Google search recently launched a product called “Google Coop”.
I’m one for the odd mashup, so this project sounded fairly interesting – you get to make little vertical searches on top of Google to which other people can subscribe, kind of a mix between RSS and Adwords.
However the documentation for this project was clearly written by people who hate me and want to make my life miserable. The Phds up in the GooglePlex clearly have no need to explain their genius to common folk like me, instead opting for really helpful headings like:
“For Contributors”, “Topics”, “Subscribed Links”
Definition of the project: “Google Co-op is a platform which enables you to use your expertise to help other users find information. This is a work in progress; over time, you can expect to see evolution in both Co-op’s structure and the means by which you can contribute your expertise to our goal of making information more discoverable for millions of people. ” (ok but what does it actually do?)
I scratched my head for a bit thinking what these things might actually mean, luckily the excellent Google Blogoscoped quickly came out with a primer, that explained what the heck google coop actually did.
This weekend, looking for another Sunday project I revisited Google Co-Op – and built a little integration between personal del.icio.us bookmarks and Google. How it works is that you visit Google Co-Op and subscribe to my ‘subscribed links’ feed.
Here is a page that explains how to set your del.icio.us bookmarks up with Google Co-Op
If you just want to integrate your del.icio.us bookmarks with Google, you can stop reading here, I won’t hassle you with Google’s screwy terminology or how the annoyances I’ve found working with Google Co-Op.
And there are a number of these annoyances, even going beyond the documentation’s sore need of a basic glossary.
Basically, what I used is the ‘subscribed links’ functionality offered by google coop. This allows you to set an XML file with preset matches to search results. If someone searches for “What is the air speed velocity of an African Swallow”, you can specify they should get a link up at top of their search to the gorgeofnothingness.com
The fun part in setting these up is that Google won’t let you have a sandbox to play around creating these subscribed links. In order to test anything, you have to upload your xml file to them and add it to your subscribed links. You then have to wait 5 minutes until they go and get your file. Then you have to study messages like “xml entity not recognized” (we don’t know how to deal with unicode). Then delete the subscription you added, re-add/re-upload and wait 5 minutes to test again. It’s a lot of fun. Did I mention sometimes if you delete a subscription and re-add it, Google retains its cache of the old subscription and won’t reload the new one?
Other than the hassle of getting started, the issues are fairly minor, mainly my opinion of their xml schema. I don’t like data in XML attributes, it’s a lot cleaner when it’s in nodes. And if you set up counting notations, stay consistant, either go 0,1,2,3 or 1,2,3—don’t mix both please.
Anyways, this was a quick project that has kept me up way too late, but if you want to see the simple source to the project you can visit the SWiK – Kibbutz wiki page.
If you want to see improvements to this project, shoot me a message, I will be looking at adding more or betterer information to the integration and I’d like feedback on if this is actually useful or what aspects of del.icio.us I could integrate with Google further.
Also if Joshua Schachter would expose an RSS feed for del.icio.us search as there is on most other things, or if there were more XML access to a user’s bookmarks, I could probably do a better job integrating the two services.