Rethinking UDS

[Stephan Hermann]

Hi Matt and others,

first of all, thanks for the summary of the UDS...as I'm still trying
to write my UDS 10.10 Wrap Up, some words from me.

A small reminder: UDS 10.10 was my second UDS in 5 years of Ubuntu
involvement. During these 5 years, the way of UDS changed totally.

In October 2005 we were only a "handful" of Ubuntu Developers (Canonical
Employed or Canonical Sponsored and some totally free of any strings),
some Ubuntu Fan boys/girls. We didn't have this session planning tool
(schedule.ubuntu.com) and everything was a bit chaotic. But after all,
we had a plan for Breezy and we wrote many specs and re-checked them
if they all had a form where people could just jump in and start
implementing some parts of those spec results.

In May 2010 we were gathering a lot more people from the Ubuntu
Developer Community, many people from external entities (ARM, NASA,
Linux Foundation etc.pp), and (that's my PoV) less Ubuntu Fan
Boys/Girls.
Everything was very business oriented, which means, the whole event
focused more on the business strategies of Canonical. (Which can be
good, but could also be bad, depends how someone will see this).

In May 2010 my colleague and I came to UDS, not because we wanted to
chat with people or drink some beers, but we came to listen, reading,
discussing the upcoming releases of Ubuntu, from a business user PoV
and from the datacenter front.

We are using Ubuntu on our Servers (here we talk about bare metal, not
(attention: buzzword) cloud or any virtualization systems. We wanted to
provide first-hand informations how we (and many many other datacenter
people) are doing their business, so that our development community can
think about top-enterprise-ready solutions for Ubuntu.

As preparation to UDS we catched up with the blueprints on Launchpad,
subscribed us and were keen to join the sessions. 

But during the very first day, we were already disappointed,
because some topics were changed to a different view.
Sadly, the outcome of those sessions was not interesting to us from our
enterprise PoV. Yes, it will be easy for people who are in need of
cheap infrastructure, but doesn't help people, like us, who are
providing the underlaying infra for the people who are in need of cheap
infrastructure.

It can be, that I do make a difference today between "Ubuntu, the OS
for the masses" and "Ubuntu, the OS for the enterprise business".

Other sessions were very important for us, and the outcome was very
high quality (e.g. the EHCache + Java Stack sessions). Sadly, still
the focus was more on the new hype named cloud and not as a generalized
solution for everyone.

That said, I did think during the UDS, that some topics were already
settled before the session started. Most of the time I had the
expression that the people who were discussing the topic do know their
job regarding the software, but not what matters in real business.

Actually I was missing the collaboration between Operations and
Development, as Flickr is proposing it, and how we are working in our
company. 
(Pointers:
 http://dev2ops.org/blog/2009/6/23/10-deploys-per-day-dev-and-ops-cooperation-at-flickr.html
 http://velocityconference.blip.tv/file/2284377/
 http://www.scribd.com/doc/16877392/10-Deploys-Per-Day-Dev-and-Ops-Cooperation-at-Flickr
)

Therefore, the outcome of most of the server sessions were quite good
regarding the "it needs to be easy for people to achieve something",
but "this is our supported enterprise solution for your goal when you
are using Ubuntu on Servers (or on the Amazon/Rackspace/Ubuntu Cloud".

That's the reason why we were disappointed, and during the
evening/night hours I talked to a lot to (Canonical) people about this.
And mostly there was only one statement: "Yes, we know, but this is
right now the direction of the Company".

Ok, business is business, and Ubuntu is business (although some will
disagree here), but I think for the future we need to make a split
between the "Desktop", the "Cloud" and the "Server".

We need a better overview what people are expecting from those sprint
plannings. Community helps a lot, but enterprises do help too, so at
some point we need to split those meetings into too...and we need to
listen more to enterprises or people who are using Ubuntu not only on
the desktop.

But there is one positive thing I took home: The most enjoyable talks
were after the official sessions, when you catch the right people at
the bar or outside while smoking. Those chats were important at least
for us.

In summary, there were 9 tracks, a too many sessions for too many
topics. Focusing on only 3-4 topics per track and trying to discuss way
of doing things for those 3-4 topics would be much more valuable at
least for me. Many sessions were follow-ups of others, but it could be
better to just increase the time for a session from 60mins to
eventually 2 or 2.5 hours to get more input and to have more output.

Regards,
\sh

PS: Flames etc. please forward it to > /dev/null
-- 
| Stephan '\sh' Hermann    | OSS Dev / SysAdmin         |
| JID: sh at linux-server.org | http://www.sourcecode.de/  | 
| GPG ID: 0xC098EFA8	   | http://leonov.tv/          |
| FP: 3D8B 5138 0852 DA7A B83F DCCB C189 E733 C098 EFA8 |