From mark@kli.org Fri Sep 17 14:01:37 1999 X-Digest-Num: 236 Message-ID: <44114.236.1284.959273825@eGroups.com> Date: 17 Sep 1999 21:01:37 -0000 From: mark@kli.org Subject: Re: Typesetting Lojban [was: Lojban word processor for Windows?] >From: David Brookshire Conner >Date: Fri, 17 Sep 1999 11:24:38 -0400 (EDT) >Cc: lojban@onelist.com > >From: David Brookshire Conner > >[ I've split this thread into typesetting and text editing ] >mark@kli.org writes: > > >Oddly enough I was thinking about just this problem this morning as I > > >was walking to work. In particular, Loglan's regular structure gives > > >the text editor lots of help in e.g., automatically typesetting (a la > > >LaTeX). > > > > I was thinking along similar lines wrt typography and typesetting for > > Lojban (I've been designing a Latin-character Klingon font, with > > appropriate ligatures, etc). But you can't improve too much on some nice > > cold monospaced Courier-like fonts for Lojban. It fits the language. > >Urgh. Okay, maybe it is just me, but I find monospaced fonts hard to >read. I don't find Lojban to be cold. Rather, I notice its rich >capacity for metaphor, its poetic rhythms, the fluid syntax. It is >certainly *precise*, no doubt about it, but cold? Not to me. But then, >I'm a hacker, so speaking in Prolog doesn't seem too strange :-) I'm a hacker too; "cold" is not an insult to me. Lojban is an astoundingly rich and poetry-capable language to me, but for all that it does still retain the distinction of syntactic unambiguity, which is decidedly a mechanistic trait. Monospace fonts do tend to be ugly, but strangely they do seem to work pretty well for Lojban. Not that there isn't something else that wouldn't work better; I just so far have been satisfied with the typewriter look (wow. That's not supposed to happen in typography. Typewriter font is hideous). >I did an experiment where I developed a mode for using Tengwar >(Tolkien's Elvish font) with Lojban. Surprisingly, it worked amazingly >well (better than it did for Brown's original Loglan). Lojban's >phonemics are so regular and uniform, and the Tengwar maps so readily >to phonemic structures, the two go hand in hand. The Tehtar for the >vowels (various little accents and such, for those not familiar with >the orthography) have a nice side effect in making syntactic structure >*visible* - any word with a tengwa with no tehta ("r" is the plural in >Quenya and Sindarin) must be a selbri (or possibly a name). Nice; obviates the need for the cmene-final period as a reminder. Someone once mailed me, out of the blue, a letter detailing how to write Esperanto using the Hebrew alphabet (actually the Yiddish alphabet, or Hebrew used in the Yiddish fashion). My response was, "Um... OK, that works... but, er, why?" Lojban in tengwar has a more legitimate right to exist, actually, since Lojban is supposed to have a bit more independence from its orthography than Esperanto. >Most of the basic little words end up being a single glyph: one tengwa >with one tehta above, maybe one below. This suggests ligatures for >these combinations, at least in Tengwar. They become pictogram-like >word-characters, yet still have all the phonemic information apparent. You really should have come up with some way to make diphthongs -- both VV and V'V -- into tehta-combinations. That way *all* cmavo would be single symbols. It would make a nice cmavo/brivla distinction. Not sure if it's possible without making a real mess of the notation, though. >Actually, the one open question I'd still muddling about for this mode >is whether to put "r" and "l" in Row 6 of the Tengwar - the >semivowels. This has a certain elegance, but it does make the writing >a bit more monotonous looking. But most tengwar modes don't put r and >l up there, but as part of the "other tengwa". Tengwar by their nature are monotonous-looking when written. Probably the weakest thing, aesthetically and practically, about the system: most of the letters resemble one another. You can't get around that, might as well live with it. There's a certain charm to teeny little Grade 6 r/l, especially given their role as hyphens/glue in Lojban. Have you bounced this off Ivan Derzhanski, Lojbanist, typographer, and tengwarist extraordinaire? > > >On another tack, a lojban font would be an interesting > > >problem. Specifically, the ligatures would probably be different from > > >an English font (as letter frequency is different), and would probably > > >emphasize the cmavo. > >> See above. I gave that some thought, but didn't come up with >> anything. Lojban doesn't feel right with much in the way of "real" >> ligatures (like fi, fl, etc). Typography to make the cmavo more >> distinctive might be nice, > >Certainly. Most of the ones starting with a c seem straight-forward, >as do ones with f, t, d, g, k. That's just off the top of my head for >the cmavo. > >Some of the consonant clusters look amenable, as well. For instance, >an English font wouldn't have a ligature for "cm", as it never (?) >appears in english. In lojban, of course, it's common. For that >matter, in "lojban" itself, "jb" and "an" look feasible, and occur >often enough that you might do it. > >I guess this seems ligature-heavy, but to my thinking, typing the >letters together emphasizes the structure, which is part of the beauty >of lojban. Come to think of it, you may be on to something here. Using ligatures for consonant-clusters emphasizes the clustering, and makes them distinctive. And noticing consonant clusters is critical in Lojban (especially as written with frequent "compound cmavo"). And maybe that's what you just said. Ligating cmavo might dilute that, but not if done carefully. >> but I'm not sure how. Except maybe to emphasize ".i", but then if you >> break lines at every jufra (which I personally don't do in writing most of >> the time) you don't need it. > >Yeah, I wouldn't break at every jufra either. So emphasizing .i gets >us right back to ligatures for little words. Not much ligating you can do with {.i} though. Except for a special font or form for the {i}. >> Though you do want to make sure your font does nice things with the >> apostrophe. Not to much and certainly not too little. Again, >> monospace seems oddly workable. The overly wide space for the >> apostrophe makes sure you can't mistake "ta'i" for "tai". > >Hmmm. True, but the overly long space also suggests "ta' i" or even >"ta i". Using a ligature for 'i (merging the dot over the i and the >apostrophe to make a bent line) would look sufficiently different that >I think it would work well. And you still have tons of white space >beneath the apostrophe. Bear in mind I'm one of those who pushed early on for permitting {h} as an alloglyph of {'}, for handwriting. So I'm in favor of pretty heavy representations of {'}, and generally fearful of losing it. Mmm, I'm torn regarding the idea of merging the apostrophe and the dot. If anything, I want to emphasize the break, not obscure it. But on the other hand, I could see some kind of big bold comma-ish thing ligated on top of an {i} that almost becomes a syllable-glyph for {'i} (the /hi/ syllable). Confusion with {ta' i} or {ta i} doesn't scare me much, since I do NOT intend for the {'} to be lost (if anything I would overemphasize it), and if we're looking to pump up the visual distinctiveness of the {.i} cmavo it will not conflict with the {i} in {ta'i}. All of which also doesn't treat the other four vowels and the need for a certain amount of visual consistency. Don't fear heavy ligatures; wait till I show you my Klingon font... ~mark