[occi-wg] How I tried to keep OCCI alive (and failed miserably)

Andre Merzky andre at merzky.net
Thu Jun 17 16:34:45 CDT 2010

Dear Sam,

apologies, but this will be my only mail on this topic.  I honestly
hope we can avoid a major flamewar.

Quoting [Sam Johnston] (Jun 17 2010):
> Date:    Thu, 17 Jun 2010 21:27:57 +0200
> From:    Sam Johnston <samj at samj.net>
> To:      occi-wg at ogf.org
> Subject: [occi-wg] Fwd: How I tried to keep OCCI alive (and failed miserably)
> I was going to let this one slide but following a [2]calumniatory
> missive to his "followers" by the [3]Open Cloud Computing
> Interface's self-proclaimed "Founder & Chair", Sun refugee
> [4]Thijs Metsch, I have little choice but to respond in my defense
> (particularly as "The Chairs" were [5]actively soliciting followup
> from others on-list in support).

The OCCI BoF was organized by Thijs and Ignacio:

Thijs is OCCI chair:

His mail to you did not mention his 'title'.

just saying ;-)

> Basically a debate came to a head that has been [6]brewing on- and
> off-list for months regarding the [7]Open Grid Forum (OGF)'s
> attempts to prevent me from licensing my own contributions
> (essentially the entire normative specification)

svn and mercurial logs attribute about 2/3 of the document contents
to you.  Mailman tells me that you generated about 20% of the mails
on this list.  Your contribution to OCCI is indeed substantial, and
obviously the single largest one.

On the other hand, for good or for worse, the current documents,
implementations, discussions etc would not exist had you been
working on it alone.  To claim that the spec *as is* is your own
(essentially), but to blame its shortcomings solely on others
(below) is simply getting things wrong.  

This *is* a community.  This is exactly the fact Thijs' mail, and
the endless threads before, have been about, and this is the point
I, personally, think you have difficulties to accept.  *This* OCCI
is *not* the product of the individual Sam Johnston.  

> under a permissive [8]Creative Commons license (as an additional
> option to the [9]restrictive OGF license)

Uhm, you are beating a dead horse here: despite our thread of close
to 100 mails your voice was the *only* one arguing for CC - no
single other group member, nor GFSG, nor the board, could follow
your arguments that OGF IPR is in fact inhibiting any valid use

Again, this is a community: there were ~10 explicit voices in the
thread to stick to OGF IPR rules, your single  voice against.  OGF
works consensus driven.  Period.  GFSG and OGF board discussed your
arguments, as we understood them, and tried to make the rules more
explicit were they were unclear.  

Yes, you can ignore the popular vote, and all arguments, all
questions, and all attempts to accomodate your requests.  But do you
honestly expect that people consider you a team player, with 'pure

> and/or submit them to the IETF as previously agreed and as
> [10]required by the OGF's own policies.  This was on the grounds
> that "Most existing cloud computing specifications are available
> under CC licenses and I don't want to give anyone any excuses to
> choose another standard over ours" and that the IETF has an
> excellent track record of producing high quality, interoperable,
> open specifications by way of a controlled yet open process. This
> should come as no surprise to those of you who know I am and will
> always be a huge supporter of open cloud, open source and open
> standards.

"required by the OGF's own policies": No, its is not.  For one,
GFD.1 is superceded by GFD.152.  Also, GFD.1 explains that OGF is
modeled after IETF, in terms of process, but not that GFDs are to be
submitted to IETF.  

I actually think, however, that submitting OCCI to IETF is an great
proposal.  I am not sure about the technical details though: for
example, it would not make sense (to me) to have two incompatible
specifications under the same name.  Anyway, OGF has strong ties
into the IETF community, and I am sure a way can be found to
accomodate the specific use case.  It has been done before IIRC.
That is my personal opinion though, not that of OGF or GFSG.  In any
case, submitting to IETF for comments will not affect OGF IPR.

> The OGF process had failed to deliver after over 12 months of
> deadline extensions

OGF does not deliver - people do.  OGF is a volonteer army (yes,
suprise: you are not the only volonteer here).  My (probably
limited) experience shows that groups which meet often face to face,
and have a good discussion culture, progress most efficiently.
Uhm... ;-)

> - the current spec is frozen in an incomplete state (lacking
> critical features like collections, search, billing, security, etc.) as
> a result of being prematurely pushed into public comment, nobody is
> happy with it (including myself), the community has all but dissipated
> (except for a few hard core supporters, previously including myself)
> and software purporting to implement it actually implements something
> completely different altogether ([11]see for yourself).

Again: you claim the OCCI spec is 'essentially' yours, but all its
failings are OGF-made?   Strange...  ;-)

Implementations: OCCI is not even a published specification yet.
Give it some time.  AFAICS, the interest to implement (and also to
use) OCCI is still strong.

> There was no light at the end of the tunnel and with both
> [12]OGF29 and [13]IETF78 just around the corner I yesterday took a
> desperate gamble to keep OCCI alive (as a CC-licensed spec, an
> IETF Internet-Draft or both).
> I confirmed that I was well within my rights to revoke any copyright,
> trademark and other rights previously granted (apparently it was
> amateur hour as OGF had failed to obtain an irrevocable license from me
> for my contributions) and volunteered to do so if restrictions on reuse
> by others weren't lifted and/or the specification submitted to the IETF
> process as agreed and required by their own policies.

Uhm, you were trying to keep OCCI alive by revoking your granted
rights?  "Either you do as I say or else..."?  My son does that,
too.  He is 7.  We are concerned actually, he should be over that
phase ... ;-)

Sam, face it: you are not a team player.  Your are technically
brillant, but it is close to impossible to work *together* like
this.  It works only for you, alone.

> Thijs' colleague (and quite probably his boss at [14]Platform
> Computing), Christopher Smith (who doubles as OGF's outgoing VP of
> Standards) 

For the records: Chris is not Thijs boss.

> promptly responded, questioning my motives (which I can assure you
> are pure) and issuing a terse legal threat about how the "OGF will
> protect its rights" (against me over my own contributions no
> less). Thijs then followed up shortly after saying that they "see
> the secretary position as vacant from now on" 

Please do understand that this is not a personal vendetta of Thijs
or Chris or anybody else.  If you want, its a conspiracy: my mail
count on that topic, i.e., on *YOU*, is 140.  Over the last 10 days.
I was not on all threads (fortunately).    There have been multiple
calls over the last days, to conspire of course.

As Thijs said: "This is not an action we take lightly".  Believe
him.  Stop attacking him personally.  Attack OGF if you must.  Its a
community, you know? ;-)

> and despite [15]claims to the contrary I really couldn't give a
> rats arse about a title bestowed upon me by a past-its-prime
> organisation struggling (and failing I might add) to maintain
> relevance.

Uhm, why *did* you pick OGF actually?  This is a serious question.
Would any other *community* have worked out any better?

Also serious: you may want to consider writing specs on your own,
completely, and submitting them as individual.  It might well be
that you are more productive that way, and that the specs improve.
Hard to tell in advanvce of course, just an idea.

In any case: you all *did* pick OGF, and its rules.  Next time,
please think about it *before* you join a community.

> My only concern is that OCCI have a good home and if anything
> Platform have just captured the sort of control over it as VMware
> enjoy over DMTF/vCloud, with Thijs being the only remaining active
> editor.

Uhm, Sam, I hate to tell you, but the fact that the two persons (out
of ~10?) you're having trouble with are from Platform does not mean
OGF is controled by Platform.  OGF is a volonteer army.  Everybody
can nominate anybody as chair or AD.  Positions are hold by
consensus of the community.  I don't have statostics about chairs at
hand, but the AD affiliation is quite diverse:

It is as often with your mails: lots of effort to distinguish
ranting from arguments :-(  Please use twitter for ranting, and this
list for discussions.

> I thought that would be the end of it and had planned to let
> sleeping dogs lie until today's disgraceful, childish, coordinated
> and most of all completely unnecessary [16]attack on an unpaid
> volunteer that rambled about "constructive technical debate" and
> "community driven consensus", thanking me for my "meaningful
> contributions" but then calling on others to take up the
> pitchforks by "welcom[ing] any comments on this statement" on- or
> off-list.

Sam, its not an attack.  It is a attempt to get the group functional
again, to avoid deadlocks, and to avoid decisions not supported by
group consensus.

Likely, this sounds like the very same rambling to you, sorry for
that - we see meaning in those points.

> The attacks then continued on Twitter with another [17]OGF
> official claiming that this "was a consensus decision within a
> group of, say, 20+ active and many many (300+) passive
> participants" (despite this being the first any of us had heard of
> it) 

Well, I can support the cited statement.  That makes 4.  Keep
counting ;-)

> and then calling my claims of copyright ownership "genuine
> bullshit" and report of an implementor instantly pulling out
> because they (and I quote) "can't implement something if things
> are not stable" a "damn lie", claiming I was "pissed" and should
> "get over it and stop crying" (needless to say they were promptly
> blocked).

Yes, we are concerned about that, too.  OCCI needs to be stable.

> Anyway as you can see there's more to it than Thijs' diatribe 

Again, please stop attacking Thijs - he is the messenger.

> would have you believe and so far as I'm concerned OCCI, at least
> in it's current form, is long since dead. 
> I'm undecided as to whether to revoke OGF's licenses at this time
> but it probably doesn't matter as they agree I retain the
> copyrights and I think their chance of success is negligible -
> nobody in their right mind would implement the product of such a
> dysfunctional group and those who already did have long since
> [18]found [19]alternatives.

Alternatives are healthy.  "The nice thing about standards is that
you have so many to choose from." (A.S. Tanenbaum).  

But really: in our opinion OCCI very much alive and kicking, and we
hope it stays that way.  But I am known to be a terrible
fortuneteller :-P

> That's not to say the specification won't live on in another form
> but now the OGF have decided to go nuclear it's going to have to
> be in a more appropriate forum - one that furthers the standard
> rather than constantly holding it back.  Protip: None of this has
> anything to do with my current employer so let's keep it that way.

Sam, honestly: I wish you all the best.  Please reconsider your
actions: the OGF OCCI group would very much welcome your continued
input!  Honestly!  Really!  Really really!  We would like to do
things our own way tough: consensus driven.  That is about the only
condition we have, really.

Thanks, Andre.

Nothing is ever easy.

More information about the occi-wg mailing list