At 11:51 +0100 2001/02/18, Javier Bezos wrote: >> Since we have been asked to provide input encoding changes for LaTeX within >> paragraphs, eg for individual words, something like this would happen >>if such >> a change appears, say, inside the argument of \section. > >A system to coordinate preprocess and "internal" process is necessary. The way I thought of it, the preprocessor should be able to handle mixed encodings. -- Thus, the extended TeX (and LaTeX) only sees Unicode characters, and nothing else. Also, I think that the use of multiple encodings in a single file is a pretty transitory thing: MacOS X, released in a regular version the next month supports Unicode fully -- so the access to editors able to handle Unicode will happen pretty soon (no more than a few years), as the availability on personal computers will push the developments a great deal. And the reason for using multiple encodings is probably the result of the lack of editors that can handle Unicode. So, I do not think it matters if one uses a seemingly complicated system, with additional files specifying encoding for now: People will probably soon want to be able to translate their multiple encoding files to single Unicode encoding files instead. (If you formerly wrote files with a mixture of say Russian and Latin encodings, where it was only possible to see the correct renderings by changing the settings of the editor, then when you get hold of a Unicode editor, the first thing that you would want is to not having the bother of changing the settings of the editor all the time. Thus, you would want a convenient way of converting your old files to Unicode so that your new editor can read them. Therefore it is best if these old mixed encodings files already have a markup that admits an easy conversion to Unicode.) It is always difficult to judge the future, but, well, this is my guess. Hans Aberg