X-VM-v5-Data: ([nil nil nil nil nil nil nil t nil] [nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil nil]) Date: Wed, 27 Jun 90 15:01:40 CET Reply-To: LaTeX-L Mailing list From: PZF5HZ@RUIPC1E.bitnet To: Rainer Schoepf Status: R X-Status: X-Keywords: X-UID: 142 I have not seen any examples of the last two: ie actions to perform whenever a font is used. Thus difficult to say anything definite, but the "out-of-math" alternative seems as useful as the "in-math alternative": is it impossible/impractical/useless to combine these two and let the user (of the scheme) make the distinction with \ifmmode \else \fi ??? We used the math action to set or reset accent characters in certain math alphabets in the amstex option. of course it would be possible to use \ifmmode; this is macros names versus space and time. I vote for different storrage if I build this. Will the existence of such an extra slow it down a lot? Obviously, actually doing a remapping will, but there is no way round that: if remapping is needed then the time has to be taken to do it. If the action is performed for every \selectfont command it will slow down things a lot. But of course one can build a scheme where remapping (for example) is done only if something has changed. In such a case every \selectfont would only add one extra call to a hook which would be normally empty. *************** Some people like to change the \fontdimen parameters font the same font for different uses or add extra \fontdimen s. the latter has to be done immediately the font is loaded, and the former is best done then. Thus these font-scheme implementors could not use your scheme as it is now. They could, of course, achieve the same by building different copies of the .TFM file with the differnt values, but should they have to??? ie. should you deny them access to some of TeX's adaptability if they use your scheme? **************** The addition of extra fontdimen parameters could be done via first argument of \extra@def provided the font wasn't loaded already. But in this case there is no chance at all to add another fontdimen (problem in TeX (misfeature)). This would apply to all fonts in a family. For those who saw the current versions of the macros: there is also a (currently) unused #5 argument to \new@fontshape which was added to extend the fontselection scheme to carry out this specific to every fontshape. I don't see why changing \fontdimen parameters can't be done with the new scheme. On the fly it can be done by defining macros as proposed by Sebastian (back to the world?) i.e., by saying ..\selectfont \fontdimen5\font 4pt ... in a macro. If it is necessary only once it can also be done in \extra@def of the corresponding family. (I did this for the euler family in concrete.sty to use the fonts in normal text. I see the above opinion as a sign that the new font selection isn't explained enough. Maybe we can help by writing a better interface description. Frank Would be fine to here some more opinions!!!!!