mcglk@scott.stat.washington.edu (Ken McGlothlen) (06/15/88)
I've been bouncing around an idea for some time, and it seems time to get some opinions from a lot of people who've been at this longer than I have. What I'd like to do is write a program to convert METAFONT source programs into PostScript. To be honest, I haven't done a lot of analysis on this topic, but it seems the advantages would be several-fold. 1) No big bitmaps. Even PK files can get messy, and you need one for cmr5, cmr7, cmr8, cmr10, cmr12, and so on. Once the font has been compiled, you need to run the compiler again to get a slightly different font. 2) Resize and reshape on the fly. With one PostScript program (albeit a somewhat large PostScript program), you could resize fonts on the fly, like 10 % pointsize 4 mag % magnification /cmr % fontname setTeXfont I assume that setTeXfont would only set up parameters, since there would be another procedure (say, draw_shape) that would actually USE the parameters in order to sketch out the character. 3) *Fast* DVI->PS conversion, *fast* downloading (hopefully). Instead of having to download those bitmaps every time you want to go to nine-point cmr from ten-point, you download the cmr_draw procedure once, and just change a couple of parameters on the fly. 4) If wanted, half-toning, patterns, etc. Why recompile in METAFONT to get an 80%-gray font? PostScript does it fairly fast. I am interested in writing a program of this sort, but I'd be interested in trying. I'd be interested in hearing your comments BEFORE I start, though--I have a tendency to get myself in too deep too quickly, and this is one of those things I'd rather do right the first time. :) Thanks in advance for your comments. --Ken McGlothlen mcglk@max.acs.washington.edu mcglk@scott.ms.washington.edu
dow@wjh12.harvard.edu (Dominik Wujastyk) (06/15/88)
In article <911@entropy.ms.washington.edu> mcglk@scott.ms.washington.edu (Ken McGlothlen) writes: >I've been bouncing around an idea for some time, and it seems time to get >some opinions from a lot of people who've been at this longer than I have. > >What I'd like to do is write a program to convert METAFONT source programs >into PostScript. > A program to do this has already been done, by Leslie Carr of Southampton University (UK). He has written up what he did as "Of Metafont and PostScript", in _TeX Users Group Eighth Annual Meeting: Conference Proceedings_, (Providence: AMS, 1988). This is number 5 in the TUG TeXniques series. Leslie worked by converting not the GF file, but the LOG file! (Verbose version.) It really is very interesting, and amazing how well it worked. Dominik Wujastyk -- bitnet: user DOW on the bitnet node HARVUNXW arpanet: dow@wjh12.harvard.edu csnet: dow@wjh12.harvard.edu uucp: ...!ihnp4!wjh12!dow
candy@umb.umb.edu (declarer/Karl B./dummy) (06/16/88)
Leslie Carr's article was interesting, and messing around with the transcript file is certainly the only way to go , but (a) Metafont pens are not converted into outlines (i.e., splines), hence MF programs that use pens (like Computer Modern) won't be readily translatable, and (b) His results are interesting from a computer scientist's point of view, but from a type designer's point of view they are, sad to say, not worth too much. His resulting type doesn't look anything like Computer Modern; it's sort of a bastardization of a modern type and an oldstyle type. The transitional serifs (with a curve, instead of meeting at right angles), are especially jarring. This is an example of a general problem in computer science. (It seems to me.) Computer people know a lot about computers; but computers can simulate almost (perhaps in a few decades we'll be able to strike the ``almost'') anything, but only if the person doing the simulating knows as much as an expert in that field. (In this case, type design.) This is not a common combination. (In the Metafontbook, Don Knuth says he hopes the best designs will result from collaborations between designers who aren't experts in Metafont (i.e., computers), and programmers who aren't experts in type design.) In other fields, such as art (the Science news article a while back about computer-generated ``Mondrians'', while good enough to fool me, didn't even cause my friend with an art background to blink an eyelash -- out of the four there, she pointed to the genuine Mondrian without hesitation), music, and on and on. As programmers get better at programming, they want to do more challenging things, which unfortunately often leads to not-so-wonderful results in terms of the field they're trying to write programs for. Until universities or businesses start funding ``computer science and X'' programs, I think the results will fall short. Karl. karl@umb.edu Kathy. letters@umb.edu
richard@gryphon.CTS.COM (Richard Sexton) (06/17/88)
In article <230@wjh12.harvard.edu> dow@wjh12.UUCP (Dominik Wujastyk) writes: >In article <911@entropy.ms.washington.edu> mcglk@scott.ms.washington.edu (Ken McGlothlen) writes: >> >>What I'd like to do is write a program to convert METAFONT source programs >>into PostScript. > >A program to do this has already been done, by Leslie Carr of Southampton >University (UK). He has written up what he did as "Of Metafont and >PostScript", in _TeX Users Group Eighth Annual Meeting: Conference >Proceedings_, (Providence: AMS, 1988). This is number 5 in the TUG >TeXniques series. Please go on. -- "Shrimp Ahoy" richard@gryphon.CTS.COM {backbone}!gryphon!richard
dow@wjh12.harvard.edu (Dominik Wujastyk) (06/18/88)
In article <4506@gryphon.CTS.COM> richard@gryphon.CTS.COM (Richard Sexton) writes: >In article <230@wjh12.harvard.edu> dow@wjh12.UUCP (Dominik Wujastyk) writes: >>In article <911@entropy.ms.washington.edu> mcglk@scott.ms.washington.edu (Ken McGlothlen) writes: >>> >>>What I'd like to do is write a program to convert METAFONT source programs >>>into PostScript. >> >>A program to do this has already been done, by Leslie Carr of Southampton >>University (UK). He has written up what he did as "Of Metafont and >>PostScript", in _TeX Users Group Eighth Annual Meeting: Conference >>Proceedings_, (Providence: AMS, 1988). This is number 5 in the TUG >>TeXniques series. > >Please go on. I have just flipped through the article, thinking about doing a precis of it for this news group. But it has diagrams, is fairly detailed and hard to summarise, and, hang it all, the whole book of proceedings only costs $15. You can order it from the TeX Users Group, P.O.Box 9506, Providence, RI, 02940, USA. Phone: (401) 751 7760 TUG will even accept credit card orders on the phone, so it's really easy. And then you have the real thing. The same volume also has papers on TeX and typesetting Greek, Japanese and Turkish, with attention to the fonts needed for each. Dominik Wujastyk -- bitnet: user DOW on the bitnet node HARVUNXW arpanet: dow@wjh12.harvard.edu csnet: dow@wjh12.harvard.edu uucp: ...!ihnp4!wjh12!dow