governance1Add my vote for this tag create new tag
, view all tags

High Level TWiki Strategy

As you know for 10 years I am trying to do the best to grow the TWiki eco-system. Over the past 6 month the community has been consumed by disagreements instead of coding new features. In the mean time our competition is now 6 month further ahead!

Let me first state where the TWikiCommunity overall violently agrees on:

  • TWikiMission
  • TWikiRoadMap
  • TWikiMarketing activities (including the urgent need to revamp our website)
  • Our TWikiReleaseProcess with by-weekly release meetings
  • The urgency to release TWiki 4.2.1
  • Delegation, responsibility, ownership: Kenneth on release management, Crawford on architecture, Arthur on site redesign, Michael Corbett on marketing (while active), Carlo on...
  • Consultants making a living with TWiki

So, why are we so much consumed by disagreements? It's none of the above. It's not event the new TWiki governance. The key issue is disagreements in the commercial aspects of TWiki, and of (what I hear) a not well defined power structure within the community. This leads to all the politics currently going on. It's my primary goal as the twiki.org founder to fix it. I can't do that alone, I need your support.

People have asked me to more clearly communicate my strategy for the project and for the role of TWIKI.NET. TWiki is GPLed software. People who do not agree with this strategy have always had their freedom. As the twiki.org founder, BDFL and trademark holder I feel strongly about the following high level strategy, which I firmly believe is necessary to propel TWiki to the next level and to bring the community together:

High Level TWiki.org Strategy

  1. Enhance TWiki with the TWikiMission in mind
  2. On commercial aspect, follow the MySQL model, not the PostgreSQL model
  3. Build up a market place for TWiki applications that enable consultants and ISVs to make money

Some clarification on point 2 and 3, both related to the commercial aspect of the TWiki project. I firmly believe that if we follow the PostgreSQL model we will not be as successful because we will create multiple distributions which brings an overhead and confusion to the end users, and dilutes efforts of the development community. For example, one group might use Ajax technology Foo, the other Ajax technology Bar. We all know that experimentation is vital, that is what the scratch branch is for (Gilmar, I am looking forward seeing your TWikiStandAlone in the svn trunk soon).

I think it is important that we have one primary open source distribution, one primary commercially supported distribution, and on top of that, many vertical solutions or bundles. That is where the application market place comes into play. This brings plenty of opportunity for consultants and ISVs. For example, Michael Daum can sell the Workbench framework and related apps, Crawford Currie his replication technology, someone else may come up with a vertical solution for supply management, inventory management, etc, etc.

My goal is to get agreement on strategy and have the TWikiGovernanceProposal1 refined before our Berlin meeting in September so that we can do actual work at the summit and focus on roadmap details.

-- Contributors: PeterThoeny - 14 Jul 2008


I agree with that article.

-- MartinSeibert - 14 Jul 2008

We had that on the phone already, Peter. You want to push through that TWiki.net is the industry one and only commercial twiki distribution, following your personal product strategy. But why should we agree to that? Why should TWiki.net get an "extra wurst" - get treated specially? There are other competitors already, e.g. jumpbox just to name one. Sure, go ahead and build the best TWiki distro you can. However, instead of hard-coding a market advantage for TWiki.net into the open source project, you should put on your TWiki.org hat again and try to foster a healthy competition of TWiki firms all over, e.g. by joining the TWikiAlliance. That would be the right way to create standards, IMHO, among other things. You simply can't stop other commercial companies from doing their job.

-- MichaelDaum - 14 Jul 2008

My observation is that Peter is unable to see clearly through his own commercial interests. He has repeatedly described arrangements that are hard coded in his favor as "fair" and seems to believe that the work of the entire open source community should funnel into his pockets via TWiki.Net.

He has shown recently that he is also willing to abuse his power as the owner/maintainer of the twiki.org server, the TWiki trademark, and his self-proclaimed title as BDFL in order to facilitate his commercial interests.

My observation is that Peter truly believes that he is "doing good". He is not "being evil" intentinally. But he does appear to be unable to see clearly past the cloud of his own commercial interests.

My observation is that Peter has been recently unable to manage his conflict of interest, which is just another way of saying that "Peter disregards community when community is against something he really wants."

And while none of this is to subtract from Peter's innumerable contributions, it should be understood and acted upon appropriately by the community, for the benefit of the community.

-- KeithHelfrich - 14 Jul 2008

to cite: "I think it is important that we have one primary open source distribution, one primary commercially supported distribution, and on top of that, many vertical solutions or bundles.". Peter, are you implying that the TWIKI.NET version should be the one and only commercially supported distribution? So, it is your intent to make TWiki == TWIKI.NET, in the same vein that MySQL==MYSQL AB?

If that's so, please just state it openly and lets discuss. TWiki is different than MYSQL in the sense that the company came years after the project matured, so perhaps the model just don't fit.

-- RafaelAlvarez - 14 Jul 2008

Hi Peter,

I joined TWikiCommunity about a year ago and most of this time I worked only on TWikiStandAlone and I didn't interact much with community. I'm worried about TWiki future and I'm sad with all this tension (and mainly with the reasons behind, that I don't understand completely). I'm watching the discussions and waiting for the decisions to decide if I still want to invest time on TWiki. (Once I merge the code, it would not be good from me to "just" disappear, at least until it's not completely stable and well-understood by others developers).

I agree with points 1 and 3 of this article (as they are in revision 6) and I don't see any reason why point 2 could not be as 3: equal opportunities to everybody.

I don't agree with "one primary commercially supported distribution" because it seems too monopolist to me. I think it's not fair and will not lead to positive results to the community. I liked the idea of TWikiAlliance and I think PostgreeSQL model is very good: everyone contribute and enjoy. I also like Ubuntu model (forked the project from Debian and had clear rules since the beginning, like stated in UbuntuAndTWiki). IMHO if TWIKI.NET wants to offer the "one primary commercially supported distribution", then it should fork. Since forks are (generally) not a good idea, then a solution like TWikiAlliance is better (IMHO).

-- GilmarSantosJr - 14 Jul 2008

Gilmar, we are all very excited about your work and would love to see you in a leading role in that area, giving it the ultimate fine grinding and conduct this work further on. Don't get distracted too much by this political hick hack. It will get resolved. Infact, the way of your contactless but very innovative work is a good example that HighLevelTWikiStrategies and high level TWikiRoadMaps are all good and fine. But what counts is actions of people like you that create great value, that, from a roadmap point of view, was totally unforeseeable before you started. Your work was not planned ahead, yet it is so great. This is a good example showing how Open Source really works out. Creative people come along and just do it. The only thing a TWikiRoadMap can do is to retrofit it, describing what you have added to TWiki-5.0. Note, that once added to the roadmap its progress indicator will immediately jump to 99% (sort of). This does not mean that your work should not have been given more attention in a proper low-level/concrete roadmap much earlier. And that's the point where the current TWikiGovernanceProposal1 discussion comes into play: TWiki as an organization failed to create appropriate roadmaps early; people thus did not get the appropriate credit; it failed to foresee changes in the market and in the community; it's efficiency to market upcoming changes is rather bad.

So please understand that TWiki on its way to mature from an organizational point will also be better suited to give you better credits and trust, not leaving you allone on this for so long. TWiki will need to reshape its process in a way to better support doers like you. Thanks for standing on track for so long.

-- MichaelDaum - 15 Jul 2008

I have to support what Michael says. The lifeblood of open source is people like you, who see a problem and do something about it. Talking is all very well, but it's the people who are brave enough to act that end up making history.

-- CrawfordCurrie - 15 Jul 2008

"one primary commercially supported distribution". I totally missed this part when I read this topic for the first time. Yet another agument for a TWikiAlliance.

-- CarloSchulz - 15 Jul 2008

I also do feel, that competition should be active all time. Also for TWiki.net. There should not be a market segmentation by definition of the BDFL or whoever. We should try to foster a segmentation by market success.

-- MartinSeibert - 15 Jul 2008

The TWiki strategy should not be related to any commercial entities at all.

-- ArthurClemens - 15 Jul 2008

As others, I agree with the proposal, except for #2. I would add that I do not really think that the success of MySQL comes from the commercial entity. I comes because mysql was very simple to use, and very fast in common web uses. And, from the open source community side, Postgres is far from a failure...

-- ColasNahaboo - 15 Jul 2008

IF we follow the MYSQL model, I would prefer to have a TWikiFoundation, not a commercial entity like TWiki.NET, behind TWiki.. otherwise, if the entity fails (goes backrupt, shift focus, or whatever) TWiki would suffer.

-- RafaelAlvarez - 15 Jul 2008

Like many others, I'm not comfortable with there being only a single TWiki distribution (implicitly TWiki.net's), nor with TWiki.net being the guiding force behind TWiki. It's great that TWiki.net, WikiRing and independent consultants are making money out of TWiki and helping to improve it, but unpaid developers are incredibly important too (e.g. Gilmar's TWikiStandalone and many others).

Commercially, the idea of a single distro plus vertical extensions would only work, IMO, if there was a formal partnership between TWiki.net and WikiRing (and perhaps other major commercial players in the TWiki market), or if one distro leaps ahead and wins in the market for commercial distros.

The MySQL model is based on dual licensing and copyright assignment, i.e. they can sell enterprise editions etc, so I don't think it applies here - there's no chance we could re-license the whole of TWiki particularly given lack of consensus.

I haven't been following all recent discussions but there's a clear meta-issue which is that this strategy is too top-down - even if one agrees that Peter is a BDFL (such leaders are "appointed" by consensus from the community and are by definition not self appointed), it's essential to develop a new strategy in an OpenSource, bazaar-like way rather than make cathedral-like pronouncements.

Having said that, TWiki.org as a community is functioning far better than it was 6 years ago when I first got involved, largely due to a transition to a more open process where Peter has handed over some key functions to other. TWiki now has a much improved process for releases, security, features, etc, and a far higher rate of new features and bug-fixing. And of course the code is far cleaner than it used to be. So I'm hopeful we can sort out the governance / commercial aspects without blocking this progress.

-- RichardDonkin - 16 Jul 2008

Bearing in mind that I have only gotten half way through the meeting's mp3 - I would like to offer the following observation.

There seems to be (what I consider flawed) attempts to equate TWiki (the OSS project) with different existing models of co-existance between commercial companies, and the GPL community such as MySQL, debian vs ubuntu, PostgreSQL.

TWiki, as distributed in the default download, is much more like the linux kernel - almost no-one installs it without adding to it, and those that do are more technically oriented. It is very similar how Linux was for the first years - Slackware and its companions where pretty much a bundle of the minimum bits you needed to get going, and then you downloaded and build, configured and got running additional softwares.

Crawfords configure work, and my installers (debian, redhat, windows, OSX) are works that began many years ago (pre WikiRing and TWIKI.NET), and we were lucky enough to get to the current stage through a contract that we worked on, are an attempt to get TWiki to the 'debian' stage of distribution - where you can trivially install TWiki on your system, and as magically as is feasible, install or upgrade plugins, their dependencies, and give us what I would consider a professional Enterprise offering (that can then be supported by anyone).

The next stage is to gain commercially created, maintained and supported distributions - TWIKI.NET being one, VirtualBox another, and hopefully more will come - as with SuSe, RedHat, Ubuntu, and heck, WindRiver etc, distributions are where the value add is.

but NOTE, the linux kernel, does not seem to contain commercials for specific distributions - in fact, the community seems to be pretty agnostic, preferring to see contributions, in any direction, not limited to a myopic vision (ie testing, writing, discussing, reviewing, and coding) as the only significant currency.

-- SvenDowideit - 16 Jul 2008

The 'troubles' the TWiki project has experienced over the years - with obvious flareups almost yearly since at least 2002, appear to me to be entirely due to a mismatch in expectations.

Peter appears to be convinced that as BDFL of the project since 1998, he is entitled to sell the TWiki project and community to the investors of the company that he co-founded.

The community past and present, have been convinced to contribute, participate, and otherwise believe in the open source TWiki project, under a perhaps misguided impression that like other true community projects, they have the right to determine the projects fate, with Peter as founder.

I see the TCC, definitions of BDFL, and all that, as not addressing the core issue.

Does the community feel that they are happy to be sold off to the investors of twiki.net?

I've found an article that compares Open source projects that are owned and driven by a company (from their inception) vs true community projects: Comparing “open source” projects? Start by asking why does the project exist. A related post by Bruce Perens I think is also relevant.

TWiki was not owned, driven or created by a company. I don't think anyone can convert TWiki.org into such a project, without disrespecting all the previous and current contributors...

-- SvenDowideit - 20 Jul 2008

Interesting article, i.e. the characterisation of Eclipse/IBM and OpenSolaris/Sun. Interesting stuff for CommerceAndCommunity. Let's repeat what Bruce Perens said in this The Register article, just so make sure people don't miss it:

"In general open source is only going to work if you let it be a community led project. Sun has had a hard time learning this, and some of their open source projects have had a hard time getting outside contributors, because Sun has insisted on owning the whole thing."

-- MichaelDaum - 20 Jul 2008

Regarding the MySQL model, the main difference is that MySQL has a long history of open source company (not community!).

MySQL is definitely one of the pioneers in the difficult task of making money with software that is free. During its 13-year history, the company has followed virtually all the known revenue models in the open source world. From a dual license for OEMs, to support, consulting and training services or the more recent subscription model, MySQL is continuously finding the optimal combination with the familiar and not-so-familiar models.

@Sven: I don't seen any evidence that Peter has tried to sell TWiki in the past years. On the contrary, as Richard is saying, the project has evolved to a more open and transparent organization since 6 years ago. Current proposals are - well, proposals.

But the market has changed. There is more competition, notably from Confluence and SharePoint 2007, and in the future perhaps from QEDWiki.

We are discussing the commercial ideas to make a fist to the competition - by making TWiki more attractive to companies (decision makers, not the bottom-up approach).

Question is whether it makes sense to discuss the MySQL/!PostgreSQL models. Peter's worry is that if we follow the PostgreSQL model we will not be as successful because we will create multiple distributions which brings an overhead and confusion to the end users, and dilutes efforts of the development community. I am not sure about this. Is this really the problem we are facing? If so, is this the only solution?

-- ArthurClemens - 20 Jul 2008

Yes, I too would rather focus on identifying the problem that a proposal is intended to solve - but I would like to highlight, that trying to convert a community lead project to a commercially driven one, isn't really plausible in this situation.

wrt selling the community - I consider having twiki.org advertising a commercial twiki distribution as diluting and probably even denegrating the open source distribution. The implication that Certified TWiki is somehow superior than the open source TWiki we work on, I find a bit insulting, and more generally, I can't see how that it would be in the interests of the community.

My worry therefore, is not ony will the 10 year old community lose its brand, as it is overtaken by a company with a marketing budget, but the community will also lose contributors, as the number of people that are willing to work on another company's open source is smaller - and the group of third party employees allowed to contribute back will also be more restrictive.

-- SvenDowideit - 20 Jul 2008

Sven, I think you have hit the nail on the head. I can honestly say looking back that if TWiki had been company-led from day one, then I would never have adopted it or contributed to it. On the other hand, a community-led project that is supported by a company would have been easier to adopt.

I think that open source can work in a company-led project, even when the original source came from a community-led project. After all, that's the Ubuntu model, and it seems to work. But it's a moot point, as TWiki.org is a community-led project, and converting it to company-led would mean that all those people who contributed in the past believing it was community-led had been deceived.

-- CrawfordCurrie - 20 Jul 2008

There is one point in the Ubuntu model that has not been mentioned... one of the first thing they did was to create a foundation to protect the Ubuntu project, even if Canonical goes bankrupt.

The main difference between MySQL, Posgress, Ubuntu and others is that they don't have competitors helping the OSS project. In TWiki's case, there are a lot of other "commercial entities" (be them companies or individual) with monetary interest in TWiki that existed long before TWIKI.NET.

Let's suppose that TWIKI.NET manage to possition itself as THE company behind TWiki (a la RedHat or MySQL model). What may happen?

  • It may mean less revenue for those people and companies that depends on TWiki (new clients are more likely to call TWIKI.NET than Rafael & Co). Note that I'm saying less "revenue", not less "work". I can do the maths if someone is interested.
  • From the marketing perspective, it makes no sense to contribute back plugins, apps and features back to OOS TWiki, as it mean a loss of a competitive edge. For example, if I develop a Project Management app based on TWIki, and contribute it back to the OOS project, then clients who want to install it will call TWIKI.NET instead of me, IF they ever call.
  • Contributing with bugfixes and new features may be still viable, as the more people like TWiki, more people will want consulting on it.
  • To counter the "TWIKI.NET Certified TWiki Release" (which, as Sven said, implies that the OSS version is not reliable), other entities would make they own releases, with some addons ("Rafael & Co Certified TWiki Release, with Project Management capabilities"). THAT is fragmentation.

If TWIKI.NET, bu sheer marketing force, manage to make TWIKI.NET==TWIKI, then so be it, that's the nature of bussiness and I bet they will absorb (one way or another) those less fortunate. BUT, using a privileged possition at TWiki.org is just unfair competition (as someone said, "it means helping TWIKI.NET compete with my company").

So I think that making TWIKI.NET==TWIKI, effectivelly converting it in a hub for consulting bussiness, is not a win/win/win situation (specially if their primary revenue model comes from "Certified TWiki").

If we put in TWiki.org a place where all commercial entities may "compete" (much like ConsultantsForHire), a sort of marketplace or whatever, then I guess we can achieve a win/win/win situation (wins for clients/entities/community). Even more, a TWikiFoundation or a TWikiAlliance may further help the development of a healthy commercial ecosystem.

-- RafaelAlvarez - 20 Jul 2008

I do fully support Svens, Crawfords and Rafaels points of view.

-- OliverKrueger - 20 Jul 2008

Me too.

-- FranzJosefGigler - 20 Jul 2008

Me too.

-- MichaelDaum - 20 Jul 2008

Me too.

-- GilmarSantosJr - 20 Jul 2008

Totally agree.

-- KwangErnLiew - 20 Jul 2008

I totally agree with Svens, Crawfords and Rafaels statements.

-- AndreUlrich - 21 Jul 2008

Me full respect Svens, Crawfords and Rafaels pow!

-- MayerEugen - 21 Jul 2008

It is madness to force a certain model on the community, if the community so much says they see another model is better fitting. The postgres model is quite good, there is a rich world of multiple companies. I think postgres is much like twiki, it is very popular but hidden. All the simple web site hosting is done with mysql, but when you go to the companies you see a lot of postgres being used.

To be honest, twiki.net would not be interesting for my workplace. We dont want some arrogant company from the united states. That scares us. If another company can offer it more close to home, with people we can relate to, that would allow us to choose twiki (apart from the current fights that are a big showstopper at this moment).

-- JoostKattegat - 23 Jul 2008

My view of OSS projects is that there are not two that are the same. The history, the people, the product are all different. TWiki needs to define the TWiki model. We can learn from the others but we need to make a solution and a strategy that fits TWiki.

There are 3 things I try to relate to in my own view.

Why I chose TWiki for my company and how I persuaded my managers

I will copy/paste one of my statements from the Q1 summit.

When I originally chose TWiki as the product in Motorola, the existence of commercial support from multiple professional consultants was a key parameter. The fact that I knew that I could buy support in the event of trouble and not having to rely on the mercy and moods of pure volunteer people was a vital decision point. The fact that if I get hit by a bus, my colleagues can go to someone and get support was an important point when convincing the bosses. The presence of Wikiring, TWIKI.NET, and many other independent consultant is what makes TWiki a viable choice for businesses. Having commercial presence on the project is essential to making TWiki the Enterprise wiki and not just a toy. So both Wikiring, TWIKI.NET and the many other consultants are a welcome factor for the customers and for the community and there is no need for any of you to hide or tone down the commercial interests. The combination of the free open source project and the commercial support is a strength for the project and makes it stand out compared to the competition.

I think many have thought the same when they chose TWiki. This was in 2004. Since then Wikiring has built a stronger brand, and TWiki.net came into play, and the list of independent consultants has grown. I do not think TWiki should go the MySQL or Ubuntu path. But Postgres does not sell any tickets either. TWiki should build on the stengths it has.

TWiki.net - good or bad?

When TWiki.net was formed I supported the initiative. I thought the company could gather some of the forces and build a strong product that would ...

  • Enable more of the existing developers to work full time on TWiki
  • Enable fresh paid resources to be added
  • Give us unpaid contributors more in return for our effort in the shape of more and better TWiki

Now a year later none of this happened. And it will take a small miracle now to make it happen. This business model required that

  • The community had worked with TWiki.net in great strength
  • TWiki.net had made a product that customer actually wanted

None of this happened

How does TWiki look like as a product?"

Naturally TWiki has to kick the competitors arse. I see the competition driving by now in the fast lane waving at us as they shows us the tail lights. Only innovation and hard work will enable us to catch up and retake the lead.

The 2nd part is - does TWiki look like a toy or like a professionel product with the support infrastructure? This is where I think TWiki.net being very visible could add this value. The more that use TWiki, the more basis for recruting developers and the more business potential for all the consultants. The obvious advantage to the community would be resources flowing into the product. It was the hope of seeing the resources flowing to the project that made me give my support.

So what do I think then?

I actually have great doubts. I have mixed feelings.

I expect to part of a free independent OSS community which is not controlled by a company. I can trust Peter easily. I cannot trust a company which may kick Peter out tomorrow and where cool cash is the only law (I understand what you mean Joost). Sponsership and visibility is OK for me as long as it is not at the expense of other business. Exclusivity (being the only allowed visible commercial player) is not a reasonable thing to ask for given that the project community and many commercial interests were already present before TWiki.net was born. This is where TWiki is much different from MySQL and Ubuntu.

But TWiki.net could very well be the only alternative distribution because I have never seen anyone that offered or planned a commercial distrubution. And it would be suicide to try once one has established itself. Visibility on twiki.org should be earned plain and simply through sponsership and significant contribution and engagement of the existing community members.

This strategy is not going to work unless TWiki.net and the key developers agree on a commercial collaboration model. And here I cannot help. I can only encourage and hope.

-- KennethLavrsen - 23 Jul 2008

I agree with you Kenneth. I think we need to support and promote commercial activities alongside the community driven project. The only problem I have with the 'only distribution' is that any agreements on this don't have a place in the TWiki strategy. Only TWIKI.NET would benefit from it.

-- ArthurClemens - 23 Jul 2008

Edit | Attach | Watch | Print version | History: r33 < r32 < r31 < r30 < r29 | Backlinks | Raw View | Raw edit | More topic actions
Topic revision: r33 - 2008-09-05 - PeterThoeny
  • Learn about TWiki  
  • Download TWiki
This site is powered by the TWiki collaboration platform Powered by Perl Hosted by OICcam.com Ideas, requests, problems regarding TWiki? Send feedback. Ask community in the support forum.
Copyright © 1999-2018 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.