Return-Path: Received: from [62.76.169.38] (HELO video) by vsu.ru (CommuniGate Pro SMTP 3.5b3) with ESMTP id 3826358; Wed, 19 Sep 2001 13:05:15 +0400 To: CyrTeX-en@vsu.ru, CyrTeX-ru@vsu.ru Cc: tex-fonts@math.utah.edu, tetex@informatik.uni-hannover.de Subject: Re: CM-Super package released References: <20010919.072219.21924139.wl@gnu.org> From: Vladimir Volovich Date: Wed, 19 Sep 2001 13:05:28 +0400 In-Reply-To: <20010919.072219.21924139.wl@gnu.org> (Werner LEMBERG's message of "Wed, 19 Sep 2001 07:22:19 +0200 (CEST)") Message-ID: Lines: 82 User-Agent: Gnus/5.090004 (Oort Gnus v0.04) Emacs/21.0.103 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii "WL" == Werner LEMBERG writes: WL> Am I right that the CM-super package tries to minimize the number WL> of used fonts? If this is the most important goal, I fully agree WL> with such big fonts. yes, the goal was to minimize the number and total size of font files. each type1 font shares all common glyphs which are present in supported encodings. e.g. latin letters, russian letters, accents and a number of other glyphs which are present in more than one font encoding are stored only once in type1 fonts. WL> Otherwise I wonder whether the Omega approach isn't probably WL> better: to have `glyph containers' with at most 256 characters WL> each -- fonts like omsela.pfb for Latin, omsecy.pfb for Cyrillic, WL> etc. but, say, omsecy.pfb is only one shape? and EC fonts (and CM-Super fonts) contain 29 multiple-size font shapes (and also 13 one-size shapes) totalling 401 font files. so if we split each font to glyph containers, we get not 401 font files, but much more (say, 401 for T1, 401 for TS1 and not smaller than 802 for cyrillic). >> * provide optimized but not hinted version (will be smaller)? a note: e.g. sfrm1000.pfb hinted is ~139kb, while simply optimized but not hinted is ~107kb. WL> Apropos smaller: How do you realize composite glyphs? Do you use WL> Type 1 facilities to make the fonts smaller (via the AFM file)? WL> I can imagine that this can greatly reduce the size of the fonts. not yet :) you mean to use the seac instruction to put accents? (or is it sufficient to not include accented glyphs in pfb files at all, but only define composites in the AFM files?) >> * add glyphs from T2D encoding WL> So T2D this has already been defined, hasn't it? I haven't seen WL> any announcement on this list... strictly speaking, A.Berdnikov defined T2D at Eurotex'99, and current LH fonts (at ftp.vsu.ru/pub/tex/font-packs/lhfnt) support T2D, but not in all shapes (and i guess that some font shapes are not even planned). but the support for T2D was not yet included into the LaTeX cyrillic bundle. as A.Berdnikov himself agreed about T2D, it is not more than a demo encoding, not suited to support Church-Slavonic language, but is to be used only for a limited demonstration of Slavonic writing. there is currently a very fruitful discussion on standardizing Ch-Sl input method (so-called HIP) and font encoding in one of russian-language mailing lists. and T2D is not consistent in my opinion (it unfortunately defines some variant glyphs which are not to be used in one font shape, e.g. both "modern" and "old" variants of CYRN are present in the same font, while this should obviously be the same glyph). That's why (and also because of not full support for all font shapes) the glyphs from T2D were not included into this first versio nof CM-Super fonts... >> * include glyphs from cbgreek ec-like fonts WL> * Include glyphs from the vnr family to cover Vietnamese. WL> * Include all missing glyphs from the Unicode extended Latin WL> sections. added to TODO list in the README :) WL> BTW, I hope you have replaced the ugly `Euro' glyph from TS1 with WL> better looking variants from either the Marvosym or China2e WL> fonts... WL> Another BTW: Do you use AGL compliant glyph names? yes, when possible (there are some glyphs which are neither present in AGL nor in Unicode, not only in T2*/X2 encodings, but also in TS1). Best, v.