[Grok-dev] Re: STORM howto

[Martijn Faassen]

Martin Aspeli wrote:
> Martijn Faassen wrote:
>> Christian (Klinger), Sebastian, and Brandon too:
>>
>> I just started a discussion on the Storm mailing list with some 
>> questions. I'm exploring whether Storm could make for a good default 
>> relational database integration layer for Grok.
> 
> Has anyone done a comparison between Storm and SQLAlchemy? The latter 
> seems to be more popular (anecdotally) and widely adopted, and I can 
> attest that it's a joy to work with, but I've never used (or hardly 
> heard of) Storm.

It's a good question. I am really on the fence between SQLAlchemy and 
Storm myself.

Storm is less high-profile in the Python world. It's high-profile in the 
Zope 3 world though: created by Canonical who do Launchpad too, and used 
by Lovely systems. Comes with Zope 3 integration out of the box - it's 
actually *in* Storm, created and maintained for us by the original creators.

My impression (not based on in-depth experience) is that:

* Storm is simpler to use than SQLAlchemy

* SQLAlchemy has more features

* SQLAlchemy looks more mature in its documentation

* Both have good integration features

* Both have good scalability features

I was at a talk by Gustavo Niemeyer about Storm last year at EuroPython, 
and he was *so* careful not to say anything negative about SQLAlchemy 
in comparison to Storm that I had to ask him what made Storm better at 
some things. :) Storm *was* written as a response to some lack of 
certainly scalability features in SQLAlchemy, though to my knowledge 
these have also been mostly lifted.

The amount of personal connections from the Grok project to the Storm 
project are bigger than our connections to the SQLAlchemy project. That 
counts for something to me.

Storm being somewhat simpler to use, being engaged in the Zope community 
and personal connections to creators makes it interesting to explore to me.

Regards,

Martijn