Fw: [Pydotorg-redesign] pytdotorg-redesign notes from PyCon OpenSpace session

[Steve Holden]

----- Original Message -----
From: "Anna Ravenscroft" <revanna at mn.rr.com>
To: "Kevin Altis" <altis at semi-retired.com>; "Mats Wichmann"
<mats at laplaza.org>; "Fred L. Drake, Jr." <fdrake at acm.org>
Cc: "Pydotorg-Redesign" <pydotorg-redesign at python.org>
Sent: Friday, April 11, 2003 3:25 PM
Subject: Re: [Pydotorg-redesign] pytdotorg-redesign notes from PyCon
OpenSpace session


> On Friday 11 April 2003 13:57, Kevin Altis wrote:
> > > From: Mats Wichmann
> > >
> > > At 10:54 AM 4/11/2003 -0400, Fred L. Drake, Jr. wrote:
> > >  >Mats Wichmann writes:
> > >  > > If people with /requests/ did the entering into the tracker it
> > >  > > might work.
> > >  >
> > >  >Of course, if there were a tracker that created issues from incoming
> > >  >emails, that would work like a charm.
> > >  >
> > >  >I'm certain this has been done before.  ;-)  I'm not familiar with
> > >  >what trackers are available that work like this, much less how to
set
> > >  >them up, though.
> > >
> > > I've seen it work fine with several systems
> > > if people follow a submission template. I
> > > suppose something could just snarf up a piece
> > > of email to webmaster and save it as the
> > > body of the issue without processing, but I
> > > think without enforcing a template it's going
> > > to be pretty tough for it to be useful. And
> > > I don't see any way to enforce a template,
> > > "webmaster" is like "postmaster", it's an
> > > address that's expected to be there and
> > > people just send to it, free-form.
> >
> > [Note that we probably shouldn't continue cross-posting to
> > marketing-python, so I've taken that list off of this reply.]
> >
> > A typical help desk system would have some stock email reply templates
with
> > FAQ answers, links to the FAQ page, tutor list, and other places to get
> > additional help. A web interface to the pending emails should allow any
of
> > the people on help desk duty to send a stock reply with a single click.
A
> > basic email can go out right away with some of the stock answers and an
> > issue tracking number.
> >
> > If the person who sent the original email requires a follow-up, the
issue
> > tracking number in the reply can be used to look at their original
> > question.
> >
> > If we expand the Help page http://www.python.org/Help.html to include a
> > form then a bulleted list of say the top 10 questions and answers can be
on
> > the page. We can try that for a few months and see whether it reduces
the
> > incoming email burden. To some extent that's what the current Help page
> > tries to do.
> >
> > Are the questions that come in actually answered on the Help page? Is
there
> > an online archive of old messages that can be scanned to look for common
> > questions...?
> >
> > Is there an off-the-shelf solution we can use? I don't know. Anyone want
to
> > volunteer to do a bit of research? We have to accomodate the straight
> > emails to help at python.org or webmaster at python.org... as well as
submissions
> > via a web page.
>
> AB Strakt has a helpdesk app that would probably handle this, programmed
in
> Python. It's possible they'd be willing to do something like this...
Someone
> would have to check with them - I'd suggest asking Laura Creighton
> <lac at strakt.com> about it.
>
> cordially,
> Anna Ravenscroft
>

Until we have something that stops us [webmaster at python.org] tripping up
over each other to do stuff and communicate with non-spam enquiries,
anything else is a YAGNI. The help-desk app might do this, I suppose.

I therefore propose that someone write a program that receives messages and
stores each one as a row in a relational database with status "unclaimed". A
CGI script can easily display all unclaimed messages, and another could
allow any user to claim a message. We could trust the community, Wiki-like,
to report their own name, or we could more paranoidly require
authentication.

Until the list of uncliamed messages grows beyond what's reasonable to view
in one page we might get away with very little else. I suppose a "completed"
indication would be nice, and possibly timestamps on both status changes.

Once you claim a message you don't have to own the task forever: simply
forward it to the input mailing address again, with your added comments to
assist the next claimer, and you're done.

Who will do this work? Some YoungTurk [I'd do it as an OldFart, but us
OldFart s get tired quickly, and must husband our resources to annoy mailing
lists].

The first part shouldn't be too difficult. If someone with a bit of time
could create a mail stream and put this in as a filter, all they otherwise
need is a MySQL database with the foillowing table defined in it:

+---------+---------+------+-----+---------+----------------+
| Field   | Type    | Null | Key | Default | Extra          |
+---------+---------+------+-----+---------+----------------+
| id      | int(11) |      | PRI | NULL    | auto_increment |
| msg     | text    | YES  |     | NULL    |                |
| claimed | int(11) | YES  |     | NULL    |                |
| handled | int(11) | YES  |     | NULL    |                |
+---------+---------+------+-----+---------+----------------+

Of course, what else the database might reasonably contain is the subject of
another YAGNI discussion, so I'm not planning to take any further part in
the debate until somebody (besides me) is succking these things into a
database.

--
Steve Holden                                  http://www.holdenweb.com/
How lucky am I?      http://www.google.com/search?q=Steve+Holden&btnI=1
Python Web Programming                 http://pydish.holdenweb.com/pwp/