From ybf2u@curry.edschool.virginia.edu Tue Jan 24 09:12:17 1995 Received: from uvaarpa.Virginia.EDU by apollo.aoe.vt.edu via SMTP (920330.SGI/911001.SGI) for yacob id AA29466; Tue, 24 Jan 95 09:12:17 -0500 Received: from curry.edschool.virginia.edu by uvaarpa.virginia.edu id aa12188; 24 Jan 95 9:12 EST Received: (from ybf2u@localhost) by curry.edschool.Virginia.EDU (8.6.8/8.6.6) id JAA60344; Tue, 24 Jan 1995 09:12:15 -0500 From: Yitna Benyam Firdyiwek Message-Id: <199501241412.JAA60344@curry.edschool.Virginia.EDU> Subject: Re: Sorting To: Daniel Yacob Date: Tue, 24 Jan 95 9:12:14 EST Cc: fisseha@cig.mot.com In-Reply-To: <9501210404.AA23712@apollo.aoe.vt.edu>; from "Daniel Yacob" at Jan 20, 95 11:04 pm X-Mailer: ELM [version 2.3.1 PL11] Status: R selam dan'El, Yonas, danEl, I'm glad to hear sorting works, or that it's a straightforward operation in MULE. The problem of placement for diqalawoc at the bottom of the table (i.e., usage of the 8 row system--aka 'desta table') would probably not be a concern that would interest the UNICODE Consortium. But it is good for us to note for ourselves. I have looked at your questions, Yonas, and I'm sure you're right in saying that programers can get around the sorting problme easily. But I think we should minimizing any further work that may need to be done in the future if we can--I guess for simplicity and elegance. Now, if there are more compelling reasons to keep the 12 column table, we should discuss those, too. wendmachu, yTna According to Daniel Yacob: > > > selam yTna, > > I just remembered your question about sorting and tried it on Mule > and yes! it does sorting in fidel lexicographical ordering. With > mixed numbers and letters and punctuation, no problem. I am sure that > it is based on the character addresses. Which means that it is > very important where ethiopic punctuation and numbers go. Sorting > will be A LOT easier for programers if the address table is already > in sorted order (since programers and computers work with characters > as numbers). > > THIS ALSO MEANS that the extended diqalawoc absolutely should not > come at the end of the fidel -what a headache it would be for this > issue! Programers would forever be scratching their heads asking > "why.. why..." was this oversight made as they do such much more work > to code sorting routines by the iso fidel table of '93. > > I think this is a strong point to argue, something abass could over- > look since he is using LISP which (by design) gets around the issue. > > all for now, > > wendmh dan'El > > > >