4. Extend the formula langauge with a "cases" construct. 7. Fix translation of slides into TeX. Tables and other figures should not float. This requires context sensitive translation and will have to wait for James Clark's replacement for sgmlsasp. (As of 1.3, there is an attribute for keeping these from floating.) 11. Add an index to the manual 21. Extend the DTD and the LaTeX mapping to support the automatic generation of indexes. Talk to Hans-Werner about this. 22. Extend biblio DTD to allow cross references starting tag, at the risk of breaking some documents. 25. The roff mapping for elements is buggy. Tables may not be included within EQN formulas. 26. The roff mapping for elements should be changed. Font requests are not allowed within equations by EQN. 27. Need some way to have successive cites be formatted properly by LaTeX. Refer does the right thing. I thought we might just be able to change the and elements to allow cdata content, instead of empty content, but this is not upwardsly compatible. When content is empty the end tag may be ommitted. Once content is permitted, text after existing cites gets interpreted as being part of the citation if the end tag was ommitted. Try something like this: This CDATA might need to be RCDATA, to allow entities for non-breaking spaces in "pg. 23", for example. 28. German quotation marks, when german style option is selected. 29. "Motto" element, for quotations at the beginning of a chapter or section. 30. Change the SGML declaration so that is uses the Latin 1 character set. Write sed scripts for converting between Latin 1 and Macintosh character sets. This has several advantages over using entities: 1) Much easier to type German using Author/Editor, at least until it supports keyboard macros. 2) Search and replace of non-ASCII characters works in Author/Editor, where entities cannot be included within search or replacment strings. 3) Only need one replacement file for the Latin 1 entities. TeX 3.0 and groff both support Latin 1. Disadvantage: Macinthosh Author/Editor files which use non-ASCII characters will need to be translated into Latin 1 to satisfy the DTD. (Leave the entities for Latin 1 characters in the DTD, for upwards compatability and to support 7 bit ASCII.) 31. Consider reducing the set of mappings, to ease maintenance. 32. Remove short references maps from the DTDs to the extent they are dependent on a particular processor or formatter, such as TeX. Instead, include formatter dependent short reference maps using the same technique now used for entity translations. This would also solve the problem of having to comment out short reference maps for use with Author/Editor, which doesn't support them. 33. Add entities for large math operators. (e.g. bigcup)