The Tao of Linden

Linden Lab has a different and better way of doing work.  It relies on the idea that if the level of transparency (everyone can see what everyone else is doing) can be made high enough, you can stop managing people by explicit authority or delegation.  Instead of being told what to do, you choose your own work by listening to your peers, making good strategic judgements, taking risks, and surfing a huge amount of information. 

As one of a very few pioneers in doing things this way, there is still a lot to learn.  We make mistakes.  But the things we have gotten right are impressive: Linden Lab has, in 6 years, had almost zero employee turnover, and our productivity, in comparison to other similar sized teams, is off the charts. 

Here is the first page of our internal handbook, designed to serve as an introduction to new team members:

TAO OF LINDEN LAB

Vision

To expand the human experience by building an online world allowing people to interact, communicate, and collaborate in a revolutionary way.

Company Principles

Work together! The problems we have faced and will face in creating Second Life are generally larger than one person can solve, and solving them is one of the defensible strengths we have as a company. We will succeed only if we collaborate with each other extensively and well. This means helping others reach their goals, joining teams, and being easy to work with.

Choose your Own Work. Given how dynamic the challenges are we face, and the opportunity for increased job satisfaction and productivity, Linden Lab places a high premium on choosing your own work, rather than being told by anyone what to do. By choosing your own work, you are more likely to have more fun at work and add more value to Second Life. By setting your own goals, you are more likely to meet them. When you commit to a team project, be prepared to be directed by the team lead, but when deciding what project you will take on next, rely first on your own best intuition and the counsel of peers.

Be Transparent and Open There are many ways to emphasize responsibility, accountability, communication and trust. We believe that the one key principle that best supports all of these values is transparency. As much as possible, tell everyone what you are doing. This transparency makes us responsible to our peers, makes us accountable to our own statements, and replaces the need for management with individual responsibility. Over time, it creates and reinforces trust. Be willing to share ideas before you feel they are ‘baked’. Report on your own progress frequently and to everyone.

Make Weekly Progress We believe that every person should make specific, visible individual contributions that moves the company forward every week. Projects must be broken down into measurable tasks so that making weekly progress is possible. This is a principle that almost no one believes is true when they first hear it, yet everyone who keeps to this principle over the course of several months is stunned by the amount of progress made during that time. Set weekly goals and report progress to everyone.

No Politics! Never act to advance your own interests at the expense of the interests of the company. This is the one principle, outside of violations of law, for which violation will likely result in immediate termination.

Might Makes Right Just kidding – wanted to make sure you’re still paying attention. Lots of things could be said here: Have a sense of humor. Have a sense of humility. Have fun. Call out inconsistency in principles when you see it. Don’t let a staid form and function become routine and boilerplate. Which leads to our last principle . . .

Do It With Style It’s not enough that we want to change the world. It’s not enough that our product is incredibly complex and our vision is vast and shifting. We’re not just going to win, we’re going to do it with style. That means a lot of different things, and a lot of what it means can’t be captured in a handbook. Find out by talking to your colleagues, by living the principles above, by exploring Second Life.

Welcome to Linden Lab.

Comments

[Ingrid Ingersoll]

If anyone would like to come visit my workplace, you are most welcome. Recently the CEO threw a huge phone directory at our IT guy. He then flipped the Controller the bird. It was all very charming.

And unlike LL, we don't get a morning visit from the bagel man, even though we probably have better bagles in Montreal.

[Michi Lumin]

Yes, Philip - and oh, the many of us who wish we could work there...

Kind of a throwback to the 'dotcom days', at some of the places I worked at from 1998-2001, I got to see and actually experience dev and business environments like this. In some ways they were unnerving, in others, they allowed for things to happen that normally wouldn't.

Though, soon after the dotcom crash, the old ways took over, and now in most places outside of the Bay Area, micromanagement, threats, and extreme high pressure along with "productivity metrics" are in full force once again.

I'm sure this post won't "help" your already-continuous barrage of Resident Applicants, Philip.

Which method is more effective for development? I don't know. It probably isn't the method, as much as it is prioritization.

To say that SL is problem-free would be a folly, in the least.

However, I've been in dev houses with rigid structure and rules (Government software dev, anyone?) and the productivity was an absolute joke.

[Caliandris]

I am astonished that you think this is an appropriate time to tell the customers how happy all the Lindens are. I'm NOT happy, and the reason is that I am having to deal with a lot of bugs which were introduced with the latest build. Seems like you don't care about that as long as things are peaceful in the office.

I know you are all used to people banging on in the forums about the changes and bugs after an update, and so maybe you have become immune over the months, but the level of difficulty which has been introduced far outstrips anything I have experienced before in SL. Texturing is infuriating, painful and frustrating, and I for one don't propose to do any more until you fix the texturing tools.

Seems to me that the "choose your work" ethos works just fine if people are choosing to do the things that need doing. But from the customer's point of view those things don't look like they are happening in an organised manner. As I blogged in SLI this morning:

Thus people who are in world and trying their best to build, are dealing with known issues, bugs which can be exploited and so are kept secret, bugs which are known about but haven't been put on the list, intended changes which haven't been documented, intended changes which are deliberately kept secret, unintended changes which just sort of ... happened. Some things, it is obvious how they are meant to work now, but they don't. For example, the changes to the way the mini map choices work. There is a search function for landmarks, instead of the alphabetical list. Except the search doesn't work. Hmmm. Well done.

I'm glad you're happy, glad the Lindens are happy. People are telling you in words of one syllable that your customers are NOT happy. Do something about it, please, don't post about how great things are for you.

[FlipperPA Peregrine]

While I often cringe a bit here and there at some of the things that go un-fixed due to this Haight/Ashbury model of business, I remind myself just how far SL has come in the three years I've been around. While I think that SL might be a little more stable with a more traditional business model, I think we'd have the feature set of about v1.2 right now if LL wasn't constantly plowing forward. The only complaint I have is that there are times Linden Lab makes the wrong decision, but for some reason, refuses to correct their course (IE: lack of registration verification) until the proper time makes it the right decision.

Quoting from this thread: http://forums.secondlife.com/showthread.php?t=86249

extensive library of stock content
help island
greeter program
frisch
private islands with extended estate controls
custom animations
added tube prim
added ring prim
light prim
full bright
added dozens of plants
improved water rendering
improved sim performance
virtually eliminated ghosting bug
L$=USD insentive
eliminated land tax
eliminated prim tax (imagine needing 2000L$ in the bank just to rez an av attachment?!)
more primtorture controls
shiney!
eliminated scarcity from the land market.
p2p teleport is back and this time it's free
music streaming
movie streaming
particles
animated attachments
shadows
new land textures,
new avatar mesh,
preview for taking pictures AND importing them into world
visual preview for importing clothes,
email via script,
sit targets on prims,
(even if it is only half finished) XML-RPC,
spacial improvements to sound,
private islands,
ban/access lists on parcels,
join and subdivide plots, was always available
improved find tools,
classifieds,
MUCH improved tporting,
vehicles
"my picks"
new welcome area
improved map
land for sale info in map
snap to grid
local and reference grids
New terraform tools!
Animated hair/clothes!
Bump mapping!
Removing recall!
llDialog!
Prim count per land owned in sim! (Bye bye, prim hogs!)
Inventory search!
Multiple inventory windows!
Multiple select!
Snapshot / postcard preview!
Group land ownership!
Freeze / eject!
IRC style chat emotes! (/me does something)
Revamped find directory!
Email to IM!
XML-RPC!
Localized sound!
Mute without restart!
Custom gestures in inventory!
Separate library for common items!
1 degree rotation increments!
Rotation snap!
Snap to grid!
Event notifications!
Land for the landless!
llGetNotecardLine - notecard reading
llGetNumberOfNotecardLines
llSetPrimitiveParams - fine prim manipulation
dataserver event - lots of world data
hexidecimal integer input - YAY for self-documenting code
llOwnerSay - less spammy attachments and debug
llParticleSystem - PARTICLES
llLog - Faster math
llLog10
working email event - objectwww communication, allows for GOM
XML-RPC - more communication
llSetParcelMusicURL - Streaming music
llParcelMediaCommandList - Streaming VIDEO!
llGetFreeMemory - Less crashy scripts
llRezAtRoot - More accurate object rezzing
llGetAttached
llListReplaceList
llSitTarget - Precise sit position
llAvatarOnSitTarget - everything that tells you to sit on it relies on this
llSetSitText - eye candy
llSetTouchText
llSetCameraEyeOffset - Camera control functions!
llSetCameraAtOffset
llUnSit - Dont steal my car!
llGetAgentSize - Psuedo-accurate avatar height correction algorythims for animations.
llGetBoundingBox
llGetObjectMass - Accurate object-object physics calculations
llRequestAgentData - Names from keys, online status, & more
llRequestSimulatorData - Sim corner from name, current status, & more
llGetRegionName - w00t, finally
llAddToLandPassList
llEjectFromLand - Ban avatars without pushing them cross the sim.
llDialog - Easier user interface
llSameGroup - Group-specific scripts.
llGetAnimationList - Animation overriders!

...and that list is far from complete. :-)

[Caliandris]

Are you building anything at the moment, Flipper? Are you talking to people who are trying? Isn't the fact that every folder in the texture picker is open when making clothes driving your partner crazy?

I love SL, and I spend a lot of time on it. I haven't complained every time there was an update. Something needs to happen. You can tell me the sun shines out of Linden Lab from now until Christmas, and what a wonderful platform it is, it won't help unless the stuff that's borked gets fixed.

We have been promised for months that bugs will be fixed before features are introduced, and it seems to me that it has been untrue. The feature voting system showed that people wanted Linden Labs to stop adding features until the bugs were corrected and they removed that from the list because it was going to happen, but it certainly doesn't seem to me that it is.

I am still seeing bugs which I reported months ago. PLUS all the new ones.

Even if they are the best damn company in the world to work for it STILL doesn't excuse what's happening at the moment.

[Siggy]

As with a lot of things in SL - there is often an either/or sort of approach by both users and developers (eg: you must fix bugs-OR-add features... we will Teleport Direct OR use telehubs), when in a lot of cases there is room for both.

There have been several updates when there has been a pleasant ratio of new features to bug fixes, and in those releases it seems that those few new features have been implimented fairly well and painlessly.

Alas we have had more releases overpacked with shiney new features (some that actually take a giant leap backward in useablility) and bugs galore - with all the comfort of a dog shitting razor blades.

The problem with choosing your own work is that a lot of people will choose the 'glory work' - the flash new feature that will wow their peers. Once the wow is gone its shelved... sometimes unresolved, and implimented half arsed and unfinished.. theres no fun and glory in doing the grunt work.

But I think there is room for BOTH - do the grunt and get rewarded, and SL will be a better place for it.

Since I've joined in 1.0 I've seen the interactive benefits of having your AV rotated to an point to an arbitary location or rotation - there was a LSL call for it too... it never worked and was labled as never working.. I think its been removed now - something as simple as moving then turning to face an object or person cannot be automated.

Since 1.4 with the advent of custom animations we've relied on animation overriders to replace our own default animations. You'll find them as attaches, HUD's, accessories, as features, or required scripts for some custom avatars - they are less than optimally written, and yet a very large portion of SL has at least one in some incarnation running at any given time.
We are incapable of overriding them via the UI - something that you would think would be a very integral part of the implimentation.

Camera Control - I can't rotate the camera and moving it to simulate a crane to to film myself (a la machinima - one of touted uses)... my expermients have shown I'm better off with a second person just looking at me.

We have this great tool - the voting tool - so developers can look and see what the people using the system (and in this case much more - making your world) think is important...

I see a lot of lovely shiney things going in - but not many that you'll actually find on the voting system at all... I highly doubt it's even looked at.

I think thats why some folks feel frustrated - I can't blame folks for thinking that they are ignored, because for all the press - joe average pretty much sees it as a 'FU we'll work on what we damn well please'... and for that, I can't fault their frustration.

In the past there has a been several releases with a nice balance between new features/old issues/bug fixes. They are few, but I think thats a goal worth striving for.

Theres less glory in the grunt work, but there is a lot more appreciation.

[Jeremy Vaught]

Human nature says that when a bunch of people get together, they start to complain. History tells us that when making history, mistakes are made. I suppose it is a good thing that Thomas Edison didn’t have an audience questioning his every move, perhaps he would never have completed the world changing inventions that did indeed changed the world. Personally, this is where I see Linden Labs. Making history. If not world changing history, then close to it.

[Siggy]

Well keep in mind that when Edisons bulb didn't work properly - he fixed it (with help) - before he went on to put together the generators to light the street.

And when they didn't work together - he fixed them so they did before continuing.

[Alpha Zaius]

That is how a business should be run! Creativity and Productivity is at its peak when everyone is relaxed and has fun. I hope that I get to work at a company that runs of similar standards, would be a blast!

[katykiwi Moonflower]

No one will argue against the benefits of having a happy work environment but things can get so loose that the business endeavor and its users suffer.

While autonomy related to choice of assignment or project makes for a happy worker who can avoid that which he dislikes, that system results in bugs and problems that go unresolved. There is no joy or glory in attending to tedious, unglorified details and it's much more pleasant to introduce fancy graphics that all will notice. The result is flashing pie charts while the inventory remains unmanageable because of the scroll to the top problem.

Several problems exist. First is the development of features that will actually benefit regular users and that will improve upon what is already in existence. This means the developers must have some meaningful involvement with actually using features they introduce, and/or meaningful communication with a statistically valid segment of the membership.

Second is careful testing of features to make sure what is introduced does not unexpectedly break or negatively affect what currently exists. This means careful testing and research by those who are trained for development carries greater weight than meeting self imposed frequent deadlines for update releases.

Third is the correction of bugs, errors, and things broken by past updates. Again, this runs the risk of neglect of attention in a total autonomy choice of project environment.

Requiring employees to make weekly "visible" progress that moves the company forward results in flashing pie charts rather than repair of less obvious problems that may each be small in comparison to the whole, but when added up result in significant frustration and inconvenience to the membership who are left to deal with them.

You stated, "Call out inconsistency in principles when you see it. Don’t let a staid form and function become routine and boilerplate." and I strongly agree with this as my vocalization on the forums has demonstrated. In this spirit, I raise the question whether LL at this point requires more structure for the priority and assignment of projects since the loose management style seems inconsistent with the solving problems that seem to result for the membership.

To resolve some of the problems that exist, I suggest the following guidelines for your consideration:

1) A system of prioritization for the repaid of bugs and errors to take precendence in selection of a project before new features are developed.

2) The responsibility to repair any bugs, errors or breaks that result from a feature developed and introduced by a developer before that same developer can take on a new assignment.

With these rules set into place the developer can still enjoy the autonomy of project selection while at the same time adding additional assurance that the membership will not suffer the problems that seem to be adding up.

[Caliandris]

Thanks for your list Flipper. I thought you were scraping the barrel a bit with some of the things you included, and some were on the list two or three times over.

I have a list of my own. None of these things were a problem before the most recent updates.

Can't move in some sims, and have to TP out after 2 minutes. Have been with other people on different machines and we have all had the same problems in the same places.

Crashing out without warning, frequently. Sometimes in groups of three or four.

Textures download after five or ten minutes, I still have textures which are grey. If I move away from and area and return the downloading of textures seems to start from scratch.

Can't control my avatar in many places.

Text is delayed and sometimes messages arrive in the wrong order.

Many friends appear online when they are offline.

I have to reset my preferences every time I log in because black on black default text is unreadable.

Pictures which were easy to take and upload are now complicated, and poorer quality unless you know what you are doing re aspect ratio and sizing.

Texturing clothing is a pain as all the folders are open in the texture window.

Texturing objects is a pain because the current texture doesn't stay selected in the list, a bug which wasn't in the known issues list despite multiple reports.

Building is difficult, because things jump around.

Starax's wand broken: undocumented change to temp on rez.

Moopf's offline notifier broken: undocumented change to email.

Falk's book broken:
undocumented change to mouselook.

Land textures still missing after 3 hours in a sim, unless light falls on them, in which case that patch appears.

Should I go on? Really?

SL is great ... if it worked properly it would be better.

[[email protected]]

it's a great way to work, but I think you need to get better feedback from residents on what are priorities.

also, quality control needs to be better managed.

[jamma]

Philip, if you won't act like the manager of a company that's interested in keeping its customers happy with the product, at least act like the political leader of a community that you fancy yourself to be and find out for yourself what concerns your constituents most. Don't just party on the hill, pat each other on the back and vote yourself new raises.

This post is incredibly ill-timed, coming as it does after a horrid run of days with bad updates, faulty patches, unprecedented downtime, incredibly poorly tested and sometimes plainly unwanted new features, and the loss or malfunction of many regularly used features.

You need to talk to us about what you're doing to fix all this before adding new shiny. You need to make sure we see you talking about this in all sorts of forums and online publications, not just here on your own blog. And you need to expand that level of transparency down low enough for us to see what's happening on an ongoing basis. Convince us, the customer, to believe in your dream again.

Second Life Views sounds like a great little idea, but I challenge you to really spend time in the forums, sift through the static, and see that things really are not going well for us. Talk to the little avatar, shake some hands and kiss some babies, and make sure the rest of us see you doing it.

Don't depend on others to feed you the information - get it yourself. Walk with us, build with us (hah! not very easy this week, is it?) talk with us and read what we say about our experiences inworld. Get some visibility and win our hearts, and not just at the chat chaos that is our regular town hall meetings.

Don't drive the content creators away with a broken interface. Without them, it would be a pretty empty world. I believe SL is a part of the future, but it's not the only possible future. It's only a matter of time before competition comes along with Havok 4 already fully implemented and providing similar or better building tools.

Other online worlds have flourished and fallen, and SL is neither immortal nor immune to the same fate. When that competition does arise, do you want us happy and loyal, dedicated to that dream of yours and the world we've built for you, or do you want us disgruntled and ready to jump ship at the drop of a broken prim?

Remember, we can't vote you out of office. but we can vote with our feet.

[Prokofy Neva]

Philip,

1. You mention that you are one of only a "very few pioneers" of this "newfangled" way of running an office and a business. Who invented this method? And why are there so few pioneers?

2. You mention that employees can't "advance their own interests at the expense of the company" and that people are to make weekly reports of their progress, as defined by them against their own chosen goal. Who defines the company interest, and how? What is that process, where does it play out, and what is the role of the board and/or venture capitaliss in this role of defining mission? In other words, you call on people to make "good strategic judgements," but how was the original strategy formulated?

3. You say, "if the level of transparency (everyone can see what everyone else is doing) can be made high enough, you can stop managing people by explicit authority or delegation."

Then why can't you publish the rest of the handbook? Don't we work here *too*?

[Elde Eponym]

The fascinating - and disheartening - thing about this post: Utterly no mention of the people who pay the paychecks - your customers, the residents of the Grid.

[Prokofy Neva]

Well, Elde, while we work here, as I put it, we really do not pay the paychecks. Like all major media, subscriptions alone never pay for the cost. Advertising does. So as with old media, new media like SL will have to have big companies, whether businesses, or universities, or government, "pay for the advertising" i.e. buy lots of private islands and consulting services. The rank and file subscription, whether $9.95 or $195 or 1995, can't begin to cover costs.

[Elde Eponym]

Do you any actual hard numbers on that Prok?

From my side, I can point to any number of MMORPG's which are surviving quite nicely without advertising. Though from a social/psychcology POV, SL isn't a MMO - from a hardware/software/developer POV there are few differences. (And it's the latter POV that matters from a fiscal standpoint.)

[Gwyneth Llewelyn]

The Zen of Linden:

1) Fix bugs or add features? How to do both?
2) if no one wants to do the dirty jobs, and everybody wants just to do the fun stuff, how will the dirty jobs get ever started (not to mention finished)?
3) How do you please the unpleasant customers? (ie. the ones that phone up every day to complain; even worse, they *might be right*)
4) It's fun to work in an environment without external threats (competition). But if it hypothetically appears, and if everybody "chooses" to ignore them, what
5) Remember the lessons from the unfinished projects: Havok 2/3, Speedtree, HTML-on-a-prim. If they're too hard and boring to work on, what happens to the thousands of man-hours spent on them?
6) How do you manage people that can't manage themselves? (ie. how do you know if someone is unable to fix their own deadlines and manage their own time to bring successful progress reports). "Managing the unmanageable" is a nightmare.
7) What is the reward for working on issues that nobody likes to work on? (If everybody is constantly rewarded by working on what they like to work, what is the incentive for working on things that nobody wants to work with?)

And this concludes this week's selection of koans :)

Please understand that I have absolutely nothing against an open, friendly, positive, creative, enthusiastic atmosphere at work — it provides constant encouragement, does not need much to build up self-esteem, and makes people willing to work well beyond their duty, just because *they love their work* (I'm quoting Jeska :) ). Actually, I'm used at work in similar environments myself :) ... and avoid micromanaged companies like the plague, even if it means getting to work underpaid. There are some things more important than a big fat check at the end of the month; that is certainly as good an incentive as anything, but providing more than that is always a good idea.

The "Tao of Linden" does not address at all the old issue of: when work has to be done, when it's boring and not welcome, who is going to "boss" people around to do that? :)

Perhaps the current answer is "nobody. Work just lies around, unfinished" which would certainly explain a *LOT* of things ;)

Still, *I'm* not complaining. Flip's list might not be totally accurate, but it certainly shows dozens (hundreds?) of issues solved over three years. It's also unfair to claim that LL has not done anything to fix bugs! Actually, many remember that we didn't get almost any features during a *whole year* — it was just bug-fixing, bug-fixing, and more bug-fixing. Well of course, there are *thousands* of bugs to fix, but SL stagnated for a whole year while all devs only fixed bugs. It was BORING. Necessary, important, but boring. After all, if a *few* features weren't introduced now and then, we would still have primitars... :)

[Prokofy Neva]

Gwyn, I think in the Tao of Linden, the Lindens have reached such a state of enlightenment, that they "just know" to do "the right jobs". There isn't any such thing as "a list of dirty work that nobody wants to do".

Elde, I think the numbers on the front page tell the story. 8000 logged on most times; 320,000 or so subscribers. Ask yourself how the bills get paid. Through land, not subscriptions. Remember so many of the subs aren't even paying nowadays anyway.

[Kweisi Mfume]

Seems to me the solution to the problem of too many issues to resolve and either too few developers to fix them, or no perceived authoritative, prioritized list of fixes, is to open source the client sooner rather than later.

I understand your concern of fixing as many bugs as possible prior to open sourcing the client, but I can't help but think that's like waiting for the right time to have kids. If you wait for the right time, you'll never have them...because the time is never right.

I'm thinking the others in the development community would help. Then, if someone wants something fixed, either they can fix it themselves, or lobby other people to do so, not just the LL developers. It gives people a sense of control and hope that their bugs will get fixed sooner rather than later. Now, confirming those fixes will still be an internal process, but...

I see you're doing more than making software. I'm going to pass this along to my VP of HR. Thanks for posting it.

[Sky Goodnight]

Hats off to you, Philip and everyone else at Linden Lab. Linden Lab has done and is doing an incredible job with Second Life.

I am quite willing to go through the growing pains and learning curves that are bound to happen with any technology platform which has to constantly evolve.

Regardless of your company principles there are obviously very talented employees at Linden Lab.

I look forward to the future of Second Life.

[Raymond Spencer Rodgers PhD]

Virtual worlds were predicted in
my l971 book, *Man in the Telesphere*. See the last paragraph
at http://www.cajunculture.com/People/rodgers.htm for a start. Then
visit http://www.vcn.bc.ca/web-prophet and, later,
http://www.WorldwideUniversity.edu

[Raymond Spencer Rodgers PhD]

Correction to a website url -
see the last paragraph at
http://www.cajunculture.com/People/rodgersr.htm

Re 1971 prediction of virtual worlds.

[Enki Stardust]

Indeed, my company uses a similar platform. However, we have an in-house meritocracy -- people get merit for doing jobs that others don't want to do, from the others who don't want to do them but know that they need to be done, and those merit points are directly translatable into a share of our in-house currency (sort of like the L$, and probably linked eventually). Also, merit can be attributed to projects and documentation, and is automatically attributed to code in our network every time it's transcluded to another project. We may not have the clout and background that LL has, however, we have learned from their strategies and we wish to become enough of an SL presence that they might learn from ours.

Cheers,
~John Ohno
President/Lead Developer,
Accela Labs
(SL: Enki Stardust)

[Sarah]

Hi Philip,

Sarah Blow here, founder of the girl geek dinners. Could you possibly send me a quick e-mail as I'd like to put a couple of ideas past you regarding a virtual girl geek dinner and also put you in touch with one of our members who would like to get in touch with you regarding a conference in Italy that you may be interested in speaking at.
:)