We’ve not afraid of embracing new technologies at Loopo, and right now this seems to have reached fever pitch. The time was ripe to try some Python development, and we’re now running both Plone and Django projects back to back.

Everyone at Loopo uses Skype, and I was very amused when I noticed that Julian, who’s heading up the Plone project, had set his status to

“Julian is diving into the depths of Plone’s Python codebase.”

Meanwhile our Django development lead had:

“Fred is surfing the shallow, warm, lovely waves of sweet smelling Django.”

I think that about sums it up really! Plone can get extremely murky once you start delving under the bonnet and poking around with Zope, although it’s well worth it for the power that lies beneath. Django however is such a transparent, well thought out and beautifully documented piece of work that it really is hard not to get down on one knee and propose marriage after a few hours of using it.

That’s not to say that it’s without faults.

Speaking with my database developer/information architect hat on, I do find the insistence on driving all of the modeling from the application’s perspective can get a bit wearing. Having said that, it’s relatively trivial to post-edit your tables and carry those changes through to models.py. I was also impressed that Django automatically adds indexes to all of your foreign keys when you run syncdb.

We are developing a fairly high availability web application with Django, which will doubtless need some further performance tuning further down the line. It’ll be interesting to see how much of that we can easily code into the models, and whether we’re left with a load of build scripts to run in MySQl as part of the install.

I’ll blog more about both these projects as they progress.