Gsoc project idea

[NAVNEET SUMAN]

I would like to work on the Improving the less popular backends supported by django for this GSOC . What are the various backend u would like to see in django/ or what are the issues related with less popular backends

--

 Thanks and Regards,

Navneet suman,

IIIT-Allahabad

+91-9305612151

[Josh Smeaton]

The contrib.gis app and the GIS backends definitely need some loving. There are a number of workarounds and hacks present in the GIS app that makes it a lot harder to work on the ORM in general. I'm not sure if improving GIS would be too ambitious for a GSOC project though, and it definitely requires a fair bit of domain knowledge and ORM knowledge.

I mention GIS because I think you'd have a hard time convincing the community that a new backend should be included in core. There are already several third party backends, like mssql and firebird, and I believe the preference is to keep them as third party.

As for the existing backends in core, I believe oracle is probably in need of the most attention, purely because it isn't used as widely, and is somewhat difficult to install and test against. I actually use oracle in production, and haven't run into that many issues, but there are several core maintainers (whose names escape me at the moment) that look after the oracle backend that may be able to point out specific pain points.

There's also a ticket open to clean up and improve TEST_ database settings here (discussion here) that you might be interested in looking at. To be honest, this is probably a good sized project with enough design decisions required to be appropriate for a GSOC. The problem might be a lack of interest in getting this done compared to some other ideas though, so you'd probably want to get some buy in before delving too deep.

I hope some of this helps.

- Josh

[Shai Berger]

Hello Navneet,

On Monday 17 February 2014 17:49:36 NAVNEET SUMAN wrote:
> I would like to work on the Improving the less popular backends supported
> by django for this GSOC . What are the various backend u would like to see
> in django/ or what are the issues related with less popular backends

The idea presented in the GSOC page[1] refers to backends both in and out of
Django; it mentions some of the deep problems of the Oracle backend (which is
in core), and refers you to a list of open tickets.

If you want to include other (that is, non-core) backends in your proposal,
you should probably mail their respective maintainers or communities. The
Django project will support work to improve a third-party backend, but you
will need to secure also the support (and probably co-mentorship) of the third
party.

With respect to more details -- like Russell said[2], part of the GSoC
"audition" process is for you to do some analysis yourself; take the initial
pointers and come up with a concrete workplan.

Josh Smeaton wrote:
> The contrib.gis app and the GIS backends definitely need some loving.

Actually, between the custom-lookup and -transform work (led by Anssi) that
landed in master recently, and Marc's Postgres project[3], I expect quite a
lot of changes in the GIS backends in the near future; I think, therefore,
that now is not a good time to create a detailed plan of work in that area.

My 2 cents,
Shai.

[1] https://code.djangoproject.com/wiki/SummerOfCode2014#Improvingthelesspopulardatabasebackends

[2] https://groups.google.com/d/msg/django-developers/q0SVuNkD1Wo/906R0N0Nq38J

[3] https://www.kickstarter.com/projects/mjtamlyn/improved-postgresql-support-in-django