I absolutely hate workarounds like using images or substituting "3" for yogh. Like James, I like UTF-8 for web pages. It makes for smaller files/faster downloads than entities, and IE (always the problem child) seems to handle UTF-8 better than it does entities for some reason. But for the "base" (presumably XML) file the considerations are different. Don't want to get into a long discussion, but perhaps the most important is not to require particular fonts or applications. It's hard to say what's best for the long haul, but Murray's solution seems very good to me. I'll just add that the Medieval Unicode Font Initiative (http://helmer.aksis.uib.no/mufi/) suggests entity names for all the Unicode characters it recognizes as useful for medievalists, and also for all the code points is claims in the Private Use Area.
And Murray--if you have questions about whether particular characters are present in the Unicode standard I may be able to help.
Peter
Murray McGillivray wrote:
James Cummings wrote:
So what is best? Obviously encoding your webpages as (say) UTF-8 is a good start. Force user to download a font for your site with appropriate glyphs? Use images of the glyphs instead of actual characters(*shudder*)? Transliterate into ascii characters/editorial marks? Use markup to allow easy replacement of different solutions on the fly?
I have my own preferences but am interested in what other people have done.
Many factors may come into play in this decision, including not just current functionality, but (I would hope for anyone putting much time and energy into a scholarly project) a plan for future accessibility of the data--humanities projects have, or should have, a long time horizon of usabilty, whatever the publication mode.
From the latter point of view as much as or more than the former, and because we are recording a variety of glyphs in our manuscript that are not (currently?) in Unicode, I have taken the tack in the transcription part of my current project of recording every manuscript glyph using an XML entity (here I simplify slightly). It's easy (here I simplify a bit more) to replace these entities on the fly with Unicode characters, transliterations into ASCII with editorial markup, or whatever, and they're reasonably convenient to deal with in writing things like search routines. Make the raw file look like doggy doo-doo, of course:
http://www.gawain-ms.caMurray McGillivray
dm-l mailing list dm-l@uleth.ca http://listserv.uleth.ca/mailman/listinfo/dm-l