L2/01-072 From: Michael Everson [everson@egt.ie] Sent: Saturday, January 27, 2001 8:51 AM Subject: Georgian Text & Case Mappings, Reduxed Ar 09:57 -0800 2001-01-26, scríobh Becker, Joseph: >as lowercase (Mkhedruli or Nuskhuri). In lowercase (i.e. normal caseless) >Georgian text, Mkhedruli or Nuskhuri are distinguished via font, >as are regular and italic forms in Latin lowercase. similar to the way in which roman and italic forms are in Latin lowercase. >Other Punctuation. For the Georgian full stop, use U+0589 ARMENIAN FULL >STOP. Most modern Georgian uses the ASCII FULL STOP. >------------------ >>Case Forms. The Mkhedruli alphabet is fundamentally caseless, and is used >>as such in most texts. However, possibly owing to the influence of case >>forms in other alphabets, modern Georgian has seen occasional experiments >>of writing with uppercase capital letters -- usually in linguistic works >>dealing with the history of the script, early saints' lives, etc. In this >>typographic departure, it is the Asomtavruli forms that serve to represent >>uppercase letters. Khutsuri texts may be found in Asomtavruli alone, in >>Nuskhuri alone, or with the two as casing pairs in which case the >>Asomtavruli represents the capital letters. This usage is quite parallel to >>the evolution of the Latin alphabet, in which the original linear >>monumental style came to be considered uppercase, while document styles of >>the same alphabet came to be represented as lowercase. > >My preference in these things is to keep the main plotline clear for the >vast majority of users of the book, namely English-speaking-if-that >programmers who just want to understand what they have to do to get through >implementing the script. I find your additions here, however true, to >clutter my ability to read through the paragraph and see what it implies for >the casing model. Maybe a point or two could be tweaked, but at the moment >I think it's OK as it was. I think tweaking is in order; sorry but I don't much like the text as it stands. The use of case forms in Mxedruli is 20th century and was experimental. >The UTC's charge was precisely to make our proposed model unambiguous at >last. This resolution may yet be rejected, by the UTC or by world opinion, >but the one thing we want to eliminate here is having the proposed solution >remain unclearly specified. Understood. >>>In those rare instances where >>>software is dealing with Georgian text that actually uses the Asomtavruli >>>forms as uppercase letters and case folding or case mapping is required, >>>this should be done via special casing rules. > >>dealing with Mkhedruli text > >Well, it's saying that (in this-here model) *all* Georgian casing has to be >done specially, so restricting the statement to Mkhedruli is not >appropriate. True, given the status quo (no separate Nusxuri). >------------------ >OK, I believe at this point we transition from editing the to-be-published >3.1 text to just normal backing-and-forthing. > >Let me skip down to the punchline: > >>The reality of Georgian is that you've got a bicameral script Khutsuri >>which is extremely analogous to Latin and a unicameral script Mkhedruli >>which some weirdos stuck Asomtavruli onto sometimes. > >I would amend this slightly, in that our Georgian customers at Voice of >America were familiar with the use of Asomtavruli caps, as is the only >Georgian website I could find on the subject: > > http://redrival.com/giasher/language.htm#Writing > >"Contemporary Georgian Typography use capitals in special cases (Headers, >Captions, Posters, Advertising, etc.)" So, it sounds like a not-too-weird >usage in display contexts. No, no, no, no, no. That's not what Gia Shervashidze (the author of that document) is talking about. Please go back to my paper N1962. On page 4 you see a discussion of the (true) mtavruli, which is analogous to small caps. Note that these are NOT capital letters, and when mtavruli is used ALL the letters in the word are ALWAYS mtavrulized, i.e. there is no concept of title case. Nuskhuri and Asomtavruli do not "undergo" mtavruli. Note, by the way, what Shervashidze has said about the three scripts and especially the derivation of Mxedruli from Asomtavruli directly. (Just as I say on page 1 of N1962.) ========== Alphabet(s) SOURCES Inscriptions: - they are saved on Stones and Church Walls (carved), Mosaics, Frescos, Goldsmith etc. Manuscripts: - (more than 10,000) Eclesiastic, Phylosophy, Literature, etc. Documents: - (about 70,000) Epistols, Juridical documents, Lists, etc. TYPES During Georgian history 3 types of alphabets are functioned: Asomtavruli, Nuskhuri and Mkhedruli. Asomtavruli is oldest one but in Medieval all three types were in use simultaneously. Asomtavruli: First existing sources are from Vth AD (sombody says IVth), it used mainly for Inscriptions. It has very strong geometrical basis and all letters considered as capitals. It consists 38 letters (5 of them abolished in contemporary Mkhedruli). Nuskhuri: First existing source 853 AD. It used mainly for manuscripts and closely related to church and church reforms in medieval Georgia. Don't know why, but I it is least attractive Alphabet for me. Think - reason is that it may be was create for hand writing from Asomtavruli without any logic related with handwriting specifics. Mkhedruli: It seems it was used from begin of Xth century and was directly created from Asomtavruli for civil use. I like it. Till now it changed very little. It contains also 38 characters (5 of them were dropped in 1860 by Ilia Chavchavadze). I don't want even talk about 'Fi' - it isn't Georgian letter. ========== >Given that adjustment, your statement expresses a kind of particle-wave >duality that is enlightening, but which, so far as I can yet see, we can't >actually encode any better than we already have. So I'd be satisfied if we >can *clarify* the model better, looking at it from the modern end because >that's where most of the encoded text will be. I agree; we can't do better until we encode Nuskhuri.... >>The thing is, lads, that casing operations between Nuskhuri and Asomtavruli >>**ARE** normal and work the same way you would expect for Latin. > >Eh?! I don't know if the question is even meaningful, given the limited >scope of usage of Nuskhuri, but are you saying that it is *normally written >with Asomtavruli capitals?!* That is not what my 17th and most recent >understanding had been. I don't know if "normally" means "usually" but it is certainly not a rare thing, apparantly. Samples in N1962 on pp. 10-11. Perhaps you would like some quantativeness to the statement, but the point is, that the funny stuff in the 1970s with Asomtavruli/Mxedruli was anomalous, while the situation with Medieval Asomtavruli/Nusxuri was much more like the situation with Latin. >> ... So now you >>have this situation where you'd want casing operations in some FONTS but >>not in others, and it's all madness till Nuskhuri is encoded. > >I think you'd want casing operations in some SITUATIONS but not in others >... not in *most/generic* situations, as the man said. If that's true, then >the current model is the best possible, which is not to say that it does the >specialcasing for you. I *think* that if we had Asomtavruli and Nuskhuri both it would be OK to casefold them. But I think this would not be what one would want to do with Mxedruli. Which as a practical matter (the 99.9% situation) is what we have done by not casefolding the two halves of the Georgian table... >As I said, I never got any closer to Georgia than a few miles, but I wish >the casing model to focus on what I understand to be the 99.9% situation, >namely plain old caseless Mkhedruli text. If we can't come any closer but >that other situations require other measures, be it special casing or fonts >or option selection by the user (as casing in *English* always does!), so be >it, c'est la vie. For ordering, though, we are mingling the two halves of the table, right, Ken? Michael Everson ** Everson Gunn Teoranta ** http://www.egt.ie 15 Port Chaeimhghein Íochtarach; Baile Átha Cliath 2; Éire/Ireland Mob +353 86 807 9169 ** Fax +353 1 478 2597 ** Vox +353 1 478 2597 27 Páirc an Fhéithlinn; Baile an Bhóthair; Co. Átha Cliath; Éire