Why not anonymous messages?

[Manuel Saelices]

Hi all,

I'd like to ask for giving "anonymous messages" [1] a last chance to
be within 1.0 version. Why not?

We've been using that for months with no problem in production sites.
I think the possibilities for that feature are enormous, enabling anon
users communication in:
* Contact form responses
* Cart items updating
* Newsletter joining
* User registration messages (before registration itself)
* All anon forms validation
* etc.

Also, you can easily homogenize user communication mechanism
(look&feel, js dialogs, etc.). For example, take my company contact
form as example [2]

By the way... if this good feature is not included, there is no
problem... current patch is non intrusive and we can (like now) path
our django deployments.

Thanks all for reading, and for this perfect framework :-)

Regards,
Manuel Saelices.

[1] http://code.djangoproject.com/ticket/4604
[2] http://www.yaco.es/contacto/

[Jacob Kaplan-Moss]

On Mon, Aug 11, 2008 at 2:18 PM, Manuel Saelices <msae...@gmail.com> wrote:
> I'd like to ask for giving "anonymous messages" [1] a last chance to
> be within 1.0 version. Why not?

Unfortunately, even if there's a bunch of great, amazing, wonderful
reasons to add this, the "why not" is "because it's too late."
Feature-freeze for 1.0 is in a couple of days, and there simply isn't
enough time to get this done. The time to propose features for 1.0 has
quite well passed; this would have been appropriate a month ago, but
not now.

Luckily, this can easily be added later without sacrificing
backwards-compatibility. Remember: Django 1.0 is in no way the last
release! Indeed, it's simply the first in what I expect to be a long
line of awesome releases.

Please tuck this under your hat, and bring it up again when we start
talking about 1.1.

Jacob

[Manuel Saelices]

On 11 ago, 22:03, "Jacob Kaplan-Moss" <jacob.kaplanm...@gmail.com>
wrote:

> On Mon, Aug 11, 2008 at 2:18 PM, Manuel Saelices <msaeli...@gmail.com> wrote:
> > I'd like to ask for giving "anonymous messages" [1] a last chance to
> > be within 1.0 version. Why not?
>
> Unfortunately, even if there's a bunch of great, amazing, wonderful
> reasons to add this, the "why not" is "because it's too late."
> Feature-freeze for 1.0 is in a couple of days, and there simply isn't
> enough time to get this done. The time to propose features for 1.0 has
> quite well passed; this would have been appropriate a month ago, but
> not now.
>
> Luckily, this can easily be added later without sacrificing
> backwards-compatibility. Remember: Django 1.0 is in no way the last
> release! Indeed, it's simply the first in what I expect to be a long
> line of awesome releases.

Of course... 1.1 targeted ;-)

Thanks for your really hard work... and also thanks to all core devs.
Django development speed in last months is amazing

[Tai Lee]

On Aug 12, 6:03 am, "Jacob Kaplan-Moss" <jacob.kaplanm...@gmail.com>
wrote:

> Unfortunately, even if there's a bunch of great, amazing, wonderful
> reasons to add this, the "why not" is "because it's too late."
> Feature-freeze for 1.0 is in a couple of days, and there simply isn't
> enough time to get this done. The time to propose features for 1.0 has
> quite well passed; this would have been appropriate a month ago, but
> not now.

It's a shame this patch has missed the boat for 1.0, especially since
it was "ready for checkin" since January. Was there or is there still
any other hesitation or opposition to the idea or the implementation
(as suggested by recent comments in the ticket) that has prevented
this from being checked in?

[Russell Keith-Magee]

Priorities. The ticket may have been 'ready for checkin' since
January, but so have a lot of other feature tickets. There is a limit
to how much work we can do for v1.0, so unfortunately, some tickets
need to be deferred.

Yours,
Russ Magee %-)

[Malcolm Tredinnick]

On Tue, 2008-08-12 at 08:31 +0800, Russell Keith-Magee wrote:
> On Tue, Aug 12, 2008 at 8:15 AM, Tai Lee <real....@mrmachine.net> wrote:
> >
> > On Aug 12, 6:03 am, "Jacob Kaplan-Moss" <jacob.kaplanm...@gmail.com>
> > wrote:
> >> Unfortunately, even if there's a bunch of great, amazing, wonderful
> >> reasons to add this, the "why not" is "because it's too late."
> >> Feature-freeze for 1.0 is in a couple of days, and there simply isn't
> >> enough time to get this done. The time to propose features for 1.0 has
> >> quite well passed; this would have been appropriate a month ago, but
> >> not now.
> >
> > It's a shame this patch has missed the boat for 1.0, especially since
> > it was "ready for checkin" since January. Was there or is there still
> > any other hesitation or opposition to the idea or the implementation
> > (as suggested by recent comments in the ticket) that has prevented
> > this from being checked in?
>
> Priorities. The ticket may have been 'ready for checkin' since
> January, but so have a lot of other feature tickets.

Note also that "ready for checkin" unfortunately rarely means that.
Larger patches often need improvements even if the triagers believe it
to be ready. In this case, the patch was potentially duplicating
functionality from other parts of Django as well as adding a feature
that wasn't necessarily an obvious win, so the decision was deferred for
a bit. That goes along with Russell said: it just wasn't the most urgent
item on the plate.

Regards,
Malcolm

[Tai Lee]

On Aug 12, 11:19 am, Malcolm Tredinnick <malc...@pointy-stick.com>
wrote:

> Note also that "ready for checkin" unfortunately rarely means that.
> Larger patches often need improvements even if the triagers believe it
> to be ready. In this case, the patch was potentially duplicating
> functionality from other parts of Django as well as adding a feature
> that wasn't necessarily an obvious win, so the decision was deferred for
> a bit. That goes along with Russell said: it just wasn't the most urgent
> item on the plate.

Thanks for the clarification, which makes perfect sense.

[Simon Willison]

On Aug 12, 8:56 am, Tai Lee <real.hu...@mrmachine.net> wrote:
> Thanks for the clarification, which makes perfect sense.

It's worth mentioning that anonymous messages is a feature which could
be added to Django without any modifications to the actual Django
repository at all - it looks like it could be done with some
middleware, a request context processor and maybe a custom template
tag. I'd suggest putting together the feature as a separate project,
then proposing it for django.contrib in time for 1.1. That way people
can start using it (and helping improve it) straight away.

Cheers,

Simon

[Julien Phalip]

> I'd suggest putting together the feature as a separate project,
> then proposing it for django.contrib in time for 1.1. That way people
> can start using it (and helping improve it) straight away.

Agreed. It is indeed pretty easy to package as a separate app.

[Carl Meyer]

On Aug 12, 9:09 am, Julien Phalip <jpha...@gmail.com> wrote:
> Agreed. It is indeed pretty easy to package as a separate app.

I already have a lightweight version of the 4604 patch separated out
into an app, which I've been using for a month or so. I've just
uploaded it at http://code.google.com/p/django-session-messages/ -
improvements welcome.