Monday, December 17, 2007

How linden lost its Core

On December 11 there was a glitch in the Metaverse as Linden Lab crashed the core of Second Life: Corey Ondrejka had left the building. Immediately a stream of rumours began to appear in the blogosphere, good for 4.500 + hits on the search "Ondrejka + Fired" at Google within a week.


The official reading is that Corey thinks its time to look for a new challenge after a couple of hectic years at Linden Lab. The word on the street however says that he disagrees with the man upfront on the future of Second Life. This disagreement could lead to a departure on friendly terms, but judging from an email that popped up at Massively the general consencus is that he was fired:



Trying to sum up 7 years of work at Linden is an impossible task. All nighters at the Linden Street office. Gaining 20 pounds but then losing 70. Flying 350,000 miles on Linden travel. Recruiting and hiring many of you. Creating a programming language that now had 2.5 billion lines of code written in it – note to self, next time spend more than one night designing language. Changing the world.


It has been an absolute thrill working with all of you on Second Life. When Philip looked across a rickety card table in November of 2000 and told me that we would do more than build a great product, we needed to build a great company, too, I knew it would be a wild ride. Through the peaks and the valleys, Philip's ideas challenged and inspired me. They often led to solutions I would never have considered and helped to make Second Life what it is today.


I continue to believe in both Second Life and Linden Lab, but Philip and my visions for the future of Linden Lab are divergent enough that he decided to lead in his own way. While I will miss all of you, I have confidence in
engineering - in all of you - to adapt and excel going forward. You are a phenomenal collection of talents and I know that both Linden Lab and Second Life will be hugely successful.


Valleywag, which first reported on the separation on Tuesday, says that Ondrejka was fired over the same differences of opinion that Ondrejka mentions in his email (especially the last paragraph).


Philip Rosedale responded:



"I can confirm that Cory Ondrejka, CTO, will be leaving Linden Lab at the end of this year, in order to pursue new professional challenges outside the company. I wanted to take this opportunity to publicly thank Cory for his tremendous contribution to the company and to Second Life, in terms of its original vision and ongoing progress.


As it grows, the needs of our company are changing, and
the role of CTO, or technical lead, has also evolved. herefore, Cory and I are in agreement that our paths, at this point in time at least, lie in different directions.


During Cory's tenure the engineering team has grown tremendously, and given the breadth and depth of our technical expertise, we do not foresee any impact on our development plans. Together, we've produced great things in the development of Second Life, and I know Cory will go on to achieve excellence in his chosen field."


Which, according to massively is a basic formula of saying "we fired him". This gets fuelled after (again) massively obtains an internal Linden Lab email reading:




8:47pm EST] I have obtained a copy of Philip's email that he sent to his employees, which you can read after the jump


...Cory is going to leave LL. He has been with us for 7 years, and was the fourth person to join. So this is a big deal. Cory has been a huge part of the company, having
designed big parts of SL, hired many people, contributed greatly to the culture, and given a powerful voice to SL and LL. Among other things, he had the original design idea for the love machine, single handedly wrote the scripting language, and got us all doing A&Os back in 2001. Losing him will be hard for the company. I will miss him a lot.


What's worse is that ultimately his leaving is my decision.Cory and I have differences in how we think Linden should be run, differences that in the past few months have become irreconcilable. These are tensions that were more manageable when we were smaller, and there have been times that they have helped us do great work together. But now, as we change and grow as a company, I feel that we need a different set of strengths in engineering leadership.


I strongly believe that this is the right decision, although not without pain, for both LL and Cory. Of course, I'm not going to go into the details of these differences. This is one of those times when, in having me as your leader, you will also have to trust me in my decision. I will hold a brown bag as soon as possible to talk about this with anyone who would like, and will schedule time both in-world and in person here in San Francisco.



Please send any external questions you get about this change to Robin who will make sure they get answered.


Philip


What will be the effect of this sudden departure? The large majority reads it as a bad omen. Tateru Nino writes:


If you asked me this-morning, "Who can Linden Lab least afford to lose" my answer would have been simple: Cory Ondrejka and Robin Harper. At least (sitting on the outside here), every other member of staff seems to be replaceable.

I tend to disagree and go with the other reading, which was clearly voiced at Ambling in Second Life:


As a friend of mine put it, small companies need uber-hackers - they ignite the process, build innovative solutions and get you up and running quickly. I may be doing Cory a disservice, but it seems to me he fits into this category. However, once you have a large (and largely successful) implementation on your hands, your focus shifts from rapid innovation and heads more towards Quality of Service and effective service delivery. The skills for
this sort of role are quite different, and this may be where Rosedale has identified a key weakness. God knows, most of us are aware that this has been a key weakness!

Second Life has exploded over the last year, from 500K registered users in August 2006 to 11 million now. Scalability and Stability of the grid has been an issue and resulted in the Project Open Letter initiative, a call for more stability by hundreds of residents. Linden Lab has done a good job on these issues in the last six months though, as the grid capacity for concurrent logins has been doubled from 30 to 60K and with Havok 4 to be implemented on the main grid soon, the issue of stability will be largely addressed as well.


If this would do the trick for Second Life, it would put Corey into the driving seat. But when you put it into the perspective of Mark Lentczner's (Zero Linden) vision for Second Life, it falls short massively:


(Headlines from Zero Linden's office hours, posted by Dizzy Banjo on Soundtracking Virtual Worlds)



  • Linden are not just talking about the sim limits we have now - they are talking truly epic scale: "to evolve the SL architecture into something that is internet wide."

  • Transition to "SL2.0" is being designed to be as seamless as possible.

  • Now for the numbers: 60Million regions; 2Billion avatar accounts; maybe 50M to 100M on-line... though admittedly hypothetical

  • And "on-line might mean something more lightweight in the future"

To achieve this, it seems pretty obvious that coding Second Life needs a different approach. With Linden Lab slowly turning away from that cowboy TAO approach and slowly implementing a more structured approach to avoid the issues that have angered the crowd in the past there would be no room for a rogue-programming approach to Second Life.


Now the speculative part. Would Mark Lentcner be LL's new CTO? I think he would be a good choice, as he is actively communicating with the Second Life community he has credit there. The only problem is, what is it with Second Life and Eastern European names? (Ondrejka out, but Zdanowski and Letcner and whathaveyougot...)

Labels: , , , , , ,

2 Comments:

Blogger dyerbrookME said...

Re: "on-line might mean something more lightweight in the future"

That's not acceptable, since what Zero has talked about is having SL be nothing more than a browser with chat and an avatar that might be able to receive inventory to dress itself or move around in one area. That's not a world; it's a browser and a glorified chat room. If I wanted that, I'd just use Yahoo Messenger where I have an avatar and can play simultaneous interactive games with one other person.

I don't feel that Zero would make the best next CTO. What is needed is not one of these extremist hacker open-source zealots, but someone interested in methodical and grown-up boring work to make the platform stable.

I would note that everybody, including Cory, talked about Havoc 2 endlessly until it grew to the not-yet Havoc 4, but it took the outside hire of Sidewinder Linden, who came in and focused methodically on the job as a program manager, to make it come to fruition. More of this is needed on every single sector, including things that don't seem technical like billing, which is totally fubar'd now.

My bet is that the disagreement between Philip and Cory was indeed about the pace and scope of open-source, but it wasn't about rogue cowboy programming versus meat-and-potatoes stuff, because there are plenty of other rogues sitll left on staff to play the same role, and for that matter, Philip himself.

I think it is more about whether you open source and dump all property values, first and foremost, intellectual property, or not, and how to do that gracefully, when a new business model isn't in place. Cory would likely say, "Damn the torpedoes, full speed ahead!" to OS. Philip would be there in spirit, but is more conscious of the community and bottom line constraints.

Prokofy

Monday, December 17, 2007 7:00:00 PM  
Blogger VeeJay Burns said...

At least we do partly agree :)

I don't feel that Zero would make the best next CTO. What is needed is not one of these extremist hacker open-source zealots, but someone interested in methodical and grown-up boring work to make the platform stable.

I fully agree on the profile for next CTO. I'm not sure Zero is out for competition here, but then again, I'm not paying close attention to programming the world, so I can't be the judge there.

My bet is that the disagreement between Philip and Cory was indeed about the pace and scope of open-source, but it wasn't about rogue cowboy programming versus meat-and-potatoes stuff, because there are plenty of other rogues sitll left on staff to play the same role, and for that matter, Philip himself.


That's actually a pretty good bet I think, though it's hard to get confirmation which story is truth.

If the Open Source case stands, Corey's farewell might be bad, if it's the rogue programming scenario, it's a good thing. Confusing.

Monday, December 17, 2007 7:15:00 PM  

Post a Comment

Links to this post:

Create a Link

<< Home