Bug in syncing: abstract too big

pfudd
Hi..

Report id: 750398029

Description of problem:
I'm just trying to sync for the first time, and the sync fails with an error about an abstract being too large. When I look at the abstract in Zotero, it is indeed a very large abstract, and I know that I didn't type it. The actual abstract from the article is only two sentences; apparently the site translator grabbed a big chunk of the article too. Perhaps the translator was written and tested by a non-subscriber to that website?

Possible solutions:
Fix translator's abstract finding ability (hard, in this world of constant change)
Make Zotero limits the same as Sync limits
Make Sync limits the same as Zotero limits
Check Zotero database for consistency with current constraints periodically
Silently truncate fields that are too large (not a great solution)

How to reproduce:
The journal article is doi:10.1016/S0092-8674(04)00045-5 ; it was added to Zotero March 19, 2009, with whatever version was available at the time.
  • dstillman
    This isn't a syncing bug. The server has limits and the client (for the most part) doesn't, and both are by design. The server limit isn't relevant until you sync, which is why that's when you see the error.

    The bug is in the translator pulling in bad data. If it's still doing so, it should be fixed.
  • pfudd
    I guess what I'd like to see is something other than an error message, since this isn't an error as much as it's an expected mismatch of design limits. It would come up if a user copied and pasted (or typed) an excessively large abstract, but I agree the translator should still be fixed if it's not already.

    I just am not keen on the idea of a perfectly valid zotero database that is not syncable: everything should 'just work', all the time.

    Thanks!
  • dstillman
    edited July 14, 2010
    When tags are too long at sync time a wizard comes up that helps you deal with it—split, edit, delete. Ideally other fields would have the same, but we haven't gotten to that yet.

    The client always errs on the side of preserving data that comes in, partly because there are a lot of different ways to get data into a Zotero database, and it's not always appropriate to pop up a notification and interrupt the user. Syncing is when the limit differences actually matter.

    Again, though—if you can reproduce the problem now, then let us know and we can create a ticket to fix the translator.
  • pfudd
    There's a zillion different ways to get the reference, but only pubmed fills in the abstract; Cell put the abstract into a note. Both gave the short abstract. I guess the bug is fixed.
Sign In or Register to comment.