Recommend RIS for EndNote® importing

[dstillman]

Now that the RIS translator has better charset support (with r3066) and a fix for EndNote®'s keyword bug, and we're not going to work around its BibTeX keyword bug (#1076), we should recommend RIS as the import format to use.

Changes won't go in until 1.0.8, though, so we shouldn't change the documentation until then, since the BibTeX translator currently has better charset support.

We should also do some testing of RIS importing on the branch (using Reset Translators and Styles to get the latest RIS translator) before then, particularly with Unicode characters.

[stakats]

The one hitch here is that their default RIS export filter does not output full author names, just first initials and last names. Not sure if there are other deficiences. In any case, I suppose we could supply our own "full" RIS export filter for use.

[simon]

Are you sure the BibTeX translator currently has better charset support for files exported from EndNote®? IIRC, EndNote® exports UTF-8 with a BOM, which the older code should have handled properly.

[dstillman]

The older RIS code? I was under the impression that it defaulted to IBM850 regardless. Was that not the case?

[dstillman]

OK, just found the comment in translate.js:

// found a UTF BOM at the beginning of the file; don't allow
// translator to set the character set

[dstillman]

The question, then, is if we think users would have trouble installing a custom RIS export filter that we distributed, or whether it's worth trying to fix #1076.

It looks like, until the very latest version (X2), you had to manually copy an .enf file into the correct directory on your system. This isn't very hard for something you do once, but it doesn't help much with interchange with any arbitrary EndNote® installation (without the custom filter installed).

Do we know of any other differences in metadata quality from EndNote® between RIS and BibTeX?

[ajlyon]

We do this already: ​http://www.zotero.org/support/kb/importing_records_from_endnote