[Systers-dev] Moving from Pickle Membership to Store in the Postgresql Database

[Jaideep Khandelwal]

> No, you probably want to make this unified record at the very beginning, or
> it's too messy (it's easy to treat the first as a distinguished case, not so
> much the second).  It's not a single signon unless you have a single
> location for user name and password.
>
> But my point was that you don't want redundancy (from there come errors),
> so you need to think about how things change when this information is not
> stored in the individual list.  Especially because you don't want to have to
> modify the code for every application that we want the single signon for (I
> think that you can assume that applications outside of mailman have some
> sort of openid or similar system hook that you will be able to use. Do you
> want to create such a hook for mailman?  Do you want mailman to be just
> another application in this set or "special"?
>

I would better prefer the mailman to be the special application as the data
would be collected from the users coming to the Mailman should only be
storing the credentials , and yes the hook will be for mailman that is to be
created and it should be restricted for the systers concerned site.


>
>> > And I think the bounce information (if I'm understanding it correctly)
>> > needs
>> > some serious thinking through.  Different applications may have
>> different
>> > ways they count bounces, so you may not be able to have a single "if it
>> > bounces this many times, drop them" metric.  And if someone belongs to
>> 10
>> > systers.org lists, and the system sends out password reminders for each
>> > list
>> > (or something similar) and they all bounce due to a temporary problem,
>> we
>> > wouldn't want to count that as 10 bounces.  So I think you need to think
>> > through the various pros and cons of keeping bounce information in a
>> > central
>> > place vs. for each list.
>> >
>> > I would like to have an IRC meeting and have a discussion about this
>> issue.
>>
>
> This sounds like a good idea.  Unfortunately, I'm about to go off on
> vacation, where I won't have reliable internet access (till next Monday).
>  What time zone are you in?  If we make it relatively short, I might be able
> to find time during the day (California time) on Wednesday before I leave.
>  I think Malveeka is in the same time zone as I am.  Who else should be in
> the meeting?
>
My time zone is : +5.30 GMT
I am busy with my Exams  so Wednesday will not be possible for me and after
that it will be too late for you and Malveeka. It would be nice if we can
fix it on Wednesday 1900 hrs (GMT) and for the meeting it would be nice if
Anna and Pinar could be present so that we can discuss the working on
Database issue.

I don't think I understand what you said here.  I'm talking about storing
> ONLY the login information, not any data needed by a specific application;
> that is, what you store rather than where you store it.
>
> While we definitely want to be able to use this for CMS, wiki, etc., I'm
> not sure how your approach does that.  I think we have to assume that we
> aren't willing to make code changes to every new application that we install
> on the systers machine.
>
> No , for that we would rather prefer the Open ID provider that will be used
for the applications and  many of the applications such as CMS , wiki now
have the modules for OpenID Clients .

Regards

Jaideep