Re: [tlug] wiki

[<burlingk@example.com>]

> ------------------------------
> 
> Message: 13
> Date: Sun, 19 Nov 2006 11:26:25 +0900
> From: "Josh Glover" <jmglov@example.com>
> Subject: Re: [tlug] wiki
> To: "Tokyo Linux Users Group" <tlug@example.com>
> Message-ID:
> 	<d8fcc0800611181826y107e8d9bh4413b2dd98244a94@example.com>
>
> Locking is but one of the database-like features that wikis 
> need. How about building and maintaining of link tables? How 
> about indexing things for fast searching?
> 
> All of this could be done with flat files and grep, but why 
> reinvent the wheel when MySQL / sqlite / PostgreSQL / etc. 
> works just fine?

I think that the biggest worry seems to be recoverability in case of a
database crash.

If a person is really totally worried about that, it is probably
possible to create a script that semi-regularly dumps the databases
information out into a flat text file.  A second script of course would
be needed to turn that data back into a database if the need were to
arise.

The same interface that allows the wiki to interact with the database
server to begin with should allow for such a script.

Another option however, is to simply back up a copy of the database
itself, and if the server moves to another box, use the same version of
the SQL  software.