Pixel Fondue's Greg Leuenberger rants on Autodesk's unfair business practices

General Forums Blender and CG Discussions
21

Financing, mindshare, education, use cases in the industry. All of these things feed to one another. The bigger Blender gets, the easier it gets to get bigger.

At the core it’s building a tool that people can use to do their work in an efficient and enjoyable manner. To build that tool Blender needs developers, and artists collaborating with those developers.

It’s how any startup grows big: pick a niche, be the best at that niche, expand into the next niche. Ebay, Facebook both started by being the biggest fish in a small pool. Once you have a solid userbase you can leverage further.

What’s the niche? For me as a game dev, personally? Indie dev, scripts, customizability in Python, freedom from licensing hell. There’s does exist the problem with GPL where you can’t really integrate things like Allegorithmic Substances or other such closed source packages into the software.

The core question to seems to be what niche does Blender serve? Can it etch out more room in a market in which new indie products are coming from every major development house.

Both games and CG made movies seem to be a growing markets, so that certainly creates more opportunities for growth.

22

Let me present you the following thesis:

Blender destroyed the entry-level market for 3D DCC applications.

Any comprehensive DCC tool would have to match up with Blender, which at this point arguably provides the equivalent of 500$-1000$ value for free. You can’t compete on price with “free”, so you’d have to deliver significantly better value, which is going to be very hard for a “high-end” program like a CG application.

At the other end, Blender’s GPL licensing makes it very difficult to build commercial solutions on top of it, so that’s not really happening either.

In essence, Blender has poisoned the well for an already unattractive market. Who profits the most? Autodesk and consorts, who have little competition to fear from what might be rising from the bottom.

So, the other “option” is to improve Blender (or create some other FOSS) as a collaborative effort from the industry players trying to be less dependent on Autodesk et al. However, aren’t we hearing all the time how tough the competition among CG/VFX houses is? Why would they collaborate on making things available for free to everyone? For one, you have the free rider problem. Secondly, the high price of commercial licenses also protect the studios from even more competition from the bottom.

I don’t see any rational (i.e. non-ideological) incentive for the situation to “improve” significantly. You’ll keep paying Autodesk (etc) or you’ll be using Blender and face (among other things) the disregard that its developers will have for your workflows with other software. You probably won’t be hiring developers to serve you either, because they’re much more expensive than just paying for software licenses outright. Accommodate yourself, accordingly!

23

The most interesting thing from these discussions is definitely how Blender has started to be viewed more as a viable alternative. It might never be able to take on Autodesk full-on, but it´ll certainly have a growing share of indie professionals and small studios over the next couple of years, if the BF keeps up it´s momentum. Alembic support I believe was a huge deal that opened up a lot of possibilities. I keep hearing about Blender/Houdini pipelines, and that would hardly be possible before (or at least very cumbersome). Cycles was important, Alembic (and FBX) was important, the next thing is going to be nodes everywhere (please)?

2016 in retrospect is going to be viewed as an extremely important year for Blender. Ton´s networking finally paid off, studios and card manufacturers (Nimble, Tangent, AMD) are starting to finally realize that supporting Blender financially is a very good idea. Blender dev fund is growing, so is the cloud.

24

Pretty good assessment. I think the main difference is that Blender not being commercial, its growth has a different path than other start ups. However of course there are certainly similarities considering Blender was originally a commercial start up. Long before its time in going for a different marketing approach and delivery system you see today with a lot of start ups.

In that light, the Blender Cloud is moving in the right direction.

The trend would seem to be that as Blender improves and the alternatives look more and more unappealing financially more people will join the cloud. It is something that could snow ball at some point.

That would be good. But it is going to take a while.

25

I am now working as a contractor at a fairly large corporation that is using Blender in one of its departments. I have been hired as the resident Blender guru to help sort out thier pipeline.

Point being. Blender has found its way in a lot of places in various industries. Things you never hear about.

There is more to life than the large studios. :wink:

26

I´m sure, but does that corporation also support Blender development? :wink: Anyway, very cool to hear!

27

Thing is, if I had the option for commercial software… I’d rather use it as the primary driver than Blender. This perhaps could be fixed if Blender wasnt GPL based.

In my opinion, if Blender were to go to the “next level” in terms of serious usage/shaking up the industry, it would need to do 3 things.

  1. Get remade (at some point) with a more closed source friendly licensing model. Cycles went Apache precisely so it could go into existing closed source applications. Blender’s licensing needs to be done in such a way where studios like Allegorithmic can safely add their substance technology to it.

  2. More streamlined industry standard interface/navigation/workflow as a means of increasing adoption and accessibility. Best way to take away autodesk users is to reduce or remove the learning curve while also making it more familiar alongside many other applications that use similar control/navigation layouts.

  3. Have a killer SDK for 3rd party development as well as a market place to go alongside with it (kind of have that now). Turn Blender into the most friendly 3rd party developer sandbox out there.

With those three things, it would be hard for Blender to not start taking a more dominant position in the industry.

28

This definitely wouldn’t be possible to do in one go.

The only way I could think when it comes to moving away from the GPL would be re-licensing areas of code as they are overhauled (so Blender would become only partly GPL and commercial vendors can integrate their stuff if they only need plugs in those areas).

For instance, the viewport code is seeing a complete replacement for 2.8, how hard would it be to treat it as its own module and license it under Apache 2? If both the Cycles and Viewport code were not GPL, we might almost get to the point where it’s sufficient for Allegorithmic to write a plugin. Re-licensing should also not be impossible as every contributor is recent enough to be possible to contact.

Other ideas in this same vein would be for the new particle and hair systems to also go down the same road when they are developed (the old system is going to be ripped out anyway).

29

Then it would be like all the others out there. Modo, C4D, Max, Maya, Houdini all competing in the same space. Blender has a great community because everybody can access it, if they spend the time to learn it.

If Algorithmic wanted they could create a custom version of Blender just like Chaos Group (VRay) and others do. Blender is the biggest SDK there is in that regards however it is not ideal for end users who must wait catch-up when a new Blender release is made. This I agree with.

Maybe a compiled plugin system could be developed for Blender so that commercial software plugin developers can hook into blender, encrypting what they want.

Your point 2 is interesting, I thought that Blender was horrible when I first tried it and can certainly understand the non-standard UI confusion. After about 5 days the penny dropped and it all made sense. But you are right, there is a need for an easier interface. Blender 101 will aim to make this easier for the eye-candy crowd out there (Unreal Engine does this to great effect). I think the problem is that Blender is not seen as industry software so many new users want icons, drag n drop and nodes instead of scripts, shortcuts and multi-use shortcuts etc. Like Houdini, it is not a toy to learn 3D as Blender is very shortcut orientated, once you get to speed with the shortcuts, memory muscle kicks in and you can really motor along. In fact I find 3DSMax a little cumbersome now, it is 99% mouse click-click-click, I never thought I would say that lol.

Anyway, keep an eye on Blender 101 (https://code.blender.org/2016/12/the-blender-101-project-and-you/)

All in my opinion :slight_smile:

Cheers

30
  • Apache 2 is still an Open License, what SaintHaven could’ve said is moving Blender from the GPL to another license that is still completely open, but is also permissive (meaning it is also friendly to commercial vendors). The re-licensing seen with Cycles for instance had no impact on how open it is as a project (but rather helped to get it out there and bring in new contributors).
  • Juggling through a bunch of custom Blender builds for the sake of using commercial plugins is not viable and not sustainable (imagine then if you wan’t to use Substance files with Vray, they are owned by different companies so you’re out of luck).
  • For better plugin development, Blender really needs a way to either make Python far faster (such as using PyPy, Cython, or some JIT solution) or have a separate API built for high performance. Face it, you’re not going to be able to easily export gigabytes of data to an external engine with the current vanilla Python solution that Blender uses.
31

Ah, ok. With you. I worried that the suggestion was for Blender to go closed, my bad. Good point about the VRay + Substance dilemma. Did not think of that.

I am all for commercial plugin support in some form or another and hope that the powers-that-be consider it in the next year or two. Different render engines + Substance support “within” Blender are top of my wish list.

32

There is a workaround to the GPL, which could be considered “legally safe” (at least for GPLv2): Going through the network or the filesystem. Both are not as efficient as you’d like, but it can work for something like a render engine. Plugin vendors generally expect a C/C++ API though, doing things completely different for just Blender is a tough sell.

Octane or Chaos Group use the aforementioned workaround. The reason that they create custom versions is that they want tweak Blender to make this workaround better, but that’s not strictly necessary. All their modifications to Blender need to be GPL’d.

Blender is the biggest SDK there is in that regards however it is not ideal for end users who must wait catch-up when a new Blender release is made. This I agree with.

Blender doesn’t have an “SDK”, it has a Python API, which is rather poor. If you actually want to (directly) use this API, your code needs to be GPL compatible.

The whole legal situation with linking code and “firewalling” yourself off with some intermediary is a legal minefield and it’s very complicated. Few courts have made decisions here (there’s an ongoing case with VMWare). Pretty much any business will just stay away from this minefield with a ten-foot-pole, unless maybe there’s serious money to be made (I don’t think that’s the case here).

What BTolputt proposes is to create a sort of “OpenCG-plugin” interface, which does not link to any particular implementation. For example, lets say that this interface has a function “createCube”, then every host application has to implement that function according to its own facilities, whereas each plugin uses just that function (nothing specific to the host application). The plugin then gets distributed separately and the GPL is not violated.

The problem is that this is a huge amount of extra work, just to define the interface properly, as well as implementing it on a host. All applications do things differently, so most likely nobody (else) would support it properly (just like with Collada). If the interface is too specific, then if Blender gets a new feature, the plugin spec may need an update as well. If the interface is too generic, it will be harder to implement and use efficiently.

Of course, this all assumes the Blender decision-makers are cooperative with this effort, every step of the way.

33

I have no idea. If they were heavily dependent on Blender for a large portion of their workflow throughout the company, I could see this making sense. As it is we are just a small team in an area that is more specific to a specialized aspect of training to do with simulation. So it is similar to preparing and viewing assets in a game engine. This one having been developed in-house by a 3rd party company. They also use commercial 3D software in other areas of training and marketing as well as commercial industrial design software. So I guess, I doubt that they are.

The point being that despite use of other software in the industry - even in this company - here Blender shows up in a relatively obscure and otherwise hidden use case.

We never hear about these things until we come into contact with them directly. And they are not sexy. But they do steal seats from the competition none the less.

34

I see a lot of you regard the GPL with a negative connotation, as if it is a burden, when in fact Blender is what it is today because of the GPL. Most of you have a problem with a proprietary vendor changing the rules, and fear that other major players will follow in that route, yet you still put you trust in proprietary software? How long until your favorite “indie” company gets bought out? You can’t have it both ways - if you prefer the status quo industry then continue the proprietary pay to play, or use Blender because you believe in the power of community and freedom from what Autodesk/Adobe is doing.

Saying a “permissive license” is just as good because “it’s still open source” is a big misunderstanding, as it opens the doors to forks and fractured communities, and can ultimately lead to the problems we want to escape. You don’t have to take my word for it - if you talk about (open source) licensing, you should at least educate yourself on it. So do yourself a favour and watch 7 minutes from a keynote by the CTO of HP, talking about the importance of copyleft licenses in an industry that dwarfs ours:

Preferably watch the whole thing, but if your attention span is shot and everything over 30 seconds is too long, the punchline is this: “Make your default license a copyleft license, such as the GPL” because “Copyleft is good, copyleft is not evil”.

If the continued use of allegorithmic/etc. software is more important to you than Martin Fink’s business insight, then by all means live in that world and accept that Blender is not supported, and accept whatever change in EULA and pricing comes down the road, Don’t try to drag Blender in to that world.

35

I fully agree with this analysis. Software like Real3D, Strata3D, Carrara, Hash AM or project:Messiah have it much harder in a market with Blender in it than they already had it before Blender got wide spread.

36

Well, arguably that’s the problem. We can only hypothesize what Blender would be, had it been licensed permissively.

You get some developers with GPL that you wouldn’t have otherwise had, but these people will be ideologues. We can debate the merits of such people, but “professionally”, I rather consider them a hindrance.

The GPL is generally toxic to business. Not all business, but in the CG or games industry, I believe it clearly is.

Saying a “permissive license” is just as good because “it’s still open source” is a big misunderstanding…

There will be tradeoffs, for sure.

watch 7 minutes from a keynote by the CTO of HP, talking about the importance of copyleft licenses in an industry…

Hah! Of course the guy comes to a Linux conference and preaches the merits of the GPL, because HP sells servers that run Linux (like everyone else) and they have to pay nothing in licensing fees. They don’t want other server vendors to have LinuxPlus© - that would set those vendors apart. That’s why they want Linux, not *BSD, to be at the top of the game (even though FreeBSD is clearly better, amirite). They want to level the playing field.

HP also makes software, but where do they use the GPL? They could be running the entire stack on top GPL software and you’d never see the source, because it’s all private server SAAS stuff. That’s where the GPL makes sense, business-wise.

“Make your default license a copyleft license, such as the GPL” because “Copyleft is good, copyleft is not evil”.

Thinking in simplistic terms like “good” and “evil” when it comes to software licenses is the root of all evil.

If the continued use of allegorithmic/etc. software is more important to you than Martin Fink’s business insight…

You know what Martin Fink would do to Blender? He’ll put it into “the cloud” and nickel-and-dime you for beveling a cube! But he wouldn’t violate the GPL, because that would be immoral!

37

Generalizations aside (seriously, stop making those and presenting them as undeniable fact), there are pros and cons to the GPL licensing. Anyone who generalizes GPL-preferring developers as ideologues should not be using Blender at all (why would you when those ‘detestable’ people are among its main drivers).

The cons have already been mostly covered (such as not being suitable for game engines, that at least being something that Ton actually agrees with). The pro argument is that the GPL prevents Blender from winding up as this minimal base that would require the purchase of dozens of paid plugins (so the point in using it becomes moot when, by the time you buy the needed plugins, you could’ve just spent the same amount on a Maya license).

The good news (or bad, depending on how radical your copyleft views are), is that Ton and co. are not forcing the GPL onto everything related to Blender (an example being the Cycles relicense). While I do like the fact that Blender is GPL outside of the game engine part, the unfortunate truth is that there really is no good FOSS alternative to solutions like Substance Designer and the Fabric Engine (the latter may eventually come about with the everything nodes project, but that will be a while yet).

Now of course, if all of Blender is to stay GPL, then who is going to create the Open Source alternative to Substance files and other proprietary tech (fortunately, we already have a few open alternatives to .fbx, but what about the rest)?

38

I will continue to make generalizations. You can not stop me.

Furthermore, in this post-factual world that we live in, there is no difference between opinion and undeniable fact. All facts are both deniable and undeniable at the same time. It is therefore a waste of time to differentiate.

Anyone who generalizes GPL-preferring developers as ideologues should not be using Blender at all (why would you when those ‘detestable’ people are among its main drivers).

I didn’t say ideologues are detestable, those are your words. In fact, those ideologues might not agree with what I say, but they would defend to the death my right to use GPL software, no matter how mercilessly I generalize them.

Also, it’s not about whether you “prefer” the GPL, but whether you refuse to contribute to a free software project like Blender, solely because it’s licensed permissively. That’s ideology.

The pro argument is that the GPL prevents Blender from winding up as this minimal base that would require the purchase of dozens of paid plugins (so the point in using it becomes moot when, by the time you buy the needed plugins, you could’ve just spent the same amount on a Maya license).

What’s bad about Blender becoming a base for plugins (like Maya)? What’s good about not having commercial plugins? You seem to imply that somehow Blender would lose features here, but that doesn’t make sense. Nobody is going to just make free plugins/features, just because the GPL forces them to.

39

What’s bad about Blender becoming a base for plugins (like Maya)? What’s good about not having commercial plugins? You seem to imply that somehow Blender would lose features here, but that doesn’t make sense. Nobody is going to just make free plugins/features, just because the GPL forces them to.

I’m not saying that commercial plugins shouldn’t be allowed, but if Blender was never on the GPL to begin with, many features today may have come in the form of paid plugins rather than in the form of a core feature (so why would you want to use Blender instead of Autodesk products if, to get the full set, Blender was just as pricey as Autodesk products)?

Now it is true that initiatives like the Blender market allowed for more sophisticated addons due to them being sold commercially, but it’s not like you’re having to pay if you want sculpting, particles, nodes, ect…).

Furthermore, in this post-factual world that we live in, there is no difference between opinion and undeniable fact. All facts are both deniable and undeniable at the same time. It is therefore a waste of time to differentiate.

Then you have lost the authority to dispute any point made by any person in any thread (don’t even try to persuade people or critique their ideas and points, including the one I made above). So be it, every point in your 1,232 posts is now to be considered null and void (as everything is valid and 2+2 can now equal SQUIRREL).

40

How do you know this is true? Is this not a generalization? Yet, you present it here, as undeniable fact? Do you not live in fear of your own words?

Then you have lost the authority to dispute any point made by any person in any thread (don’t even try to persuade people or critique their ideas and points, including the one I made above). So be it, every point in your 1,232 posts is now to be considered null and void (as everything is valid and 2+2 can now equal SQUIRREL).

You are beginning to understand…