.ruhr-uni-bochum.de> To: David Kastrup Cc: tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Reply-to: rama porrat Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII X-Authentication-warning: shum.cc.huji.ac.il: rama owned process doing -bs > What you might be thinking of is doing changes immediately in > the LaTeX classes, which is neither prudent nor allowed according to > the licence. If this is so, there is no chance of us having a possibility of writing Hebrew latex articles. Take a small example: In article.cls you have \renewcommand\thesubsection {\thesection.\@arabic\c@subsection} In a right-to-left document, the order should be reversed; that is, something similar to \renewcommand\thesubsection {\@arabic\c@subsection.\thesection} This is because, as the direction of typesetting is not part of the code, ALL the line is simply outputted backwards, so numbers come out reversed. Another example from article.cls -- \newenvironment{quote} {\list{}{\rightmargin\leftmargin}% ..... This should become {\list{}{\leftmargin\rightmargin}% ..... How is this to be solved without changing article.cls? And the class files are FULL of those things. Rama. 25-Sep-1997 10:01:11-GMT,3292;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id EAA26576 for ; Thu, 25 Sep 1997 04:01:09 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 25 Sep 1997 10:01:08 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IO1QYUA440001UZ8@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Thu, 25 Sep 1997 05:25:49 EST Received: from pillar.elsevier.co.uk ([193.131.222.35]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Thu, 25 Sep 1997 09:25:39 +0000 (UT) Received: from snowdon.elsevier.co.uk (snowdon.elsevier.co.uk [193.131.197.164]) by pillar.elsevier.co.uk (8.8.5/8.8.5) with ESMTP id KAA21619 for ; Thu, 25 Sep 1997 10:24:58 +0100 (BST) Received: from SRAHTZ (actually host srahtz.elsevier.co.uk) by snowdon.elsevier.co.uk with SMTP (PP); Thu, 25 Sep 1997 10:24:53 +0100 Date: Thu, 25 Sep 1997 09:40:00 +0100 From: s.rahtz@elsevier.co.uk (Sebastian Rahtz) Subject: Re: TeX--XeT In-reply-to: <970924201325.10a60@vms.rhbnc.ac.uk> To: P.Taylor@vms.rhbnc.ac.uk Cc: TEX-IMPLEMENTORS@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <3302-Thu25Sep1997094000+0100-s.rahtz@elsevier.co.uk> MIME-version: 1.0 X-Mailer: VM 6.33 under Emacs 19.34.4 Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit References: <970924201325.10a60@vms.rhbnc.ac.uk> Philip Taylor (RHBNC) writes: > Sebastian advises me that the Web2C/Win32 system _does_ include e-TeX, > and if Rama can get that to run on his system, it should be quite > large enough to deal with his documents (we hope!). users of web2c 7.0 can, of course, change memory allocation in the run-time texmf.cnf file, so Rama is going to have to work much harder to run out of memory > ** Phil (who wants to see TeX continue into the 21st Century, and who > would therefore like to encourage an ethos in which users such > as Rama get help rather than flak...). This is silly. If any query goes to any list, madness ensues. In this case, Rama was misled by the title `tex-implementors', I think, but usually the purpose of mailing lists is pretty clear, and its perfectly reasonable to expect people to stick to those purposes. Whether the words were harsh or not, I think in fact Rama got the required answers, and indeed the LaTeX issues raised about hebrew (and czech, privately) were passed to the LaTeX team. I fail to see, actually, the relationship between the longevity of TeX and the helpfulness or otherwise of tex-implementors list.... I would peraps suggest the longevity of TeX is determined by factors outside TeX. That is to say, it will survive until someone makes something better. Possibly that something better will come from inside the TeX world, but lets be honest, there are not many signs of it. I discount things like e-TeX, which are just tidying up the edges, and probably Omega too unless it has a fairly radical change of heart. We'll have to see what NTS produces, won't we? Lets hope something more than the stated aim of `reimplementing TeX in Lisp'.... Sebastian 25-Sep-1997 10:10:39-GMT,2634;000000000001 Received: from math.ams.org (math.ams.org [130.44.210.14]) by csc-sun.math.utah.edu (8.8.5/8.8.5) with SMTP id EAA26748 for ; Thu, 25 Sep 1997 04:10:38 -0600 (MDT) Received: from axp14.ams.org by math.ams.org via smtpd (for csc-sun.math.utah.edu [128.110.198.2]) with SMTP; 25 Sep 1997 10:10:37 UT Received: from gate1.ams.org by AXP14.AMS.ORG (PMDF V5.1-8 #1) with SMTP id <01IO1R2KS0R40020II@AXP14.AMS.ORG> for Beebe@Math.Utah.edu; Thu, 25 Sep 1997 05:28:49 EST Received: from pillar.elsevier.co.uk ([193.131.222.35]) by gate1.ams.org via smtpd (for axp14.ams.org [130.44.1.14]) with SMTP; Thu, 25 Sep 1997 09:28:40 +0000 (UT) Received: from snowdon.elsevier.co.uk (snowdon.elsevier.co.uk [193.131.197.164]) by pillar.elsevier.co.uk (8.8.5/8.8.5) with ESMTP id KAA21768 for ; Thu, 25 Sep 1997 10:27:59 +0100 (BST) Received: from SRAHTZ (actually host srahtz.elsevier.co.uk) by snowdon.elsevier.co.uk with SMTP (PP); Thu, 25 Sep 1997 10:27:58 +0100 Date: Thu, 25 Sep 1997 09:57:42 +0100 From: s.rahtz@elsevier.co.uk (Sebastian Rahtz) Subject: Re: Right-to-left TeX and LaTeX. In-reply-to: To: rama@cc.huji.ac.il Cc: dak@neuroinformatik.ruhr-uni-bochum.de, tex-implementors@MATH.AMS.ORG Errors-to: tex-implementors-request@MATH.AMS.ORG Message-id: <6711-Thu25Sep1997095742+0100-s.rahtz@elsevier.co.uk> MIME-version: 1.0 X-Mailer: VM 6.33 under Emacs 19.34.4 Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit References: <199709241308.PAA03700@maloche.neuroinformatik.ruhr-uni-bochum.de> > In a right-to-left document, the order should be reversed; > that is, something similar to > \renewcommand\thesubsection {\@arabic\c@subsection.\thesection} I don't really see the issue. create a `hebart.cls' which loads article.cls, and then redefine what you need. > How is this to be solved without changing article.cls? > And the class files are FULL of those things. > write your own classes? or do you mean you want \documentclass{article} to produce proper results in Hebrew? how do you indicate you are going to use Hebrew? with a \usepackage{hebrew}? but i concede you have a point, it might be nicer if the standard classes exposed more hooks for people like you. But: Sebastian's Prediction: latex2e standard classes will not be changed for Hebrew. I think you should a) write some decent classes of your own b) lobby for LaTeX3 to give you more hooks sebastian