[Tutor] Search theory for luddites?

[Alan Gauld]

> Is there a more efficient approach? I don't think a relational DB
> really suits this sort of task at present, 

Why not?

If its a big data set then a proper database will be more effective.
The advantages are that you can define many more parameters for 
searching - author, date, number of accesses, version, etc
Also it allows easy use of wildcards and ranges in your searches.

And search times will be more close to constant for large volumes.
It will also make generating stats for your own use easier.

If you keep the structure simple and only store paths/filenames rather 
than trying to store the files then a relational database seems 
perfectly reasonable.

Another approach to consider would be an LDAP Directory, but 
I'd go with a database for this one if volumes are to grow to more 
than a few hundred items.

Alan G