December 13, 1992 ############################################ # # # WELCOME to the METAFONT list # # # ############################################ Dear friends, it is a great pleasure for me to welcome you to this new discussion list. Yesterday, Dec. 12, you were already 110 having joined this list, from all over the world: France, Germany, US, UK, the Netherlands, Austria, Australia, Greece, Norway, Sweden, Danemark, Canada, Israel, Italy, Ireland, Japan, Taiwan, and other countries... Thank you! I would like to start by some personal thoughts on the past/present/future of Metafont (which I will call MF for brevity). First of all some historical background: when D. E. Knuth (whom I will call DEK for brevity, and without any loss of respect) first released TeX and MF, in the days where TeX was not yet frozen, they were quite a succesfull pair: TeX needed a lot of fonts to do a correct job, and MF provided all of these fonts thanks to the great idea of metaness. There was, though, one small problem: plain and latex TeX were build upon Computer Modern fonts (which later evolved to DC fonts, without loosing their shapes and meta-properties). It seemed like an infernal dilemma (or a Kobayashi-maru problem, less optimistic people would say): CM/DC fonts had metaness but didn't look nice, other fonts (for example PostScript fonts) looked nice but had no metaness; and what would you prefer: a TeX file with optically corrected but otherwise medium fonts, or with good-looking fonts without optical correction? Depending on their personality and on their needs, people were choosing one or the other solution. Things got more complicated when Blue Sky and Y&Y released PostScript versions of CM fonts. Now one could even use Computer Modern fonts in other programs, like Illustrator or Corel Draw. But this didn't make them look nicer. But it proved however, that MF and CM fonts are two different issues. The bottom line of these three paragraphs is that people tend to confuse the use of MF and the one of CM fonts (said in another way: the contents of volumes B versus E of ``Computers and Typesetting''). This confusion comes from the fact that they were always forced to use the one in conjunction with the other... But what is so wrong with CM fonts? Well, in my opinion, DEK wanted to show to the academic community that it was *possible* to create a complete meta-family. He wanted to put all of MF's strength into this one and only example. And he certainly suceeded. And what happened after that? Thousands of people were using this one-and-only example to write hunderts of thousands of papers, books, reports etc etc in the whole world... now is that a way to treat an example? certainly not. It is a fine thing when a font has different optically corrected sizes. It is even nicer when it has different weights and styles (italics, small capitals) all coming from the same code. But is it really necessary to have sans serif, typewriter, funny, dunhill, fibonacci, etc etc out of it? The answer is NO. Sans serif fonts are not so old in typography, and typewriter ones even less. There hasn't been enough time to establish a tradition on these and on their relationship with regular serif(ed) fonts. What both TeX and MF desperately need are new good-looking *efficient* meta-families. And by *efficient* I mean that metaness should be restricted to what is really necessary: optical correction and weight. At this very moment there is a new technology pretending to cover these issues, namely Multiple Masters by Adobe. This is good for MF (because people get sensibilized to the problem of optical correction, which was a real tabou at the times of pure PostScript) and bad at the same time (they already have Minion, and Myriad, and preparing Bodoni---we still have *only* CM...) So I sincerely believe that it is time to start working on a new font environment for TeX, a new use of MF. Typographers with which I have discussed (Richard Southhall, Ladislas Mandel) are always concerned by rasterization; they are still thinking in terms of pixels. MF solves rasterization problems better than type 1 hints; but this will soon be an empty problem: resolutions of printing devices get higher and higher, and soon nobody will care about pixels anymore (since they won't be visible anyway...) Our main concern should be metaness. We must find efficient ways of (1) coding shapes of good-looking fonts (either already existing, or new ones) in MF (2) adding metaness to obtain all possible weights (3) adding metaness for optical correction (4) kerning. A big step concerning point (1) has been done by Erik-Jan Vens and his (newest version of) ps2mf package. He obtains control points and Bezier curves in MF notation (z1, z2 etc) out of PostScript type 1 fonts. One may argue if that is legal, one may also argue if these outlines are good enough. But Erik-Jans' utilities exist and they *are working*: they are a beautiful starting point. Next we can take a font out of Erik-Jans ps2mf and perform steps (2) to (4). It certainly is hard work, but the result will be a new meta-family for TeX, TEN YEARS after DEK having released CM fonts!! Will the result still be under copyright protection after all these changes? another open question. I would be extremely happy if this list serves as a starting point for the revival of creative MF; of course many technical and legal aspects remain to be discussed... but remember: glory and fame await those who will give the TeX community the gift of a new meta-family!! ############################################################################# Second issue: MF for beginners. The first contact Ben User has with MF is for compiling fonts he/she needs for his/her document. But he/she has still to think about quite a few things: -- resolution of the printer/fax/screen -- other characteristics of the very same machine (mystically known as mode_def) -- pointsize of the font -- magnificaion of the font not to mention those terrifying error messages he/she might get: Strange path (I wonder if DEK was reading Lovercraft when he wrote that pool file) etc etc Font organisation is a complex thing and we are not going to change that... but we can establish a clear introduction to these issues to help newcomers find their way out of the "strange path" of font compilation. Geoffrey Tobin (who is on this list [Hi Geoffrey!], not to be confused with Georgia Tobin) has written a nice introduction to the use of MF for compiling fonts. If he agrees I would like to make this paper of his the starting point of a FAQ file which TUG will put on every possible server. Suggestions concerning this file are welcome; for the moment, you can find it in /pub/tex/yannis/MF-list, by anonymous ftp at spi.ens.fr (IP 129.199.104.3), under the name tobin.tex. ############################################################################# Final issue: a (Christmas) gift for all of you (having patiently read all of my chatter...) Two files: multicma10.mf multicm.tex to be found in directory /pub/tex/yannis/MF-list/multicm (same server spi.ens.fr, IP 129.199.104.3 --- for those who don't have ftp access, just send me a message and I'll send you the files through regular email). It's supposed to be a surprise, so I won't tell you what these files do; I hope you'll like the result. Here is what to do: Run file multicma10.mf with MF and GFtoPK, at mag:=3.583; put the TFM and PK files where they are supposed to be; then run multicm.tex. If you like the result, you can make other files multicme10.mf etc etc, you'll find the instructions inside. Enjoy! ############################################################################# Sincerely yours, Yannis Haralambous 187, rue Nationale 59800 Lille, France Email: yannis@gat.citilille.fr (Internet) Fax: (33) 20.40.28.64