[Wikitech-l] Re: Back to IPA/SAMPA and pronunciations

The last discussions on the lists about how to represent pronunciations on Wikipedia didn't end with any definitive conclusions. One of the proposed suggestions was to provide a way input IPA in an easy-to-use format and then have it automagically get converted into Unicode IPA. As a first step I wrote a Lex analyzer to convert X-SAMPA to Unicode IPA (which results in C code that I compiled to an executable). This will enable editors to enter pronunciations in the ugly-but-easy-to-type X-SAMPA format and have them appear in IPA using the Unicode IPA extensions. I also dove into the Wikipedia code and patched OutputPage.php to support <IPA> </IPA> tags that surround SAMPA and output it as <SPAN class="IPA"> </SPAN>, with the Unicode IPA HTML entities inside the SPAN tag. Finally, I patched style/wikistandard.css to have a .IPA style that explicitly sets the font-family to a list of fonts that are known to contain the IPA Unicode extensions. This is necessary because some broswers (namely Windows IE) don't display IPA Unicode characters even if they're installed unless the currently active font has those characters. I envision the use of <IPA> tags on Wikipedia to be fairly limited, in that IPA/SAMPA will only appear on pages discussing pronunciations, but I think it will make a good starting point, perhaps, for representing pronunciations on Wiktionary. I can attach the diffs for OutputPage.php and style/wikistandard.css to a future message (if I can ever get the sourceforge cvs server to respond), but what should I do with my .lex file and Makefile for building the parser? Should I post them too? I would guess most list readers would not be pleased by my spamming them with such tediums. I understand the hesitation developers have to handing out CVS access to people whose code they have never seen, so email me privately if you want me to send you what I've done. Thanks! - David [[User:Nohat]] Note: for most purposes, X-SAMPA is backwards compatible with SAMPA for various languages, but the Lex analyzer can be modified to support something like <IPA lang="French"> or something like that for the language-specific SAMPA encodings, if that is deemed desirable.