Transcript of Roundtable with Andrew Linden

Very interesting talk. Thanks much for posting the transcript Iron.

Very nice. I feel so much more enlightend.
great work Iron, for making it happen.

about what? or are you joking? that would make sense.

Well, the bit about the quantum analogy, physics FPS, and interpolation resolution makes things much more clear for me.

I can now revisit my Bowling Lane and upcoming basketball game with a lot more insight.

Well that's good. Is that why you kept steering the confab back to physics. I suppose that's what we all need to start doing. Talk to Lindens about existing features. All we ever get back when we talk about future ones is "great idea," "yeah we've looked at that <insert possible technical problem>," and "it's on the table not on the list on the ceiling not on the board."

Now I understand why we always get that response. What I don't understand is why the Lindens have hidden behind the idea that people are working in solid dev teams on lists of features in priority order based on our input and votes as much as possible. That's simply not true. They just hack at features until they lock a release then graft the branches in until it's ready for release. The new dev routine should work a lot better for them. Understanding what their doing without the marketing facade will definitely be better for us.

Thanks, Iron, for the transcript. That was some great stuff!

As a new SL convert, and longtime software guy... I feel their pain of short-term/long-term tradeoffs, reworking first-architectures and backwards compatibility, API's, branching strategies, &c &c &c.

So, Lindens, if you read this: carry on, keep up the freakin' great work!

Snapshot of the discussion

Thank you so much for making this available. With your permission, I would like to include this material on http://history.secondserver.net

Great stuff! Thank you for taking the time to post it. It's the kind of information that the General forum was made for.

~Ulrika~

The admitted lack of structure in planning, development, and project management is not a surprise, but it is really quite disappointing.

LL is going to have to tighten up if they want to grow.

To tell the truth, that lack of structure is by design. We have deliberately tried to keep company culture flat and the development process design-light.

By "flat culture" I mean each employee bears most of the responsibility of prioritizing what are the important thing to work on each week, and we are all encouraged to participate in the conception/design/implementation/deployment of features we find important and interesting -- there is almost no middle management. In particular, there is no one whose sole job is to design SL and delegate the jobs to others. Of course some features and tasks get labeled as "high priority" and then a search goes out for the right person to get it done, but there is more overlap and less structure than a typical company of the same size.

By "design-light" I mean we try not spend a lot of time thinking about how we're going to implement some particular thing -- we get something working quickly, then evaluate it and change it as necessary. Smaller steps with more frequent evaluation rather than big ponderous steps. The theory is that despite a lot of pre-ponderance the step would not be in the most ideal direction and the smaller steps will get there faster.

The main advantage besides speed is that most everyone is usually working on something in which they are interested and/or passionate, so they have fun doing it.

The system does have its drawbacks. The first one that comes to mind is that such a development process is resistant to complete reconstruction of its sub parts. They happen, but typically not when they are first recognized as necessary, rather when the need is painfully clear. That this is a drawback is debatable since some might argue that this is a way of prioritizing the high nails -- waiting until the nails are clearly higher than the rest and then nailing them down.

The lack of a road-map is a characteristic of the system. It is no one's job yet to write the long detailed road-map into the distant future. For someone to embark on the responsibility would be counter to the system. Under the current system it could only be done in the form of a list of details that are suggestions and proposals, or as a vague outline of true generalities.

Growth is not our main problem at the moment. With growth comes changes in group dynamics ==> the challenge is to maintiain a flat culture and an agile development process in spite of growth. Will LL remain a fantastic and challenging place to work? Or will it become just another big, better than average, tech company?

I think this is great, in fact I think it is the ideal way to run a company working on a project with the leading edge nature of SL.

Because let's face it, you can not waterfall SecondLife, it has to be hyper-evolutionary.

Now that we realise this, I think going forward there will be less frustration in general as we understand that we're all apart of that evolutionary process and our lobbying will begin to reflect that.

How will it reflect that? I see a type of lobbying that involves less requests for specific pet features and more attempts at educating one another and the lindens about what's going on in terms of facts and figures regarding our present day SL experiences.

Out of that mix of enlightenment, hopefully the prioritization will become more obvious for everyone involved.