Outgoing

Live Wiki

Well, the WorldCat wiki functionality went live over the weekend.  Although the concept seems fairly straight-forward, there were a lot of decisions to be made (and a bit of code to be written) to make it happen, and this first version got done in what seems a very short period of time for a major piece of functionality.  It really wasn't much more than an idea when I posted about it in May, although some of the code was starting to work.

Some here at OCLC have said we really shouldn't call it a 'Wiki' since it all act the way a standard Wiki would.  Well, maybe, but there's more there than meets the eye.  Taking some liberties with what a 'review' should be I've created a little review of the Laughing Policeman.  Here's what I put in the review to make this work:

Here is an ISBN link: ISBN: 0752850938
A link into [[WikiPedia:Martin_Beck]] WikiPedia
A link to a discussion of links: http://outgoing.typepad.com/outgoing/2005/10/live_wiki.html

After I finish the Laughing Policeman, I'll go back and make it a real review.

--Th

New Dewey Browser collections

There are two new databases that can be viewed via the Dewey Browser.  Previously we just had the e-books from WorldCat.  Now you can also search a database of 2.2 million of the most widely held WorldCat records, and, possibly even more interesting, selected data from the Abridged version of the Dewey Decimal Classification (DDC).  Note the Collection list in the bar:

at the top of the screen that lets you select which database you want to look at.

The abridged version has 4,300 categories, rather than the 30,000+ in the full Dewey (of course DDC numbers can be built up from various tables in the classification to give an unlimited number of possibilities, even from the abridged).  We think this should be useful; the abridged DDC covers a lot of topics and there are 36,000 terms leading you to the classes.  Try looking for peat bogs.  You should see that there is a topic in Earth sciences (551.41 Geomorphology) and one in Life sciences (578.768 Riparian biology).  The DDC does a good job of pointing out different aspects of a topic and leading you to a useful classification of it.

Disclaimer: I had little or nothing to do with getting these new collections up.  Diane and Harry share the primary responsibility for this.

--Th

A string of PURL servers

Andy Powell, from UKOLN, visited OCLC last week and we had a lot of discussions with him.  Some of them were about PURL's, a method of creating 'Persistent URLs' that we both think gets too little notice.  Currently OCLC maintains the main PURL server at purl.org, although the software is available for others to run independent PURL servers.  Andy thinks we should distribute responsibility for purl.org:

1. Several organizations around the world to agree to replicate the PURL software and database
2. We use DNS to direct users to the most appropriate PURL server
3. OAI-PMH is used to keep the servers in sync, so that PURLs created or modified on one server are rapidly available everywhere

This would make the PURL service more robust, reduce its dependence on OCLC, and give PURLs even more reason to be used.  Stu Weibel says that federation was considered from the very beginning of the PURL service.  Maybe now is the time to try it.

--Th

Note: the image at the top was found on Sprott's Gateway at wisc.edu.  Lorcan suggested 'string of PURLs' which I used even though I don't think a string is quite the right metaphor.

If you are interested in PURLs you might be interested in POIs.

Screencast update

Some have reported problems viewing the screencast I published earlier.  In the interest of portability here's an mpeg version which might be easier to view, especially for Mac users.  Unfortunately it is something like 50x the size of the .wmv file and the quality is much worse, but it might give you an idea of what the system looks like.

We're working on getting the actual system up for public use and I'll announce that here when we do.  In addition to porting the interface from Python into Java/Tomcat, we've been doing some work on the files themselves.  We are now constructing the indexes using our map-reduce prototype and have figured out how to squeeze about three times as many unique phrases into memory, so the system works a bit better.  It will be interesting to see how good the response time is and whether this meets a real need or not, once we get it public (off of the ResearchWorks page).

--Th