March 04, 2005
Being BileBlogged

I've been "BileBloged". This is my bile blog back. :)

Getting "biled" is an honor. It means Hani Suleiman has dedicated one of his
BileBlog entries to me, bitching and moaning as only he can. In this case
he's biling about a talk I gave on open source at the TSS Symposium in Vegas
this afternoon.

But sadly -- instead of a good healthy lampooning of my actual beliefs and
words -- Hani put words in my mouth. That's just not right. I've gotta bile

Here's what Hani said:

> TSSS Day 2 open source from the inside This talk was fairly disturbing. The
> whole thing seems to be one big giant apache fapfest as conducted by none
> other than captain one hit wonder, Jason Hunter.

The talk was "Open Source from the Inside". It's about why people do things,
based on my Apache, JDOM, and experiences. It's about helping
people curious about open source understand what makes it tick at an
individual contributor level.

> Jason is, it turns out, a deranged opensores lunatic. He's not just drunk the
> koolaid, he sounds like he's been receiving koolaid enemas as well as
> injecting it intravenously for quite a while.
> For example, he thinks that it's 'cowardly' for companies to consider their
> source to be part of their IP, and if they 'really wanted to compete', they'd
> hand over the source and then compete!

Ah, the fun you can have with quote marks. You quote a word out of context,
then a short phrase, and you can imply I said what you want me to say. What
Hani attributed to me was the exact opposite of what I believe. Here's my
actual point:

There's a powerful dynamic that applies to leaders of open source projects
due to the threat of external forking. It forces leaders into benevolence
because anyone at any time can take the same source code you have and
compete with you. Imagine if corporate companies had to work that way.
How responsive they'd be! Imagine if companies couldn't use patents and
copyright law to retain customers. That's what open source projects live

> He then talks through the trials, tribulations and motivation behind writing
> JDOM, and as expected, it basically boils down to 'it makes me look cool', so
> take that Cedric, more ego-driven opensource! The anecdotes are all about
> going to interviews and saying 'I wrote the book! I wrote the api you're
> using!'

Now that's a flat out lie. That was the punch line of a joke, and it got
laughs. The actual point I made was "you share your tools so that others
might sharpen them". You solve a problem initially because it helps you, then
you release it because it's no skin off your back and others might contribute
changes that help you later. I said that's the primary reason I did JDOM.
Then I said it doesn't hurt with job interviews either.

> To his credit, he does proclaim that the GPL is the ultimate evil, with its
> sick 'fuck you all! This is MY code and you can't do diddly squat about it!'

God, I never said and I don't believe GPL is the ultimate evil. What I said
is that the difference between BSD people and GPL people is BSD people see
reinventing the wheel as the ultimate evil while GPL people see someone
selling you a product that contains your own code as the ultimate evil. I
said I fall on the reinventing the wheel is evil side. But that doesn't make
GPL the ultimate evil.

> To be fair, the talk is a good overview of some of the issues surrounding
> opensores and the various licenses and models involved, even if it is from the
> one of the enema eager crowd.

Darn, Hani got that part right. :)

> Unsurprisingly, it all goes horribly wrong in the end. Jason totally loses it
> and puts up a slide of 'guidelines for open source projects'. This list has
> absolutely nothing to do with open source, and it's pretty offensive how he's
> managed to hijack a bunch of common sense guidelines (use a source repository!
> Have mailing lists! Use email instead of meetings!) and somehow make them open
> source specific issues rather than standard stuff any project should employ.

There's a lot of good ideas to be learned from open source. Some that Hani
conveniently missed mentioning but which I know from experience are sorely
needed at companies:

* Make each commit small and a fix to just one thing. Don't check in all of
last week's work as a whole.

* Fully document every commit. Don't say, "Bug fix". Say what bug and how
you fixed it and how app behavior might change because of the fix.

* Use mailing lists to track the commit messages as well as the debate
beforehand. Have 20% of your engineers work at home each day until email
based communication has replaced meetings.

* Encourage all interested parties within the company, regardless of group, to
subscribe to the mailing lists. Don't say it's none of their business.

* Use the commit archive as the basis for detailed release notes. Don't use
engineer's memories as almost all companies do.

I had ten like this. Hani, if you think these concepts are well understood
and used in corporate America, then you may want to look at yourself for that

> The delusions continue in full force shortly thereafter, where Jason proclaims
> that there are more talks about Jakarta projects than JCP ones.

I didn't say, "Jakarta projects" I said "open source" projects. Hani, you're
trying to make me into an Apache zealot, but you have to put words in my mouth
to do it.

> I don't know what conference you're at Jason

Hani, I specifically quoted in the talk which conference I was talking about
-- the No Fluff Just Stuff conference. And let's take a look. Look at the
next show agenda page you'll see the three promotional talks are on: "SWT",
"EJB, JDO, and Hibernate", and "Tiles and SiteMesh". We're three for three in
talking about open source projects.

> but the only Jakarta project that has its own
> talk is Tapestry, which all of 3 people use. Contrast that with a talk about
> the JCP, twenty nine EJB3 talks, a JSR168 BOF, JSR208, JSF, and a smattering
> of random open source crap that is most definitely not in Apache (always by
> deliberate choice, given how broken the Apache committee and voting schemes
> are).

This conference has more corporate involvement and company-provided speakers
than NFJS.

> Some more drug induced gibberish comes out later, where we're told that the
> JCP should accept every proposal submitted to it, without any consideration to

> existing solutions that might already fulfill that need.

Bzzt! Gutter ball again. Game's almost over.

My point was that the JCP has no checks and balances, and once someone puts a
flag in the ground promising to work on an area, no one else can put a flag
close to it to compete. Even if the company is sitting on the JSR and not
letting it progress to community review, you can't within the JCP propose an
alternative. You can vote out the lead person, but the lead company remains.
It's a terrible roadblock. There's no internal forking allowed, so that's why
you see innovation outside the JCP.

> Basically Jason
> thinks that the real world is insane enough to build an industry around the
> zoolike loonybin model that is apache.

That's what I believe, huh? It's news to me. And Hani should know better
since he was there for the slide where I talked about how people are happy
with the JCP and how I didn't see anyone wanting an open source style (aka
Apache style) decision making process for Java standards. What I said people
wanted was more openness in the JCP, not a model where if you propose a JSR
you get to own that space within the JCP and permanently retain dictatorial
control over the spec without any checks or balances.

> Really, I must say I was quite surprised. You'd think that at a talk like
> this, it'd be a lot more even handed and representative of 'the other
> viewpoint', rather than how childishly naive and simplistic it actually turned
> out to be. (2005-03-04 19:07:42.0)

I'm quite surprised too. You'd think that in a review like this -- even a
BileBlog review -- it'd be a lot more representative of what really transpired
rather than the overly simplistic one it turned out to be.

Ooh, this "biling" is fun. :)

But something tells me I'm going to be "BileBlogged" a lot more now.

Posted by Jason Hunter at March 04, 2005 07:03 PM
To be notified when new weblog entries are posted, click here.