It’s getting a bit tired now Mark. People don’t “need” to do anything unless it suits them. The weak minded people you drag into your little cabal aren’t going to be of much value to you in the long run.

http://brainstorm.ubuntu.com/idea/23566/

i think it would be great to find a way for users to pay for “soon-to-be” GPL-code.

The argument you are presenting here is pretty interesting — it was quite refreshing to read. Yet I must admit that I do not entirely agree with some of your premises…

You are rightly observing that all revenue created with GNU/Linux distributions goes to a few big players — and in fact mostly one in particular. While one might argue they they simply understand the free software market better than their competition, this would be only half the truth at best. Indeed it is quite obvious why the standard distribution market favours big players: such distributors are pretty much offering the same software to all their customers — so clearly there is a strong economy of scale.

But is it really true that almost all the revenue generated with free software goes to these few distributors? I’m not so sure of that. While every GNU/Linux customer obviously needs some standard distribution, they also need other things. So they usually contract some kind of consultancy, to look at their requirements, and provide them with a solution — which will typically include services such as designing a software architecture; selecting appropriate components; making necessary customisations; setup; training; support. The companies providing these services are not as big, and thus also not as prominent as the distributors; but they are very successful, and growing in number and size. They don’t need to be big: they are providing individual solutions — so there is no strong economy of scale here.

I believe that discussing the free software economy, it’s usually not helpful to single out the standard distributions as a distinct group; but rather to look at the larger group of integrators, which includes both the standard distributions and the solution providers. With that in mind, I think a much better balance of power becomes evident, than from looking at distributors only…

Let’s continue on that theme. After pointing out the importance of a balance of power, you conclude that we need stronger upstreams (in the form of companies), to compete for revenue against integrators. However, I wonder whether this kind of competition is actually helpful? And is the dominance of the integrators really a problem? Indeed it seems to me that development driven by integrators is actually *better* for the quality of free software at large: while component owners are interested in making their components stand out as much as possible, integrators are more interested in making the *complete* offering more attractive to users — which actually means components being *less* perceptible, so they form a more consistent whole. In fact I believe this to be one of the major strenghts of the free software economy. (I touched on a related issue in one of my favourite rants: http://tri-ceps.blogspot.com/2009/06/one-shell-to-rule-them-all.html )

This point is actually a bit moot though, as I don’t believe that development driven by integrators necessarily means weak upstreams… An integrator generally won’t provide all the services himself — this would be very inefficient. Rather, for each important component, we need a few companies focusing on this component — gathering expertise around it, and thus being able to offer relevant development and other services much more efficiently. An integrator will pass on any non-trivial requirements regarding specific components to the companies specialising in these. We already see such companies forming around major independent free software projects: be it GNOME, KDE, Linux, GCC, Drupal, PostgreSQL…

So all in all, I think the free software economy is actually set up pretty well — we just need more of it! And honestly, I can’t think of any fundamental obstactle that would prevent that. I think the major missing piece right now is awareness: among businessmen in general; but even among free software enthusiasts, and companies already dipping their toes in free software — many simply don’t realise yet that gathering expertise, and offering services around a free software component, is a very viable and successful business model. That lack of awareness is what we need to change. Once people at large understand the possibilities, I have no doubt more companies will invest in free software development.

Note that copyright ownership doesn’t figure anywhere in these considerations. In a business based on expertise, it’s simply not necessary. Indeed it would be detriminal to the success of the component in question, and thus any company offering expertise around it: a project with an asymmetrical copyright situation is less interesting both to developers and users. Other companies and individual developers can’t participate on an equal footing, and thus are not likely to participate at all. In such a scerario, other companies can only really compete by offering completely distinct software packages — whereas one of the major advantages of a true free software economy is the ability of companies to be competitors, while sharing the codebase! It is not only more efficient regarding the amount of useful software produced from a certain investment, but IMHO also makes for a more effective balance of power. Indeed the whole *point* of free software is users not being dependent on the developers — and for commercial custormers, the ability to switch the service company while staying with the same software is possibly the strongest manifestation of the freedom. An asymmetrical copyright situation effectively negates what I believe to be our strongest selling point.

This is a crucial aspect, which I fear you have somewhat neglected in your considerations: in the end, the strongest motivation for companies to participate in free software, is being able to offer a convincing selling point for the customers, which can really drive demand in the long run. “Open Core” and other proprietary relicensing models based on asymmetrical copyright are mostly a sham — and smart customers are able to see through it. Let’s not bet the commercial success of free software on them.

So what is Canonical position about following upstream advice?

In some more radical way, such a license could ensure users that the software will become free software after a given period of time or after generating a predefined revenue. So a software company could openly publish on its website the revenue generated by its software, and upon reaching a predefined goal, releasing the product as free software.

I think that such licenses would broadly encourage free software users to put money into the free software ecosystem.

How can you prioritize shipping at certain dates higher than readyness? Users don’t care about arbitrary dates. They care about polish. I don’t get it.

listen to this quote:
“On that note, we simply must stop shipping incomplete products that aren’t ready for the end user. It is hurting our brand tremendously.”
– Anonymous senior employee at RIM
http://www.bgr.com/2011/06/30/open-letter-to-blackberry-bosses-senior-rim-exec-tells-all-as-company-crumbles-around-him/

Jon Loldrup

One of the biggest reasons why Ubuntu hasn’t yet become a mainstream success is because the software selection is not nearly as good as that for Windows or Mac OS X. The big problem here is not that there just isn’t any open-source competitor to the commonly-used Windows and Mac apps (MS Office, Adobe CreativeSuite, …), but it’s that most of these competitors provide a sub-par user experience. For example, the feature set of the GIMP should satisfy any normal Photoshop user, yet, whenever I introduce this piece of software to anyone new to open-source, they’re always discouraged by its unwieldy user interface. And the reason why ODF adoption is so slow is because the suites that have ODF as their default format provide an inferior user experience to Microsoft Office.

So what can Canonical do about this? I don’t think merely funding projects is enough, because most of these projects have had enough funding to improve the UX more radically and not much happened. Canonical could collaborate with the UX teams of certain projects and fund initiatives it finds interesting. I also like what elementary seems to be doing — tweaking and collaborating with projects to get their desired experience (Midori, Nautilus elementary, even Ubuntu). It’d be great if Canonical offered suggestions to UX teams of important projects, and if the developers refused to cooperate, then perhaps Canonical could create their own shell to the software (like it has done with Gnome), assuming that it had enough funds for this (and I think this is actually much more important than, say, bringing Ubuntu over to Wayland).

I found your one point particularly important – namely “Ubuntu’s success is linked to the success of Canonical”

Now you went on to share that Ubuntu’s current revenue stream (support-related services) is probably not sufficient to keep Canonical afloat going forward. However I recently read somewhere that Canonical is steadily approaching break-even. Mark – perhaps you can clarify this?

Regardless of this, in my mind, Canonical’s financial health is a given (be it through Mark’s pocket or support-related fees). I think the main point of discussion here is how do we give more power to the upstream FLOSS providers?

Back to Tumbes2000′s comment – the idea of using a bounty / karma to encourage the elimination of bugs / implementation of new features strikes me as pretty neat. If this was formalised and globally recognised, it could make for great decoration on one’s CV – which, apart from the actual bounty fee – ought to be sufficient incentive for hobbyist open source programmers to get the work done.

Kosta

It seems like there are 3 FLOSS business models. 1) sell support, 2) sell the code for non-FLOSS use, 3) sell ads. Model 2 use is essentially making money from proprietary software, model 3 is not applicable to many cases and is particularly difficult for an upstream. I really like László Monda’s idea. However, again, it would serve to increase the power of the distribution.

I think a better model would be to create an independent, somehow stakeholder owned, platform for cash distribution. Kevin Kelly has a very insightful article on what people will pay for in the network economy. I propose that one or more of those ideas must be at the foundation of a stronger FLOSS ecosystem.

For example, I would be willing to put money behind the bugs or features that really affect me, as László Monda proposes. The stakeholders would be open source projects and users (individual people or businesses providing cash). If the stakeholders were able to influence the direction of the organisation, and it was truly independent, it could serve to both strengthen and diffuse power in the FLOSS ecosystem.

A throwaway thought, if the whole FLOSS ecosystem was 10 times larger, many upstreams might reach a critical mass and be able to formalise commercial structures around their work.

How would having a well-known brand help us? Users may demand “Banshee”, but Canonical is giving them Banshee, just modified to take a 3/4 revenue cut. The only legal power branding gives is via trademarks, a rightfully contentious issue in FOSS land.

The only way I can see FOSS projects having more power is by being their own distributor, eg Firefox on Windows. With FOSS giving everyone the right to modify, it’s the distributor who decides what code users actually receive.

I completely disagree with your characterization of our negotiations; Banshee has clearly defined leaders/maintainers (see http://banshee.fm/about) who spoke with a single voice (not at UDS, the negotiations did not happen there, they happened in e-mails and conference calls in the months after). The mess we ended up in resulted from Canonical not speaking with one voice, offering a deal that it later revoked.

Gabriel