> tomi wrote:
>
> Hi guys,
>
> This topic (creating a SQL backend) has been added to the TODO list. I
> studied a little the problem and I had a little discussion on it with
> Geoff Hutchinson. It is a great step forward creating databases
> directly in SQL based formats:
> 1) it is much more reliable;
Why? You make a statement without proofing it. Any stable database
system
can be said to be equally reliable to another one. However, since SQL
systems
allow the database to reside on *remote* machines and thus require the
backend
to have a net connection to the SQL host, it might even be *less*
reliable than
approach used by Ht://Dig.
> 2) considering many cases a webmaster or a system enginer could face
> (very huge databases, Clustering, RDBS, etc.), SQL is the best way to
> overflow any problem, thanks to the engines of databases as Postgres
> or mySQL (just to cite only GNU projects).
SQL is surely a more general approach, favours bigger databases and
distributed
processing of queries. It also allows for concurrent updates of the
search
engine databases more easily than BerkelyDB. But AFAIK are neither
PostgreSQL
nor mySQL GNU Projects (GNU is running its own database project which
has its
goal set on implementing a SQL-92 database system. I have not seen a
useable
distribution of it, however).
> Creating an SQL database is not difficult:
> infact, if we want to create, for example, the table employee, we
> should simply follow the following statement:
>
[...]
>
> Will follow other specifics to insert and modificate datas if You'' ll
> require. Yhis e-mail was only to demonstrate the ease of creating SQL
> tables and databases.
You used a rather trivial example here. One of the main problems of the
Ht://Dig search engine with regards to an SQL backend implementation is
the use of document excerpts which in nearly all cases are too large for
many popular SQL database engines. Not only are those excerpts pretty
large, but they are of a dynamic length. In order to make such a
document
database portable to all SQL engines, you will probably need to store
each document in a single file in a document-db directory, causing
lookups
to be extremely slow for larger databases - let alone that you (again)
have to stick with single host databases and cannot truely utilize the
power of the SQL engine unless you also use networked file systems for
this document-db directory.
just my 2 cc,
Torsten
--
InWise - Wirtschaftlich-Wissenschaftlicher Internet Service GmbH
Waldhofstra�e 14 Tel: +49-4101-403605
D-25474 Ellerbek Fax: +49-4101-403606
E-Mail: [EMAIL PROTECTED] Internet: http://www.inwise.de
------------------------------------
To unsubscribe from the htdig3-dev mailing list, send a message to
[EMAIL PROTECTED]
You will receive a message to confirm this.
