You are viewing a plain text version of this content. The canonical link for it is here.
Posted to derby-dev@db.apache.org by Rick Hillegas <Ri...@Sun.COM> on 2008/07/01 16:10:08 UTC

Re: simpler api to the Derby store

Hi Dan,

Thanks for responding. Some comments inline...

Daniel John Debrunner wrote:
> Dag H. Wanvik wrote:
>> Bryan Pendleton <bp...@amberpoint.com> writes:
>>
>>> Aren't there some pretty good packages for this already? E.g., BDB-JE,
>>> JDBM, Perst, etc.?
>>
>
>> JDBM is being incubated as an Apache project.
>
> That would take care any licensing concerns.
>
> Maturity wise, it's unclear to me what state JDBM is in, started at 
> source forge in 2000 but last release in 2001? I only looked briefly.
>
> The download seems to be around 250k, which could be a better starting 
> point than Derby. I found a blog entry that said its actual jar file 
> size was 85k.
>
> Community, well maybe not much community currently around JDBM, but 
> coming to Apache could change that, especially if folks from Derby got 
> involved. Seems strange to me for Derby to start up its own lite 
> project while another is in incubation at Apache. Could be a great 
> opportunity for two communities to work together rather than against 
> each other.
For some problems, Apache offers multiple solutions and communities. For 
instance, Maven and Ant are both build tools. Here's another example: 
Apache is incubating Empire-db now--that's another ORM framework which 
will sit alongside Apache JDO and Apache Torque.

It's hard for me to predict how this will play out. However, I don't see 
JDBM competing with Derby Lite. The JDBM jarball weighs only 82K. That 
suggests to me that it's not packed with features but it's a good fit 
for tiny devices. I don't see Derby Lite playing in that space.

>
> If JDBM graduated from incubation at Apache, then why would anyone 
> want to work on slimming Derby down if there was a working solution 
> under the same license?
In my mind, footprint is only one of the issues. Derby Lite would be 
feature-rich, high performance, and multi-user. For instance, Derby's 
multi-column keys would be an interesting differentiator. So would 
replication.

A lot of new databases are cropping up these days. Different problem 
spaces are demanding very specific solutions. I think that JDBM and 
Derby Lite will find different sweet spots.

Regards,
-Rick

>
> Dan.
>