You may have noticed that we have recently started linking Dewey numbers to dewey.info in blog posts and @DeweyTeam. The data refresh and adding of Table 2 notation that happened over the weekend may be a good opportunity to look closer at what dewey.info is providing (and how it is different from other sources of DDC data).
First and foremost, dewey.info was conceived as being a web service, which means there is a good chance that most members of its target audience are not breathing (being computers and such). Yet, because dewey.info publishes data according to linked data principles, data are very much being treated like regular web pages, so when a human looks at a Dewey class via dewey.info, the system conjures up a HTML page for the class, e.g., T2—43 or T2—629.
The HTML page corresponds to the raw data a machine would see, but is in some ways richer than, in other ways not so rich as the machine representation. Both the data and HTML views contain at their core number and caption information. The data view adds a lot of metadata about the class(es) and is very precise in terms of referencing other classes.
The HTML view is more compact, but instead of showing raw URLs (i.e., hyperlinks) to super- and subordinate classes, it actually looks up their captions to save the user several trips. Also, in cases where there are several versions of a class available for a specific edition, only the most current will be shown (unless expressly requested by, e.g., adding a date component to the URL).
In short, web pages in dewey.info show captions in their immediate hierarchical context with minimal frills. Dewey.info allows allows a quick caption look-up but is rather sparse compared to services aimed at classifiers like WebDewey. For example, dewey.info does not contain notes, and only includes a small subset of Relative Index terms (those terms used as substitute captions for built numbers). To see the full data associated with each Dewey class it is still necessary to consult WebDewey or your trusty print volumes of the DDC.
On the other hand, dewey.info (being linked data and all) connects up much better to other datasets. A nice example is the huge amount of structured data that was recently added to WorldCat.org. This new release of dewey.info even helps you find a related number in the same hierarchical array if the requested class is not part of the dataset (yet). For example, let's say you want to find out more about the rover missions exploring Mars (919.92304; depending on the emphasis of the actual work, the topic could also belong somewhere in 629 or 559, as explained by the Manual note at 629.43, 629.45 vs. 559.9, 919.904, and exemplified by the recent tweet of my colleague Juli Beall about Curiosity before its landing.) This specific number has not been built, but instead of just responding with a "the requested resource was not found" message, the service will provide a redirect to the most specific number that is available (in this case: 919). A quick glance at one of the narrower classes (919.9204) may give you good idea what the originally requested class was all about and how it was built (namely, using 91 Geography and travel + T2—9923 Mars + 04 Travel from the table under 913–919). As you may have noticed, the latter two classes are not available in dewey.info. Except for Table 2 notation, dewey.info only includes "assignable" classes, which exclude internal add table notation or number spans.
There is lots more to explore in dewey.info, but we may want to leave other topics (e.g., using the SPARQL search interface for querying the data) for another day.
Do you have any video of that? I'd want to find out more details.
Posted by: Dawn | 29 September 2012 at 09:44 PM