File naming best practice

Question

Hello
We are building a CMS system that uses BDB XML to store the individual xhtml pages, editorial content, config files etc. A container may contain tens of thousands of relatively small (<20 Kb) files.
We're trying to weigh up the benefit on meaningful document names such as "about-us.xml" or "my-profile.xml" versus integer/long file names such as 4382, 5693 without the .xml suffix to make filename indexing as efficient as possible.
In both situations the document remain unique: appending '_1', '_2' etc where necessary to the former and always incrementing by 1 the latter.
There is a 'lookup' document that describes the hierarchy and relationships of these files (rather like a site map) so the name of the required document will be known in advance (as a reference in the lookup doc) so we believe that we shouldn't need to index the file names. XQuery will run several lookups but only based upon the internal structure/content of the documents, not on the document names themeselves.
So is there any compelling reason not to use meaningful names in the container, even if there are > 50,000 documents?
Thanks, David

Answer 1

George,
I was interested in finding out whether document names made of integers would be much more efficient - albeit less intuitive - in the name index than something like 'project_12345.xml'.
We may need to return all documents of type 'project' so putting the word in the document name seemed like a good idea, but on reflection perhaps we're better off putting that info in the metadata, indexing that and leave the document name as a simple integer/long such as '12345'.
If so, is it worth rolling out my own inetger-based counter to uniquely name documents or am I better off just using the in-built method setGenerateName()? Is there likely to be much of a performance difference?
Regards, David