Perhaps I was not clear.

A "single" file for the database has limitations in size. For example some file 
systems have a 2 GB limit..

The concept is to create tables spaces where each table space is a single file.
However you can have multiple tables spaces.  Now these files can either be 
cooked or raw.

So yes. You are correct that this should be the direction.  But there is a 
catch.  Who is going to do this?
Are they really qualified to do this?  How do you resolve issues when the 
change in structure effects other efforts?
How do you workout design philospohiesthat will effect the future of the 
product?

This is no small task.
-----Original Message-----
From: "Kervin L. Pierre" <[EMAIL PROTECTED]>
Date: Wed, 14 Dec 2005 08:51:03 
To:Derby Discussion <[email protected]>
Subject: Re: single file database?

Michael Segel wrote:
> On Wednesday 14 December 2005 2:42 am, Roger Keays wrote:
>>I saw on the derby todo list[0] that there were plans to store a derby

[...]

> That would be highly inefficient and wouldn't scale.
> 

Why SQLite does this ( http://sqlite.org/ )
and is very fast.  May be faster than Derby.

Application users are used to having a single
file manage a 'set of work'.  Eg. a PSD file
for work with an image, or a XLS, or DOC file
etc.

If a developer uses Derby for storage this
would get cumbersome.  Zipping the derby DB
directory is one approach, but that
complicates application saves, crash recovery,
etc.

Regards,
Kervin


Sent via BlackBerry.

-Mike Segel
Principal
MSCC
312 952 8175

Reply via email to