At 7:43 PM +0200 7/23/00, tomi wrote:
>We' ll I would do, if I' ll be able to do it to htdig too... maybe
>with your help ('cause I don' t know all the code...) ... always if
>you want...
I don't know how clear this has been, but we'd like to see a good
quality SQL alternative for ht://Dig. However, it wouldn't be easy
for several reasons and certainly I have my hands more than full with
other tasks. :-(
>Anyway... I would first know the use of htmerge in creating
>batabases... because it could be trivial once created the SQL
>indexing procedure...
In the 3.2 code, htmerge isn't used for creating databases. It's only
used for merging them. The databases are made on-the-fly by htdig.
>I' ve heard too htdig could generate an ASCII file where puts the
>informations about the site retrivied (not sure), to be inserted
>then in the database... Well this could be usefull if we (or
>you..:)) decide not to change the indexing procedure, but only the
>format of the ASCII text file (according to SQL specs)...
Well, you can generate an ASCII version of the database, but it's not
exactly space-efficient or easy to use for queries. Certainly a
well-designed SQL schema would be the way to go for a SQL backend.
>Even more... there is a program were to look for an implementation
>such this: UDM Search... that has supports for every database
>formats... just to look at what they did...
Let's just say that UDM Search has rather different ways of doing
things. In most cases I cannot see that they are any better. For
example, I don't see a very good reason to store the word database in
a SQL index because it's very space inefficient.
In any case, let me not dissuade you from working on SQL code. We'd
all be more than happy to take a look at it and test it out.
-Geoff
------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
[EMAIL PROTECTED]
You will receive a message to confirm this.
