Lutz Donnerhacke <lutz(_at_)iks-jena(_dot_)de> writes:
Furthermore it's easy to implement a keyserver, because you notice how many
new keys/signatures/userid etc. pp. occured. So merging keys is very simple.
Yup, and real-time status queries are also incredibly trivial, the user
submits a request (e.g. via HTTP, http://......?keyID=abcd), you translate it
into "SELECT COUNT(*) FROM validKeys WHERE keyID=abcd" and return the output
to the user (there's an RFC draft on this floating around somehere on
ietf.org).
You forgot:
- Maintaining 10 tables at once is entertaining for the software developer.
Yep. Obtaining standard query strings from you local DB developer (it's fun
for him) and using them as black box is plain recovery. ;-)
While we're rhapsodising about the joys of SQL keyservers, it also makes
debugging incredibly easy, you just drop your favourite SQL browse tool onto
the database and you can check everything that happens, arranged/filtered by
whatever criteria you want.
Peter.