[DotGNU]The DotGNU FIXIT at LCA2004

[Rhys Weatherley]

The DotGNU FIXIT was held at LCA2004 on Friday, January 16, 2004.
In this message, I am going to describe some of the things that were
discussed and their outcomes.  Representing DotGNU were Chris Smith,
Andrew Mitchell, and myself.  There were about half a dozen other
attendees.

The purpose of the "FIXIT" sessions at the conference was to identify
problems in a particular area, and potential solutions.  So, I started
the session with a quick 10 minute overview of where we are now and then
opened it up for comments from the room.

The first issue to be raised was HOWTO's and the initial impression.
When people visit our Web site, it isn't immediately apparent how to
get the software, install it, what it is good for, etc.  The lack of
Mac OS X packages and HOWTO's was identified as a particular problem here.
Mac OS X people are very interested in pnet, but simply don't know where
to start.  DGEE has similar "how do I get started" problems to pnet.

This is a very fair comment to make.  We have a lot of knowledge on how
to install, fix problems, etc, scattered throughout our mailing list
archives and the source trees, but we haven't bothered to collect it
up into HOWTO's on the Web site.

The second issue raised was about general advocacy.  We need to get a
lot better about getting the word out.  And once again, this has to
start with our Web site.  It does no good to announce something on
Slashdot or in .NET Magazine if users are still left bewildered when
they visit the site for more information.  There's a lot of talk about
"why" we are doing this (freedom, etc), but little on "how".

We also lack success stories.  Which doesn't necessarily mean customers.
It also means describing what will work out of the box so potential
end users know what they can do with it now.  A suggestion was made to
change pnetlib-status' report structure: instead of showing just what
is broken, also show what works!  e.g. "85% of System.Drawing is
implemented" or something.

One person commented that some GNU/Linux people are interested in
learning C#, but are not interested in using Microsoft's tools to do it.
We don't have any beginner tutorials for newbies to the C# language.
We can point people at the spec, but that is no substitute for a
well written tutorial, and then a pointer to the spec for more info.

There were a number of technical comments.  People would like to see
Winforms "look like Unix, not like Windows"; i.e. theming to fit in
with KDE/GNOME desktops.  There was a positive response to our plans
for a "Compact Framework for GNU/Linux handhelds".  And they are interested
in our language integration plans to allow other languages to run on
the engine, or be accessed via PInvoke callouts.

One attendee mentioned that a WinCE 3.0 version would be nice.
Apparently Microsoft didn't bother to backport their Compact Framework
to older versions of WinCE, leaving developers on those platforms out
in the cold when it comes to C#.  We have an opportunity to provide
infrastructure on commonly-used Microsoft platforms that Microsoft
themselves don't provide!

One of the questions I raised was whether we should put a lot of
effort into the Whidbey and Longhorn API's.  A response was that most
customers are using the .NET Framework 1.1 and it will be quite a
while before they bother to change.  So we should make that work in a
"just drop it in and it works" sense first.

In conclusion, the number 1 "FIXIT" that we have is to fix the Web site.
I didn't expect advocacy to be our biggest problem, but in hindsight,
the FIXIT attendees were right.  There's no point us being the best
.NET alternative if no one knows how to find us, install it, and use it.

Cheers,

Rhys.