From: Greg K. <gr...@ma...> - 2004-08-11 21:20:07
|
On Wed, Aug 11, 2004 at 03:15:24PM -0500, Alex Roitman wrote: > > > I think we may have an option "Attempt to strip non-ascii information" > > > in a plain text format. If the user choses so, the characters would be > > > deaccented. We may even have another option "Use TeX commands > > > for non-ascii" to go with \l instead of slashed ell and the like. > > > > Yes! That's all that I've been asking for. Like you, I don't > > mean to just be arguing; maybe we don't really disagree after all. > > Alright! ... > To summarize, my opinion so far is that we may attempt limited lossy > conversion from unicode text using latin scripts into ascii/iso-8859-1 > as a user-adjustable option. I would not extend this into anything else > than the plain text reports though :-) Let's start with that then. I've never even liked iso-8859-1 all that much. I would be just as happy if the only down-conversion that you supported was Unicode to iso-646 = 7-bit ASCII. In any case it's your project and my only direct request as a user is text reports in the iso-646 character set. Let me just note that you can kill four birds with one stone: You can use almost the same functions to generate iso-646 plain text, iso-646 LaTeX, and ANSEL, and adequately alphabetize words in all Latin-derived alphabets. (I don't need LaTeX from GRAMPS at the moment, but if I did, I would ask for iso-646.) -- /\ Greg Kuperberg (UC Davis) / \ \ / Visit the Math ArXiv Front at http://front.math.ucdavis.edu/ \/ * All the math that's fit to e-print * |