Wiki and tagged bookmarks integration · Curiosity is bliss

Curiosity is bliss    Archive    Feed    About    Search

Julien Couvreur's programming blog and more

Wiki and tagged bookmarks integration


Here's a quick idea: allow a wiki to pull a list of links from a bookmark store using the WikiWord as the requested tag.
The links could be broken down into multiple buckets depending on other specific tags like "read" or "unread" for example. The tags on these imported links could be updated directly from the wiki, using some javascript and DHTML. This would be quite useful for organizing information during the research phase.

Looking at for the bookmark+tag store, it turns out the APIs don't allow modifying tags on an item (it looks like you can only add new tags). Re-tagging is important so that you can mark items "read" or "unread" as you progress in your research.

Another problem with is that it doesn't offer any private "space". But the wiki and tagged bookmark integration makes more sense with a private wiki and private bookmarks ("read" or "unread" tags should be personal, not shared, and wikiwords need to be scoped to a single wiki, not global). Unfortunately, I couldn't find any project for private bookmark+tag store, on freshmeat or sourceforge.

Trying to think about other integration scenarios, another case seemed to have potential: using urls as wikiwords. This would turn a private wiki into a personal annotation store.

Update (2005/02/01): Matt Biddulph posted an related hack: integrating tagging into a website (via Boing Boing). He's using DHTML and javascript for adding tags to the current page and fetching and displaying other urls with the same tags.


I think you're mistaken about the API. I think it's perfectly valid to use /api/posts/add to re-post an existing URL with a different set of tags, so you can add or subtract tags with this technique.

There are still several operations which are missing, like fetch a post by URL, and fetch "all users posts" on a URL.

I like what you're suggesting about using tags to track workflow; there has been some discussion along these lines on the delicious-discuss list:

related to this discussion of a use case:

I have been wanting to make an inside out version of what you're suggesting out of wikalong + delicious, so that for each page I visit, a Firefox sidebar shows tagging information from delicious. I suspect that Joshua wouldn't like this idea much, though, because it would result in an expensive hit on the database for every page loaded in my browser, which is several thousand a day.

It could work easily enough with a local copy of your own information, but it would be cacheing of the "all users" data scalable.

PS I found your post via pubsub:

Posted by: Brian Del Vecchio (February 1, 2005 05:53 PM) ______________________________________

What I meant was:

It could work easily enough with a local copy of your own information, but it would be [difficult to make] cacheing of the "all users" data scalable.

Posted by: Brian Del Vecchio (February 2, 2005 12:27 PM) ______________________________________

Thanks Brian. I see what you mean about being to update a tag list using the API.
I also just realized that the "recent" API only returns results from my bookmarks (not from other users).
Cool! It seems like these APIs are sufficient for scenario I described. The only problem is the scenario is not very community friendly, as it tags items with wikiwords from my private wiki...
Since the APIs return xml, it seems like javascript+XmlHttpRequest is a good candidate :-)

Posted by: Julien (February 27, 2005 09:48 AM) ______________________________________

Actually, remote integration using javascript doesn't work, XmlHttpRequest only lets you load a url in the same domain as the current page :-(
See the "permission denied" comment at

I tried the iframe approach as well, but the "same domain" security policy in javascript also prevents access to any data returned by within the iframe... :-(

Posted by: Julien (February 27, 2005 03:20 PM)
comments powered by Disqus