Hotclub or Couchdb Backend
thauber
had a couple ideas and have narrowed them down to two.
First I have talked to James Tauber and he said that he would like to
see some hot club apps worked on for GSoC. This means I would be
working on django-friends, django-mailer, and django-notification.
Also he wanted to see a django based opensocial container. This would
be my first choice, as hotclub has always been something I thought was
important to django.
Second I would like to work on a couchdb backend. I have been working
on couchdb on another project and I think that it is a really cool
database. While it may not exactly fit with couchdb document based
database to have a relational mapping, I still think that there would
be benefits. It would require some sort of checking to see if what
goes in and what comes out fits the model because there is not tables
or rows in couchdb.
If anybody would be interested in this I would love to work on it this
summer. Please let me know what you think.
Malcolm Tredinnick
On Sun, 2008-04-06 at 10:45 -0700, thauber wrote:
> I am going to apply for the GSoC for the Django organization. I have
> had a couple ideas and have narrowed them down to two.
>
> First I have talked to James Tauber and he said that he would like to
> see some hot club apps worked on for GSoC. This means I would be
> working on django-friends, django-mailer, and django-notification.
> Also he wanted to see a django based opensocial container. This would
> be my first choice, as hotclub has always been something I thought was
> important to django.
Hopefully any such application will include an explanation of what this
"hotclub" thing is for those of us, like me, who don't have the faintest
idea.
Malcolm
--
Atheism is a non-prophet organization.
http://www.pointy-stick.com/blog/
Russell Keith-Magee
>
> I am going to apply for the GSoC for the Django organization. I have
> had a couple ideas and have narrowed them down to two.
>
> First I have talked to James Tauber and he said that he would like to
> see some hot club apps worked on for GSoC. This means I would be
> working on django-friends, django-mailer, and django-notification.
> Also he wanted to see a django based opensocial container. This would
> be my first choice, as hotclub has always been something I thought was
> important to django.
Keep in mind that the GSOC guidelines specifically exclude projects
other than coding. Feel free to correct me, but a lot of the work
required by Hotclub involves documenting interfaces and building a
community website on which to share application resources.
I agree that the Hotclub idea is an important one - but unless you can
frame it in terms of a specific coding goal, you will probably find
your application rejected.
If your project is to work on django-friends, django-mailer et al,
then your project isn't really "Hotclub" - it's to build/extend some
useful Django applications. Hotclub compliance (whatever that means at
the moment) is a feature on your to-do list, not the goal of the
project itself.
> Second I would like to work on a couchdb backend. I have been working
> on couchdb on another project and I think that it is a really cool
> database. While it may not exactly fit with couchdb document based
> database to have a relational mapping, I still think that there would
> be benefits. It would require some sort of checking to see if what
> goes in and what comes out fits the model because there is not tables
> or rows in couchdb.
I haven't played with couchdb myself, but from what I hear on the
grapevine its an interesting resource - but a completely non-SQL
resource. This would certainly be an interesting contribution.
However, keep in mind the time available for a GSOC project. Malcolm
is an extremely talented developer, yet it has taken him the better
part of 18 months to get the QuerySet refactor code to the point where
it is nearing merge quality. Integrating CouchDB has the potential to
be just as significant a change to Django. I submit to you that unless
you have a specific plan for constraining the complexity of this task,
rewriting the Django query engine to support a non-SQL query language
is perhaps a little outside the scope of a 4 month student project.
Yours,
Russ Magee %-)
thauber
> then your project isn't really "Hotclub" - it's to build/extend some
> useful Django applications. Hotclub compliance (whatever that means at
> the moment) is a feature on your to-do list, not the goal of the
> project itself.
on some apps that James Tauber has been saying would be good additions
to hotclub. You are also right that if I am going to apply I need to
be specific about what it is I am planning on working on (ie django-
friends or django-notifier or both etc) I'm not sure what the best
scope for this project would be.
> However, keep in mind the time available for a GSOC project. Malcolm
> is an extremely talented developer, yet it has taken him the better
> part of 18 months to get the QuerySet refactor code to the point where
> it is nearing merge quality. Integrating CouchDB has the potential to
> be just as significant a change to Django. I submit to you that unless
> you have a specific plan for constraining the complexity of this task,
> rewriting the Django query engine to support a non-SQL query language
> is perhaps a little outside the scope of a 4 month student project.
impossible for me to code in 4 months. I appreciate your input I
think its really helped me with some obvious weaknesses in my
application.
- Tony Hauber
James Tauber
On Apr 6, 2008, at 7:31 PM, Malcolm Tredinnick wrote:
> Hopefully any such application will include an explanation of what
> this
> "hotclub" thing is for those of us, like me, who don't have the
> faintest
> idea.
The Hotclub of France is a project to help build a series of reusable
Django apps. The goal is to have them follow certain conventions and
play together nicely. "hotclub" apps don't necessarily have to have
been written with the HCoF in mind -- django-registration, django-
tagging, etc would be candidates.
It was born out of conversations I had at PyCon 2007 with Jacob, James
B and others. There was a discussion on django-users about it around
that time which lead to the mailing list[1] being formed. The list
incidentally now how 160 members although hasn't had much activity.
Also see http://code.google.com/p/django-hotclub/
It would be nice having someone spend some dedicated time to building
out a couple of the apps that have already been started.
I could act as a mentor or co-mentor if you guys have enough slots to
go ahead with it.
James