user support (was Re: [mapserver-users] AGG_Q + raster problem)

[Frank Warmerdam]

Brent Fraser wrote:
> Frank (and all),
> 
>   Will do (as soon as we narrow it down).
> 
>   This problem got me thinking of ways to improve our
> current support methodology.  I've only asked and answered
> questions on this mailing list. Are there other methods
> being used (reading the doc, irc, paid support, etc)?

Brent,

I certainly *hope* there are people reading the docs!  I do think a
"debugging mapserver problems" howto would be helpful, but of course
only if people read it.

There are people who come to IRC for questions, and it is more
efficient for back-and-forth debugging sort of stuff.  One downside
to IRC is few eyes - there are usually only a half dozen people in
#mapserver who answer questions and their expertise is by no means
comprehensive.  Another downside is no permanent record that are
easily searched by new people with problems.  IRC plug:

   irc://irc.freenode.net/#mapserver

There are certainly paid support arrangements.  I think DM Solutions
"Premiere Support" is one of the few "standard price list" type
support arrangements.  Lots and lots of folks will provide support
as part of a larger service arrangement.  Some folks will also provide
support on an hourly basis.  The OSGeo Service Provider Directory is
a good way of identifying organizations claiming to offer some sort
of MapServer support and expertise but exactly what they offer varies
quite a bit and it can take time to narrow down on a likely service
provider.

   http://www.osgeo.org/spd_help

I certainly have concerns about using this list for a great deal
of "back and forth questions" style debugging of problems.  With
around 2400 subscribers every question and answer is hitting a lot
of mailboxes.  Is it really a useful approach?  To some extent it
is good because it gives others a chance to chime in at different
points with suggestions, and hopefully some people on the list
are picking up hints and new knowledge that helps them in the future.
But ... well, I often prefer to take debugging session off-list with
the intent of summarizing final findings to the list. I do this
despite the prevailing KEEP IT ON THE LIST! mantra of some of our
leading contributors.

> Is this list an effective way of solving user problems?
> Should/can we improve it?  How?
> Perhaps a Summer-Of-Code opportunity?

Summer-of-code is very software development oriented.  For instance,
a SoC effort cannot be primarily focused on writing documents, providing
support, etc, except as a by product of software development.  So I don't
think this problem is very much a SoC solvable one.

If we ever put a MapServer sponsorship program into place, and were able
to use it to hire a paid maintainer of some sort, I'd like to suggest
some of their time be spent providing support to but of course even this
would still likely only be a drop in the bucket.

Ultimately, I think we need to create better documentation, especially
starter and troubleshooting documentation so folks can "self serve" through
the standard problems, and just fallback to other people for the harder
stuff.  Helping to write such documentation might well be something a paid
maintainer could effectively do.

I will say that user to user support is very important, and scales up better
than user to core developer support.  That's why it is so valuable when you
and other super users (or even just average users with experience in some
stuff) step up to the plate.  I think this has been a notable success of
the MapServer community though there is always room for improvements.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | President OSGeo, http://osgeo.org