[ajmas]

Although already posted by other means here is an idea for a plug-in architecture for allowing to search other sites:

Stealing an idea already used by Konfabulator and Dashboard, it would be to allow plugins written in Javascript along with XML. The Javascripts would supports a number of DL specific keywords, or even have a DL 'object' witht the terms 'title', 'creator', 'ean', etc. This code would then be able to go off, do a search using whatever means it wants to (could even call an external application) and then compile the result in a XML format that would be compatible with the one used by DL for storing its data.

At a later time DL could be extended to support other languages, but I think keeping it simple to start is what is important.

What do you think?

[Delicious Monster]

This is a pretty common request, the most compelling argument being the idea that if we implemented the architecture people around the world could make and share plugins expanding Delicious Library's functionality and lowering our workload.

I have my doubts, and I'll tell you why.

Delicious Library's data is wide open in XML, a format understood by even the most novice of programmers. Even without a plugin architecture, there's nothing to stop anyone with a decent knowledge of any programming or scripting language from watching the data file, picking out failed lookups, looking up the information, and inserting the results into Delicious Library.

The fact that, in two years, nobody's done that leads me to suspect that nobody will write plugins either.

Of course, that's not to say we wouldn't ever write a plugin architecture, even if for our own use. It's just good design. However, I don't think plugins are the panacea they're made out to be.

[ajmas]

If there is a general work around to a lack of a plugin architecture, for achieving search results from alternative sites, then I would ask anyone to post a sample on how they achieved it.