encoding problem

zazie243

My Library was working just fine until the end of last semester (may) but I went to check it again and see that the encoding for japanese titles is all messed up. Where might I go to adjust those settings - it may be as easy as making sure to use unicode, or not.... The settings don't seem to have an option for this. Advice?

dstillman

The data in Zotero is always stored and displayed as Unicode and doesn't change. Did you export your library in some format and reimport it? Are you using your library on a different computer without the necessary fonts?

zazie243

Thanks, Dan.
The problem appears to be with only some entries, suggesting that there was an import problem. Other entries display just fine. I'll experiment with the import settings to make sure everything is unicode.
Cheers,

zazie243

Encoding problem, part two: the plot thickens
OK, so I have checked first to make sure all necessary fonts are loaded on my machine. Here's the issue: the non-Roman script in WorldCat is garbled in Zotero. Try to import the same record twice (once in Roman letters, once in non-Roman characters) and you should get the same problem.
Example: search for Kotengeki to zeneigeki : Furansu to Nihon (OCLC: 25549028) and import it to zotero in Roman characters. No problem!
Next, click on the Show non-Roman characters option in Language.
Import it to zotero. Problem! The information comes into zotero as garbage. Any suggestions?

fbennett

I don't see a "Show non-Roman characters" option. Where should I be looking?

(Might this be in an OCLC FirstSearch view of WorldCat, rather than Open WorldCat?)

adamsmith

it sounds like this is WorldCat through OCLC First Search?

zazie243

Yes, you both are correct. I was using the Firstsearch view.
Thank for looking,

adamsmith

finally figured this out. A new version of the oclc translator is now out - it will update itself in your copy of Zotero within 24hs or you can manually update form the general tab of the Zotero preferences.

(For those interested in the tech details - the oclc translator has iso8859 hard-coded into it, because OCLC usually encodes in ISO and if you try treating that as utf-8, accented letters get mangled. But when you click to show non-roman characters, the catalog switches to utf-8. The new translator should now be able to detect that switch and also switch to utf-8, but we're not sure how stable that is, so let us know if you run into problems (if you do, switching the "display non-roman characters back and forth may help)).