bug-gne
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-gnupedia] Architecture Questions


From: Bryce Harrington
Subject: Re: [Bug-gnupedia] Architecture Questions
Date: Sun, 21 Jan 2001 01:44:30 -0800 (PST)

On 21 Jan 2001, Mike Warren wrote:
> What exactly are the advantages of using a MySQL database? I'm not
> extremely familiar with it, but I don't see why human readable files
> (in XML or SGML or whatever) aren't a Good Thing.

Performance, mostly.

If you're a Perl head, think of a database as akin to a permanent hash.
If you're a C/C++ adherant, think of a database as like to a permanent
memory buffer.  If you're an HTML hacker, think of a database as a sort
of "automatic" search engine.  

There's actually _three_ options here that we're considering:

A)  No database.  All files and their headers are stored in the file
system and nothing else.

B)  All database.  All data including identification info is stored in a
database.  

C)  Hybrid.  An index is generated from the original files and fed into
the database.  The database is used for doing the searchin' stuff, and
the original files used to provide the complete info.

I favor C.  This is probably clear.  A is sufficient initially, and I
would support it from a starting point, however it will not scale well.
B is ok in theory, however databases do have limits, and they do crash
from time to time, and putting all faith in them is unwise.  Databases
are good for doing indexing and searching.  File systems are good for
storing.  Let each do what it does best, IMHO.

Bryce




reply via email to

[Prev in Thread] Current Thread [Next in Thread]