[Portland] Summarize current plan?

[Bryce Harrington]

On Sun, Mar 12, 2006 at 03:38:52PM -0600, Jeremy White wrote:
> > If I recall correctly, Jeremy posted a glowing endorsement of the
> > Portland project here for use in John's email, but
> > his comments were about the "script to invoke
> > the email client" use case.  I don't think it would be right
> > to use his comment until some Portland deliverable addresses that in a
> > way he can use.
> 
> Hmm.  No, I actually was fairly cheerful about the
> glowing comment; the Portland Project has helped me in
> a very direct way.
> 
> Of course, it was the analysis/documentation phase that
> provided direct help to me; I'll gleefully point that out
> to folks that are rushing to a new implementation rather
> than working on documentation.
> 
> This is an area, though, where I think OSDL could
> make a huge difference.  We're mostly volunteers,
> and rational volunteers want to voluteer to do something
> fun (like implement dapi).
> 
> But if OSDL can *pay* someone to do it... <grin>

I always wince when I hear people say this.  ;-)

However, I'm more than happy to help with docs.  Just for the record,
OSDL's having me put in about 10% of my time, to split between the
developer portal, and to help organize making a test suite for the
Portland project.  The former is on pause waiting for feedback, and it's
probably a bit early still for the latter, but I plan to put some time
in it when I get back next month.


I have a couple other observations related to OSDL funding for work on
Portland.  I think often it's assumed OSDL has a huge coffer of money
that is thrown semi-randomly at projects.  Obviously, in reality this is
definitely not the case, else OSDL would have disappeared in a puff of
improbability four years ago.  ;-)

OSDL invests efforts in areas where it gives value to its member
companies.  That's why companies become sponsors.  In some cases, this
is very direct - member company A wants to see work done on code B, so
an engineer gets assigned to work on delivering that.  However, for many
things companies need, it's not as simple as that; usually true success
requires a cooperative community effort, and trying to work around that
just leads to extreme grumpiness all around.

In this latter situation, money alone isn't enough.  You have to gain
community concensus, and you need community participation in the
implementation as well.  We've also learned that while
analysis/requirements documents can be useful, they are not a good fit
for how FOSS actually works.  You might recall the DTL Capabilities
Document from early last year.  In that effort, OSDL *did* pay several
engineers to work on it, along with people from member companies.  Yet
we failed to get community participation in the process, and when it was
finally released, the only thing it generated was funny looks. ;-)


This time around, the approach being taken is fascinatingly different.
OSDL is taking a more participatory role, and focusing its efforts on
*facilitating* the community, with the hope that things can be figured
out and solved bottom-up.  Resources have focused on helping the
community organize, meet, plan, and communicate, and on encouraging
companies to become part of the community as well.

OSDL's primary motivation is to help Linux become a good solution for
enterprise desktops, as this gives value to OSDL's members.  But I think
an underlying motivation could be that by assisting the community to be
successful at what *it* chooses to do, companies will see these
successes and increase their membership levels in OSDL - which would
mean more resources for the community.

I've seen this particular model work firsthand with the NFSv4 testing
project.  We helped the existing community figure out testing plans and
how to get distros to test and include NFSv4, and so forth.  These
successes resulted in at least one company increasing membership from
bronze to silver, plus another $85k in donations for equipment, travel,
and a new hire to further support that community's efforts.  I'd bet we
can achieve similar sorts of successes with Portland.  (I hope it does
prove highly successful, because this is much more enjoyable than the
top-down driven model.)

So anyway, this is why it's exciting to see the prototyping work under
way, and why I think we hope the press release John is putting together
will be very well received.  :-)

Bryce