Skip to end of metadata
Go to start of metadata

XML databases:

eXist is "market leader" and most production-ready. After creating a plausibly sized and scoped data set many of the extra features which eXist buys you had to not be used for efficiency. Ended up with something which just resembled a regular database, but more cumbersome! Probably not ready yet as a technology for this size app.

Open Collection

OpenCollection: not very customisable, particularly for a non-a programmer, or someone without a lot of training and knowledge of its workings. Minimal documentation available at a high level. Overall experience and navigation are good, though. Should look at navigation flow and see what we can use in Collection Space.


For publishing museum collections on-line. Fairly straight forward MVC structure allowing some plug-in functionality and customisability, though some obvious features such as authentication seem to be hard-wired. Probably only useful to CSpace as an exhibition plug-in but lacking in code documentation.


There are implementations of Lucene in multiple languages, though the original was written in Java. This looks like a computationally intensive part of the system, so we ran it against both Java and PHP implementations. The java implementation was 50 times faster than PHP. This seems to suggest that computationally expensive ops are still best performed in java or similar language. The java Lucene handled a large data set in minimal time (250,000 records in less than half an hour), and suggests it's going to be a valuable tool.


Not really principal target domain of application, but a lot which it looks like we can use. Need to see, like eXist, if the benefits of the builtin functionality exceed the cost. Costs include it being reasonably painful to install, complex, and poorly documented, so might be a time sink. Still under active investigation with plausible dataset.

  • No labels