The Free UCS Outline Fonts Project — an Attempt to Create a ... - TUG
The Free UCS Outline Fonts Project — an Attempt to Create a ... - TUG
The Free UCS Outline Fonts Project — an Attempt to Create a ... - TUG
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
<strong>The</strong> <strong>Free</strong> <strong>UCS</strong> <strong>Outline</strong> <strong>Fonts</strong> <strong>Project</strong><strong>—</strong><strong>an</strong> <strong>Attempt</strong> <strong>to</strong> <strong>Create</strong> a Global Font<br />
Primož Peterlin<br />
University of Ljublj<strong>an</strong>a<br />
Faculty of Medicine, Institute of Biophysics<br />
Lipičeva 2, SI-1000 Ljublj<strong>an</strong>a, Slovenia<br />
primoz.peterlin@biofiz.mf.uni-lj.si<br />
http://biofiz.mf.uni-lj.si/~peterlin/<br />
Abstract<br />
In February 2002, the <strong>Free</strong> <strong>UCS</strong> (Universal Character Set) <strong>Outline</strong> <strong>Fonts</strong> project (http://<br />
sav<strong>an</strong>nah.gnu.org/projects/freefont/) was started.Exercisingtheopen-sourceapproach,<br />
itsaimis<strong>to</strong>providease<strong>to</strong>ffreeTimes-,Helvetica-<strong>an</strong>dCourier-lookalikesavailableintheOpen-<br />
Typeformat,<strong>an</strong>dprogressivelycoverthecompleteISO10646/Unicoder<strong>an</strong>ge.Inthisstageofthe<br />
project, we focus mainly on two areas: collecting existing fonts that are both typographically <strong>an</strong>d<br />
license-wise(i.e.,GNUGPL)compatible<strong>an</strong>dc<strong>an</strong>beincluded<strong>to</strong>covercertainpartsofthecharacterset,<strong>an</strong>dpatchingupsmallerareasthatarenotyetcovered.Pl<strong>an</strong>nedfutureactivitiesinvolvetypographicrefinement,extendingkerninginformationbeyondthebasicLatinarea,includingTrue-<br />
Typehintinginstructions,<strong>an</strong>dfacilitatingtheusageoffontswithvariousapplications,includingthe<br />
TEX/Ω typesettingsystem.<br />
Résumé<br />
Le projet de fontes vec<strong>to</strong>rielles libres pour <strong>UCS</strong> (http://sav<strong>an</strong>nah.gnu.org/projects/<br />
freefont/) adémarréenfévrier2002.Àtraversl’OpenSource,lebutdeceprojetestdefournir<br />
un ensemble de clones libres de Times, Helvetica et Courier, disponibles d<strong>an</strong>s le format Open-<br />
Type, et couvr<strong>an</strong>t progressivement l’ensemble des caractères d’ISO 10646/Unicode. À ce stage<br />
du projet nous nous concentrons sur deux domaines : la collection de fontes exist<strong>an</strong>tes qui soient<br />
compatibles avec notre projet, aussi bien du point de vue typographique que du point de vue de<br />
lalicence(GNU GPL),quipuissentêtreintégréespourcouvrircertainespartiesdel’ensemblede<br />
caractères; et d’autre part, faire des ajouts de petites régions qui n’ont pas encore été couvertes.<br />
Nos activités futures prévoient le perfectionnement typographique, le crénage au-delà de l’alphabet<br />
latin, l’incorporation dehints TrueType,et lesupport deplusieurs applications, dont aussiles<br />
systèmes TEX et Ω.<br />
Introduction<br />
<strong>The</strong> aim of the <strong>Free</strong> <strong>UCS</strong> <strong>Outline</strong> Font project is <strong>to</strong><br />
provide a st<strong>an</strong>dardised set of glyphs which make a harmonised<br />
design despite including glyphs from different<br />
scripts(Latin,Cyrillic,Greek,Armeni<strong>an</strong>,etc.).Itisclear<br />
thatthisrequirescompromisesatthecos<strong>to</strong>ftypographic<br />
finesse(see[11]foradiscussionoftypographiccompromises<br />
regarding the Greek alphabet), yet the end result<br />
mustlookacceptableforgeneraluse,includingelectronic<br />
mail,world-wideweb, <strong>an</strong>dtext edi<strong>to</strong>rs.<br />
Whilethisisclearlynotthefirstattempt<strong>to</strong>createa<br />
typefacecoveringglyphsbeyondLatin,previousattempts<br />
are not as numerous as one might imagine. Bigelow [5]<br />
quotes Romulus by J<strong>an</strong> v<strong>an</strong> Krimpen from 1931 as one<br />
of the first examples. Another example is Nikola<br />
Nikolaeviq Kudrxov [Nikolaĭ Nikolaevich Kudryashov]’s<br />
Encyclopaedia (Kudrxovska enciklopediqeska<br />
[Kudryashovskaya entsiklopedicheskaya])<br />
family (1960–1974), originally designed for the third<br />
edition of the Great Soviet Encyclopaedia which contains<br />
thous<strong>an</strong>ds of glyphs, including Cyrillic, Latin, <strong>an</strong>d<br />
Greeklettersinserif<strong>an</strong>ds<strong>an</strong>s-serifstyles,aswellasother<br />
specialsigns<strong>an</strong>dsymbols.Mostpopulartypefacessuchas<br />
St<strong>an</strong>leyMorison’sTimesRom<strong>an</strong>,MaxMiedinger’sHelvetica,Adri<strong>an</strong>Frutiger’sUnivers<strong>an</strong>dEricGill’sGillS<strong>an</strong>s<br />
wereoriginallydesignedfortheLatinalphabet<strong>an</strong>dlater,<br />
astheygainedpopularity,extended<strong>to</strong>Greek<strong>an</strong>dCyrillicalphabets.<br />
AtypefacecoveringthewholeISO10646/Unicode<br />
r<strong>an</strong>gehas<strong>to</strong>bedesignedflexiblyenough<strong>to</strong>allowaddition<br />
of scripts not belonging <strong>to</strong> the Western typographic tradition.<strong>The</strong>taskisnoteasy,<strong>an</strong>ditisnotsurprisingthatso<br />
farthereareveryfewaestheticallysatisfyingtypographic<br />
solutions. Among those one has <strong>to</strong> mention Bigelow <strong>an</strong>d<br />
Holmes’s LucidaS<strong>an</strong>sUnicode [5] <strong>an</strong>d Haralambous <strong>an</strong>d<br />
Plaice’s Ω [12]. Other solutions like ArialUnicode [16]<br />
<strong>an</strong>d James Kass’s compendium Code2000 [14] are <strong>to</strong>o<br />
variedinstyle <strong>to</strong> form aunifiedtypeface.<br />
<strong>TUG</strong>boat,Volume24 (2003), No.3<strong>—</strong>Proceedingsof EuroTEX2003 545
PrimožPeterlin<br />
Design issues<br />
His<strong>to</strong>rical <strong>an</strong>d cultural context. It is clear that a typeface<br />
aiming <strong>to</strong> cover “all the scripts of the world” c<strong>an</strong><br />
not rely on his<strong>to</strong>rical styles (e.g., Renaiss<strong>an</strong>ce, Baroque,<br />
etc.) known from Western typography, as most of the<br />
worldhasnotexperiencedtheseperiodsintheevolution<br />
of typography. Trying <strong>to</strong> extend them <strong>to</strong> non-Europe<strong>an</strong><br />
scripts 1 isasinappropriateas,say,trying<strong>to</strong>designaLatin<br />
alphabetinKufi style.<br />
Eventheapparentlylogicaldivisionin<strong>to</strong>seriffed<strong>an</strong>d<br />
s<strong>an</strong>s-serif typefaces is eurocentric. To see why, one only<br />
has<strong>to</strong>thinkoftheoriginofLatincapitalletters.Serifsare<br />
<strong>an</strong> invention of Rom<strong>an</strong> s<strong>to</strong>ne-carvers<strong>—</strong>a smaller stroke<br />
perpendicular <strong>to</strong> the main stroke was added <strong>to</strong> provide<br />
a uniform smooth finishing of the stroke. Here, the visual<br />
effect followed the available technology. <strong>The</strong> modulated<br />
stroke<strong>—</strong><strong>an</strong>other <strong>an</strong>cient Rom<strong>an</strong> typographic invention<strong>—</strong>is<strong>an</strong>oppositeexample,wherethetechnology<br />
followed the visual effect [13]. <strong>The</strong> need <strong>to</strong> strengthen<br />
vertical strokes arose once the Rom<strong>an</strong>s started <strong>to</strong> erect<br />
monuments of monumental proportions, where the inscriptions<br />
were no longer in the eye-height. While the<br />
rain <strong>to</strong>ok away the paint from the vertical strokes, dirt<br />
<strong>an</strong>d dust accumulated in the horizontal strokes, which<br />
thus appeared optically heavier. Physical broadening of<br />
verticalstrokes wasintroduced as acompensation, inorder<br />
for the horizontal <strong>an</strong>d vertical strokes <strong>to</strong> look optically<br />
equivalent. <strong>The</strong> Rom<strong>an</strong> technological innovations<br />
survivedfor 2000 years inEurope<strong>an</strong> typography. However,<br />
in countries where the letters were painted with<br />
brush on<strong>to</strong> silk or carved with a needle on<strong>to</strong> a palm leaf<br />
ratherth<strong>an</strong>carvedin<strong>to</strong>thes<strong>to</strong>newithmallet<strong>an</strong>dchisel,<br />
suchtypographic developmentneverhappened.<br />
Similar concerns about eurocentrism are valid for<br />
differingtheupright<strong>an</strong>dtheitalicforms,orforthatmatter,<br />
even between majuscules <strong>an</strong>d minuscules, capital <strong>an</strong>d<br />
smallletters.Evendifferingbetweenupright<strong>an</strong>dsl<strong>an</strong>ted<br />
forms isquestionable,as sl<strong>an</strong>tedforms make no sense in,<br />
say, most nativeAsi<strong>an</strong>scripts.<br />
<strong>The</strong> least questionable seems <strong>to</strong> be the differences<br />
basedontheweigh<strong>to</strong>fthetypefaces,whichappears<strong>to</strong>be<br />
almostuniversal.<strong>The</strong>onlyexceptionknown<strong>to</strong>theauthor<br />
isEthiopic, 2 wherethewordsweretraditionallyemphasisedbyprintingthemin<strong>an</strong>othercolour(redinreligious<br />
texts, blue in imperial decrees) or by underlining them<br />
or enclosingtheminovals.<br />
Legibility. A typeface which includes non-Latin scripts<br />
offers <strong>an</strong> opportunity for exploring the “universal” parameters<br />
of legibility. While there has been a wealth of<br />
1. Eventhoughhis<strong>to</strong>ricalstylesdifferthroughoutEurope,Europeisneverthelesstreatedhereasahis<strong>to</strong>rical<strong>an</strong>dculturalunity<br />
when contrasted with the rest of the world.<br />
2. D<strong>an</strong>iel Yacob, personal communication.<br />
publicationonlegibilitycentredonLatinscript[20,22,<br />
9, 15], exploring the fac<strong>to</strong>rs like weight, serif vs. s<strong>an</strong>sserif<br />
faces, x-height, capitalisation etc., it is clear that<br />
suchdifferencesareminorcompared<strong>to</strong>thedifferencesin<br />
shapebetweenLatin<strong>an</strong>d,say,Hebrew,Arabic,Dev<strong>an</strong>agari<br />
or Tamil, which nevertheless provide roughly the<br />
sameleveloflegibility.Lackingcomparativecross-script<br />
studies,theexperimentsinlegibilityremainintherealm<br />
ofthetypographer’sintuition.Onthebrighterside,makingmulti-scripttypefacesavailable,albeitnotperfect,is<br />
astep<strong>to</strong>wardstheworldwheresuchcross-culturalstudieswillbe<br />
easier<strong>to</strong> achieve.<br />
Methodology<br />
<strong>The</strong>basicideabehindthefree<strong>UCS</strong>outlinefontsproject<br />
was <strong>to</strong> collect various available free outline fonts, covering<br />
single national scripts, <strong>an</strong>d <strong>to</strong> compile them in<strong>to</strong> a<br />
largefontusingtheISO10646/Unicodecodedcharacter<br />
set [21], taking in<strong>to</strong> consideration typographic <strong>an</strong>d legal<br />
compatibility,<strong>an</strong>dfillinginthemissingareasontheway.<br />
<strong>The</strong>wholedevelopmentwaspl<strong>an</strong>ned<strong>to</strong>becarriedinthe<br />
open-source m<strong>an</strong>ner, with m<strong>an</strong>y developers using a central<br />
reposi<strong>to</strong>ry.<br />
<strong>The</strong> general requirement for technical realisation<br />
wasthattypefacesneed<strong>to</strong>beavailableasscalablevec<strong>to</strong>r<br />
fonts.<strong>The</strong>actualtechnicalrealisation(PostScriptType1<br />
[1] uses cubic Bézier splines, while TrueType [3] uses<br />
quadratic ones) was considered secondary, because at<br />
leastinprincipleitispossible<strong>to</strong>tr<strong>an</strong>sformthefontsfrom<br />
oneform<strong>to</strong><strong>an</strong>other.Inreality,though,noknowntr<strong>an</strong>sformationsiscompletelylossless<strong>—</strong>kerning<strong>an</strong>dhintsare<br />
usuallythemost volatile.<br />
Licenses of used sources. We <strong>an</strong>ticipated that our result<br />
will contain glyphs from m<strong>an</strong>y different sources. Thus,<br />
special attention was given <strong>to</strong> the license under which a<br />
font is released. <strong>The</strong> license we looked for should allow<br />
redistribution,modification<strong>an</strong>ddistributionofmodified<br />
font files. M<strong>an</strong>y free <strong>an</strong>d open-source licenses fulfil this<br />
requirement. As the URW++ core PostScript fonts, the<br />
Ω-Serif<strong>an</strong>dsomeothermajorsourceswerereleasedunder<br />
the GNU General Public License [10], we adopted<br />
it for our project as well. <strong>The</strong> license itself is suited<br />
for programs rather th<strong>an</strong> typefaces<strong>an</strong>d itsapplication<strong>to</strong><br />
fontsmaybelegallydubious,eventhough,say,PostScript<br />
Type1fontsareperfectlylegalprogramswritteninPost-<br />
Script. We were not able <strong>to</strong> find <strong>an</strong>y license in the same<br />
spirit pertaining specifically <strong>to</strong> fonts, though. It may be<br />
worthseeingthefinallicenseoftheBitstreamVerafonts,<br />
whichwere<strong>an</strong>nouncedinJ<strong>an</strong>uary2003<strong>to</strong>besoonavailableunderalicenseinthe<br />
open-sourcespirit.<br />
We thus limited our search <strong>to</strong> fonts which were<br />
no<strong>to</strong>nlytypographicallycompatible,butalsospecifically<br />
released under the GNU GPL. In some cases, where<br />
546 <strong>TUG</strong>boat,Volume24 (2003), No. 3<strong>—</strong>Proceedingsof EuroTEX2003
<strong>The</strong><strong>Free</strong><strong>UCS</strong><strong>Outline</strong><strong>Fonts</strong> <strong>Project</strong><strong>—</strong><strong>an</strong> <strong>Attempt</strong> <strong>to</strong> <strong>Create</strong> aGlobal Font<br />
the fonts were released under less clearly defined terms<br />
(“free”, “public domain” etc.), we contacted the authors<br />
<strong>an</strong>d asked for permission <strong>to</strong> use their work under<br />
the terms of the GNU GPL. Most authors agreed, <strong>an</strong>d<br />
whilenonedisagreed,someoftheemailsremainedun<strong>an</strong>swered.Insuchcaseswewereunable<strong>to</strong>confirmwhether<br />
the recipient received our email at all. <strong>The</strong>se fonts still<br />
wait <strong>to</strong> be included in the free <strong>UCS</strong> outline font collection.<br />
Choice of typefaces. Even though Bigelow <strong>an</strong>d Holmes<br />
claim that their primary reason for choosing a s<strong>an</strong>s-serif<br />
fontisbecauseitcarriesleasthis<strong>to</strong>rical<strong>an</strong>dculturalassociations[5,p.1003],<strong>an</strong>ddespitetheexpressedconcerns<br />
about eurocentrism, we decided <strong>to</strong> develop in parallel<br />
three different families, modelled after Times Rom<strong>an</strong>,<br />
Helvetica <strong>an</strong>d Courier, <strong>an</strong>d specifically derived from<br />
URW++ typefaces Nimbus Rom<strong>an</strong> No. 9, Nimbus S<strong>an</strong>s<br />
<strong>an</strong>dNimbusMono.Reflectingthenatureoftheproject,<br />
wedubbedthemas<strong>Free</strong>Serif,<strong>Free</strong>S<strong>an</strong>s<strong>an</strong>d<strong>Free</strong>Mono.<br />
Aside from covering three different letterforms<strong>—</strong><br />
one monospaced <strong>an</strong>d two proportional, one with modulated<br />
strokes <strong>an</strong>d <strong>an</strong>other with unmodulated strokes<strong>—</strong><br />
the primary reason for choosing these three typefaces<br />
wastheirubiquity.Sincetheyhavebeenextremelypopularform<strong>an</strong>ydecades,<strong>an</strong>dpresentinthedesk<strong>to</strong>ppublishing<br />
world-widefor over twentyyears, they inspired numerouslocaldesignsaroundtheworld,wherenon-Latin<br />
glyphs were designed specifically <strong>to</strong> blend with one of<br />
the above-named typefaces.Whilewe realise that m<strong>an</strong>y<br />
of these designs introduce typographic practice alien <strong>to</strong><br />
a traditional local design, we also recognise that the designsdonebynativespeakersreflectthelocaltypographic<br />
knowledge<strong>an</strong>dits evolvingtypographicrules.<br />
Tools. While we paid special attention <strong>to</strong> the licensing<br />
of the used sources, no particular attention was given <strong>to</strong><br />
the licensing issues of the <strong>to</strong>ols used, as the font <strong>to</strong>ols<br />
generallydon’t implythe licensesunderwhichthefonts<br />
are released. Still, we ended up using exclusively opensourced<br />
<strong>to</strong>ols. <strong>The</strong> choice was clearly influenced by the<br />
fact that most particip<strong>an</strong>ts of the project use Linux, <strong>an</strong>d<br />
while none of the numerous popular font edi<strong>to</strong>rs from<br />
other platforms have been ported <strong>to</strong> this platform, we<br />
havesome excellentfree<strong>to</strong>ols available.<br />
PfaEdit Without George Williams’s excellent font edi<strong>to</strong>rPfaEdit,<br />
3 thisprojectwouldnothavebeenpossibleatall.PfaEditisavisualfontedi<strong>to</strong>rthatallows<br />
copying<strong>an</strong>dmodifyingglyphs<strong>an</strong>dmos<strong>to</strong>therthings<br />
expected from a font edi<strong>to</strong>r, reads <strong>an</strong>d writes most<br />
popular outline font formats, <strong>an</strong>d has proved <strong>to</strong> be<br />
stable enough <strong>to</strong> allow working on large glyph sets.<br />
Eventhoughtheprogramwasmatureenough<strong>to</strong>be<br />
3. Since renamed <strong>to</strong> FontForge:<br />
http://fontforge.sourceforge.net/<br />
usedasa<strong>to</strong>olbackinlate2001,thepaceofitsdevelopment<br />
has not diminished over the past year, <strong>an</strong>d<br />
itsauthor isveryresponsive<strong>to</strong> bugreports.<br />
ChoosingPfaEditasourmain<strong>to</strong>olalsoinfluenced<br />
the decision of the native font format used in the<br />
free <strong>UCS</strong> outline fonts project. Since we use a CVS<br />
reposi<strong>to</strong>ryforthebookkeepingoffontadditions<strong>an</strong>d<br />
modifications, we were looking for a format suitable<br />
<strong>to</strong> differential ch<strong>an</strong>ges. A format was considered<br />
suitable if local ch<strong>an</strong>ges in typeface design remained<br />
localised in the font file rather th<strong>an</strong> propagating<br />
themselves across the file. This requirement<br />
ruled out compiled binary formats, as well as Post-<br />
ScriptType1eexecencoding.PfaEdit’snativeformatSFD,similar<strong>to</strong>Adobe’sBDFformatinitsstructure,<br />
proved <strong>to</strong> be a suitable choice. <strong>The</strong> fonts are<br />
thusarchivedintheSFDformat,<strong>an</strong>dPfaEditisused<br />
<strong>to</strong> create fonts inother outlineformats.<br />
PfaEdit’s only major drawback may be its weak<br />
supportforTrueTypeinstructionsor“hints”.Since<br />
TrueType is not the native format used by the<br />
free <strong>UCS</strong>outlinefontproject,wearelookingeither<br />
for au<strong>to</strong>mated generation of TrueType hints, or<br />
forseparatefileswithTrueTypeinstructionswhich<br />
couldbecompiled<strong>to</strong>getherwiththeglyphoutlines<br />
in the SFD format <strong>to</strong> produce a hinted TrueType<br />
font.Anysolutionrequiringm<strong>an</strong>ualmodificationof<br />
binary TrueType fonts is not acceptable, as True-<br />
Typefontsarecreatedau<strong>to</strong>maticallyeachtimefrom<br />
theSFDsources.WhilePfaEditwillprobablyeventuallyproduceacceptableTrueTypehintsin<strong>an</strong>au<strong>to</strong>mated<br />
way, other possibilities have <strong>to</strong> be investigatedintheme<strong>an</strong>time.<br />
TrueType <strong>to</strong>ols Rogier v<strong>an</strong> Dalen’s TrueType <strong>to</strong>ols 4<br />
mightbethepropersolutionfortheTrueTypehinting<br />
problem. V<strong>an</strong> Dalen’s approach uses a separate<br />
file with TrueType instructions written in a highlevell<strong>an</strong>guagenotunlikeC,whichc<strong>an</strong>becompiled<br />
<strong>to</strong>gether with <strong>an</strong> unhinted TrueType font file <strong>to</strong><br />
produce a hinted TrueType font file. If this approach<br />
is adopted, each TrueType font file will be<br />
au<strong>to</strong>matically produced from two source files: the<br />
SFD file containing the glyph outlines <strong>an</strong>d the TTI<br />
file containing the TrueType instructions for hinting.<br />
At the moment of this writing, the main concern<br />
is the estimated amount of work needed for<br />
producing TrueType hintinginstructions in a m<strong>an</strong>ual<br />
way.<br />
TTX While the SFD format fits present needs just fine,<br />
the future might require distributing font sources<br />
in some other open format. For this purpose we<br />
4. http://home.kabelfoon.nl/~slam/fonts/<br />
<strong>TUG</strong>boat,Volume24 (2003), No.3<strong>—</strong>Proceedingsof EuroTEX2003 547
PrimožPeterlin<br />
are considering TTX, the TrueType <strong>to</strong> XML converter.<br />
5 Given the pace at whichPfaEdit currently<br />
evolves, it may well be that PfaEdit itself will support<br />
XML before we will actually feel the need<br />
for it.<br />
Towards ISO 10646/Unicode compli<strong>an</strong>ce. As of version<br />
3.2, Unicode defines 95156 encoded characters. Takingin<strong>to</strong>accountthreefamilies(<strong>Free</strong>Serif,<strong>Free</strong>S<strong>an</strong>s<strong>an</strong>d<br />
<strong>Free</strong> Mono), two weights (normal <strong>an</strong>d bold) <strong>an</strong>d two<br />
shapes (regular <strong>an</strong>d italic/oblique), this me<strong>an</strong>s designing<br />
almost 1.2 million glyphs for the free <strong>UCS</strong> font project.<br />
Beingavolunteerproject,free <strong>UCS</strong>outlinefontsproject<br />
growsmainlyin<strong>an</strong>on-systematicway<strong>—</strong>whenasuitable<br />
set of glyphs is found or donated, we add it <strong>to</strong> the font.<br />
Often, a wider community c<strong>an</strong> benefit from scratching<br />
one’sown itch,e.g.creatingaset of APL glyphs[7, 8].<br />
Still, we felt the need <strong>to</strong> add the glyphs in a more<br />
systematic way in parallel with the spont<strong>an</strong>eous growth<br />
oftheproject.Forthatpurpose,weusethemultilingual<br />
Europe<strong>an</strong>subsetsasdefinedbyComitéEuropéndeNormalisation<br />
[6] as a guideline:<br />
• MES-1, a Latinreper<strong>to</strong>ire basedon ISO/IEC<br />
6937:1994 (335 characters)<br />
• MES-2, a Latin, Cyrillic, <strong>an</strong>d Greek reper<strong>to</strong>ire<br />
based on ENV1973:1996 (1062 characters)<br />
• MES-3, a reper<strong>to</strong>ire needed <strong>to</strong> write all the l<strong>an</strong>guages<br />
of Europe <strong>an</strong>d tr<strong>an</strong>sliterate between them.<br />
MES-3A is a script-based, non-fixed collection,<br />
while MES-3Bisafixedsubse<strong>to</strong>f 2819 characters.<br />
Character-<strong>to</strong>-glyphtr<strong>an</strong>slation. As the ISO 10646/Unicode<br />
st<strong>an</strong>dard encodes characters rather th<strong>an</strong> glyphs, it<br />
is clear that for acceptable rendering of m<strong>an</strong>y l<strong>an</strong>guages<br />
more glyphs th<strong>an</strong> those corresponding <strong>to</strong> Unicode<br />
charactersareneeded.Thisisparticularlyimport<strong>an</strong>tfor<br />
rendering Indic l<strong>an</strong>guages <strong>an</strong>d Syriac which contain numerousligaturesnotencodedintheISO10646/Unicode<br />
st<strong>an</strong>dard.<br />
During the 1990’s, several initiatives for creating<br />
“smartfonts”werestarted,i.e.,fontsbeingno<strong>to</strong>nlycontainersofglyphsbutalsoincorporatingsomelogicalrules<br />
forglyphsubstitution,glyphposition,contextualrendering,etc.AmongthosewereAppleAdv<strong>an</strong>cedTypography<br />
[2], SILGraphite[18]<strong>an</strong>dAdobe/MicrosoftOpenType<br />
[17].Fromthesethree,OpenTypeseems<strong>to</strong>begrabbing<br />
thelargest market share.<br />
Results<br />
<strong>The</strong> project was conceived in December 2001 <strong>an</strong>d was<br />
approvedforhostingontheSav<strong>an</strong>nahWebsite 6 inFebruary<br />
2002. As of March 2003, a <strong>to</strong>tal of 17794 charac-<br />
5. http://www.letterror.com/code/ttx/<br />
6. http://sav<strong>an</strong>nah.gnu.org/projects/freefont<br />
ters are encoded. We are just a couple of dozen glyphs<br />
short of MES-1 compli<strong>an</strong>ce in <strong>Free</strong> Serif <strong>an</strong>d <strong>Free</strong> S<strong>an</strong>s,<br />
while <strong>Free</strong> Mono is already MES-1 compli<strong>an</strong>t, <strong>an</strong>d approximately<br />
3500 glyphs short of MES-2 compli<strong>an</strong>ce.<br />
Kerning. Aside from the character set compli<strong>an</strong>ce <strong>an</strong>d<br />
thealreadymentionedTrueTypehintingproblem,there<br />
remain some other tasks. One of them is kerning, which<br />
iscurrentlypresen<strong>to</strong>nlyintheLatinportionsofthefont.<br />
Kerning non-Latin scripts requires typographic knowledgewhichwecurrentlydonotpossess.Webelievethat<br />
the open nature of the project will eventually attract<br />
somebodywiththenecessaryexpertisewhoiswilling<strong>to</strong><br />
contribute.<br />
Discussion<br />
<strong>The</strong> open-source development model has not been previouslytestedforfontdevelopment.Whileitmaybearguablewhethersuchadevelopmentmodelc<strong>an</strong>beapplied<br />
at all <strong>to</strong> creation of works of art, it is also worth noting<br />
that most contribu<strong>to</strong>rs haveatechnical <strong>an</strong>d scientific<br />
background, <strong>an</strong>d probably none of them perceives himor<br />
herselfas <strong>an</strong> artist.<br />
While there is no doubt that a skilled typographer<br />
wouldbeable<strong>to</strong>designamulti-scriptfontvastlyexceeding<br />
what we have done in typographic beauty <strong>an</strong>d consistency,<br />
we must not overlook the fact that designing<br />
1200000 glyphs is probably beyond the capability of a<br />
singleperson,evenifweneglectthefin<strong>an</strong>cialpar<strong>to</strong>fsuch<br />
<strong>an</strong> endeavour.<br />
<strong>Free</strong> <strong>UCS</strong> outlinefonts project<strong>an</strong>dTEX. So far, the <strong>Free</strong><br />
<strong>UCS</strong> outline fonts project has taken from the TEX community<br />
more th<strong>an</strong> it has repaid. Th<strong>an</strong>ks <strong>to</strong> Péter Szabó<br />
<strong>an</strong>dhisTEXtraceprogram(http://www.inf.bme.hu/<br />
~pts/textrace/) [19], it is easy <strong>to</strong> tr<strong>an</strong>sform Metafonts<br />
in<strong>to</strong> PostScript Type 1 fonts which c<strong>an</strong> be edited<br />
by PfaEdit. Karel Piška demonstrated the technique for<br />
Indi<strong>an</strong> Metafonts during the <strong>TUG</strong> 2002 conference in<br />
Thiruv<strong>an</strong><strong>an</strong>thapuram, India.<br />
Even though the free <strong>UCS</strong> outline fonts were designedprimarilyforscreenuse,wec<strong>an</strong>certainlyraisethe<br />
questionofusingthefontsfortypesettingwithTEXor Ω.<br />
As of March 2003, this would require extensive work,<br />
suchassplittingthefontsin<strong>to</strong>smallerfontscontainingno<br />
more th<strong>an</strong> 256 glyphs. However, the current development[4]promisesthattheremightbeamoredirectway<br />
of usingOpenTypefonts with Ωin thefuture.<br />
Acknowledgments<br />
I would like <strong>to</strong> th<strong>an</strong>k the following authors who contributed<br />
<strong>to</strong> the project by allowing their work <strong>to</strong> be includedinthefree<strong>UCS</strong>outlinefontsproject.Inalphabetical<br />
order: Berh<strong>an</strong>u Beyene, D<strong>an</strong>iel Shurovich Chirkov,<br />
548 <strong>TUG</strong>boat,Volume24 (2003), No. 3<strong>—</strong>Proceedingsof EuroTEX2003
<strong>The</strong><strong>Free</strong><strong>UCS</strong><strong>Outline</strong><strong>Fonts</strong> <strong>Project</strong><strong>—</strong><strong>an</strong> <strong>Attempt</strong> <strong>to</strong> <strong>Create</strong> aGlobal Font<br />
PrasadA.Chodavarapu,VyacheslavDikonov,DMSElectronics,ValekFilippov,ShaheedR.Haque,Y<strong>an</strong>nisHaralambous,<br />
Angelo Haritsis, Jeroen Hellingm<strong>an</strong>, Maxim<br />
Iorsh, Mohamed Ish<strong>an</strong>, M<strong>an</strong>fred Kudlek, Harsh Kumar,<br />
Sush<strong>an</strong>t Kumar Dash, Olaf Kummer, Noah Levitt,<br />
Jochen Metzinger, Anshum<strong>an</strong> P<strong>an</strong>dey, Hardip Singh<br />
P<strong>an</strong>nu,ThomasRidgeway,YoungU.Ryu,VirachSornlertlamv<strong>an</strong>ich,<br />
M.S. Sridhar, Sam Step<strong>an</strong>y<strong>an</strong>, URW++<br />
Design&DevelopmentGmbH, Fr<strong>an</strong>s Velthuis,<strong>an</strong>d the<br />
Wadalab K<strong>an</strong>jiCommittee.<br />
References<br />
[1] Adobe Systems Incorporated. Adobe Type 1 Font<br />
Format. Addison-Wesley,1990. Also available on<br />
the Web: http://partners.adobe.com/asn/<br />
developer/PDFS/TN/T1_SPEC.PDF.<br />
[2] Apple Computer, Inc. Apple adv<strong>an</strong>ced typography.<br />
Available on the Web: http://developer.<br />
apple.com/fonts/, 1994.<br />
[3] Apple Computer, Inc. TrueType reference m<strong>an</strong>ual.<br />
Available on the Web: http://developer.<br />
apple.com/fonts/TTRefM<strong>an</strong>/, 1999.<br />
[4] Gábor Bella, Anish Mehta, <strong>an</strong>d Y<strong>an</strong>nis Haralambous.<br />
AdaptingOmega<strong>to</strong>OpenTypefonts. InEuroTEX2003(thisvolume),2003.<br />
[5] Charles Bigelow <strong>an</strong>d Kris Holmes. <strong>The</strong> design of<br />
a Unicode font. Electronic Publishing, 6(3):999–<br />
1015, 1993.<br />
[6] CEN. Information technology – multilingual<br />
Europe<strong>an</strong> subsets in ISO/IEC 10646-1.<br />
Technical Report CWA 13873:2000, Comité<br />
Européen de Normalisation, Technical Committee<br />
304, 2000. Also available on the Web:<br />
http://www.evertype.com/st<strong>an</strong>dards/<br />
iso10646/pdf/cwa13873.pdf.<br />
[7] Phil Chastney. An APL Unicode font. Vec<strong>to</strong>r,<br />
16(1):75–85, 1999.<br />
[8] PhilChastney. APLUnicodefont–extended. Vec<strong>to</strong>r,17(3):139–142,<br />
2001.<br />
[9] Rudi W. de L<strong>an</strong>ge, Hendry L. Esterhuizen, <strong>an</strong>d<br />
Derek Beatty. Perform<strong>an</strong>ce differences between<br />
Times <strong>an</strong>d Helvetica in a reading task. Electronic<br />
Publishing, 6(3):241–248, 1993.<br />
[10] <strong>Free</strong> Software Foundation. GNU General Public<br />
License, version 2. Available on the Web: http:<br />
//www.gnu.org/copyleft/gpl.html, 1991.<br />
[11] Y<strong>an</strong>nis Haralambous. From Unicode <strong>to</strong> typography,<br />
a case study: the Greek script. In Proceedingsof14thInternationalUnicodeConference,pages<br />
b.10.1–b.10.36, Bos<strong>to</strong>n, 1999.<br />
[12] Y<strong>an</strong>nisHaralambous<strong>an</strong>dJohnPlaice. Ω,aTEXextension<br />
including Unicode <strong>an</strong>d featuring Lex-like<br />
filtering processes. In EuroTEXProceedings, pages<br />
154–167, Gdańsk,1994.<br />
[13] GeorgesJe<strong>an</strong>. L’écriture,mémoiredeshommes. Gallimard,Paris,1987.<br />
[14] James Kass. Code2000. Available on the Web:<br />
http://home.att.net/~jameskass/, 1998.<br />
[15] OleLund.Knowledgeconstructionintypography:the<br />
caseoflegibilityresearch<strong>an</strong>dthelegibilityofs<strong>an</strong>sserif<br />
typefaces. PhD thesis, <strong>The</strong> University of Reading,<br />
Departmen<strong>to</strong>fTypography&GraphicCommunication,<br />
1999.<br />
[16] MicrosoftCorporation. ArialUnicode MS. Available<br />
on the Web: http://office.microsoft.<br />
com/downloads/2000/aruniupd.aspx, 2000.<br />
[17] Microsoft Corporation. OpenType specification.<br />
Available on the Web: http://www.microsoft.<br />
com/typography/otspec/, 2001.<br />
[18] SILInternational.Graphite.AvailableontheWeb:<br />
http://graphite.sil.org/, 1997.<br />
[19] PéterSzabó. ConversionofTEXfontsin<strong>to</strong>Type1<br />
format. InProceedingsofEuroTEX2001conference,<br />
Kerkade,September24–27,pages192–206,2001.<br />
[20] Miles A. Tinker. Legibility of Print. Iowa State<br />
UniversityPress,Ames,1963.<br />
[21] <strong>The</strong> Unicode Consortium. <strong>The</strong> Unicode St<strong>an</strong>dardVersion3.0.<br />
Addison-Wesley,Reading,Massachusetts,<br />
2000.<br />
[22] Bror Zachrisson. Studiesinthelegibility ofprinted<br />
text. Almqvist&Wiksell,S<strong>to</strong>ckholm, 1965.<br />
<strong>TUG</strong>boat,Volume24 (2003), No.3<strong>—</strong>Proceedingsof EuroTEX2003 549