On Jan 26, 2007, at 2:47 PM, Simon Spiegel wrote:
>> On Jan 26, 2007, at 12:11 PM, Simon Spiegel wrote:
>>> I think that's a very exciting development and I was wondering
>>> whether BibDesk could maybe profit from the work done by the Zotero
>>> team. Zotero (http://www.zotero.org/) is an open source Firefox
>>> plugin which recognises and parses a big variety of web page
>>> from microformats to Google Print, Amazon and Aleph. If this code
>>> could be integrated into BibDesk it would make this new type of
>>> search group very powerful.
>> Simon, thanks - I read the parsing code from Zotero a while back.
>> in a database that Firefox reads, and
>> it writes to that database to store its information. You have to
>> unzip a few things, dig around - it's a real
>> educational adventure :)
>> FF2 plugins seem like kind of a wild frontier. There are so many
>> crazy layers of software between writing their code and seeing it
>> work, it takes a lot of fortitude to develop them. (IMHO)
>> Certainly we could learn from them how to get metadata from other
>> sources, but I can't imagine that we
>> could directly use their code, since it is written for such a
>> different situation.
>> Probably we can look at what they do to build conversion dictionaries
>> that map field names between types -
>> maybe that is the most we can get out of it.
> Ok. Another question along the Zotero lines: Let's say several
> metadata sniffers are added to BibDesk, would it be hard to create a
> Safari plugin from this? Most of the time we do our surfing in the
> browser, so a Zotero-like solution is useful in many situations. But
> we don't want the actual data managment to happen in the browser. So
> the best of both worlds would be a Safari plugin which recognizes
> sites with citation data and offers the possibility to move this data
> over to BibDesk. Surfing in BibDesk is itself is certainly also
> useful, but I think most users would profit more if they could do the
> surfing in Safari.
I agree that most browsing is now done in a browser - but I think
there's a chance that BibDesk's web support might be able to be
specialized in ways that would make using it preferable to a regular
browser for citations searching.
I think about how NetNewsWire's built-in browser gets heavy use from
people who want to keep their recreational newsreading separate from
work browsers. I think that having BD be a separate research-only
browser might be a win along those lines.
Certainly, some more thought needs to go into how to fit it into the
usual workflow, and I'm expecting some good discussions on that once
it's in the builds and we talk about it on bd-users.
Something along the lines of a plugin is possible, sure. I can't say
how hard it'd be, because the last time
I looked at building a Safari plugin was a while ago, and they were
underdocumented and hard.
This has surely improved, but I can't comment on that just yet.
I think it might be easier in Firefox, but I'm not sure how to
communicate with other apps - can a Firefox plugin call AppleScript?
Do I want to learn enough to figure that out? I'm not so sure.
Another issue is that everyone uses a different browser - a solution
that avoids customizing for a browser would be great. I still hate
that my browsing history isn't shared by all my browsers - for an
example of why this is annoying, how useless is the recent safari
downloads submenu for the local-url attachment if you use OmniWeb