They though to have a solid technical basis to move forward. One that would not need breaking changes. This was confirmed by the fact that it was also usable in many places. That’s why they said, it was stable/production ready or however you want to call it.
As already explained, that was a mistake, because it needs breaking changes. If it didn’t need them, it would have been find to call it stable and usable (or production ready, depending on how you define it).
That is what I said since the very beginning, not sure how I should be all over the place. However, the only reason why I joined this part of the conversation is because some made a connection between this and why Tangent Animation had to shut down. As this is clearly not the case, I don’t intend to continue this discussion in this thread, because it doesn’t belong here.
And how do you know their exact thoughts? You do know that quote is from a recent post?, right. if it was in such case as you claim they wouldn’t put that part their.
So from their POV it was ready and it’s being used in their production so you can’t call it a mistake or whatever based on your opinion.
Second. We don’t know exactly what happened with Tangent, so It could be true or could be not.
The point about “production ready” in terms of Blender is a valid one because the way they define it ,is completely different from what’s expected or known for.
I think the closest definition of “production ready” I can think of is saying “We, the creators of the tool are claiming the tool is robust enough you can rely on it in commercial projects where money is at stake. Not just now, but also in (not only close) future.”
And making such a statement:
when releasing as complex of a system as Geometry nodes
when releasing it as soon as it was just fresh out of an oven
knowing that prior to that, it was used by just a single studio whom needs it was tailored to
tested on a very isolated, narrow use case by just that single studio
…is IMHO just reckless
There’s obviously no single entity which defines what exactly production ready is, but between a game development corporation, which has shipped numerous AAA game titles and provides game engine that many more AAA game titles run on, and a single Blender developer who was just only recently appointed to some product management role, it’s not that hard to guess who has gotten their interpretation of what “production ready” means more right
This conversation of Geometry Nodes is really a bit of a digression. I suggest you take it to another thread. I can split relevant posts from this one if you’d like. Just let me know.
Yes, perhaps split off the production ready related posts to something along the lines of “Production ready features in Blender” or something like that.
In it a standard practice today and even more so in agile software development to publish features as soon as possible, as soon as they provide value to the user. The goal is to get a minimum viable product which is useful and a solid basis for further development and then ship it.
No mind reading needed, that’s simply how it is done pretty much everywhere.
It was not ready because they had to break the compatibility, so it was not solid enough contrary to what they thought. It was a mistake and in hindsight it is obvious that it would have been better to make it experimental.
Pretty simple explanation, not sure why you are trying to give it a twist.
They were a pretty big player and they certainly test their pipelines in depth before they go into production and usually when such big studios are in the production mode, they won’t change their pipelines. You are never relying on what anyone says, you are testing whether it works for your use case. Even smaller studios are doing this, though they are usually more flexible to adjust their workflows if they see fit.
Could you give me a definition that I understand what you mean by production ready?
Maya is the leading tool at almost all large animation studios. Or at least a big part of the pipeline.
And then RenderMan and Arnold. Both get the testing required.
Testing in this case does not have to be all the beta testers. Beta testers are not even supposed to use it in a studio.
There is another much larger pool of testing that either happens through back channel support accounts with large companies or simply the fact that when they release a feature they have a client base that does the “testing”.
And because Maya has been accepted for so long at this level, and because it originated at a large studio to begin with. Disney I believe… It goes without saying that they get the needs of a large studio and develop accordingly.
Blender does not have this “client base”.
We will not see the Blender developers “get it” untill they have worked on a feature film or two, or three…
For that it is quite possible Blender would have to be completely refactored from the ground up…4.0?
It is human nature and completely expected that no one can just “imagine what it would be like if”. And get it, and then assume they have made the right decisions. You have to walk through it for real. Until they do that, they won’t really get it.
I think to much importance is give to blender being used in large productions.
The fact is it will never beat Maya in that area, Back when Softimage (XSI) was being developed by Avid it was arguably a better package out of the box than Maya was. But it didn’t dethrone Maya.
Blender is the perfect 3d App for the small single freelancer to the mid sized studio and that is what it should focus on. Who cares if it is used on the latest King kong movie or what ever. I use it in production by myself every day so it is production ready as far as I am concerned.
Correct… Too many details to try and remember by heart when typing…of course it was wavefront… I was referring to the development of it as a tool within a studio. Disney in this case. The Maya that was later released we all know and love…lol
Blender very definitely fills that gap you mentioned. There was a time, oh over 5 years ago, I was making this case to a room of skeptics over at the LightWave forums.
If course Blender was not having any impact at all… And so on.
When I knew for a fact it was. And that this was going to grow. And did it ever!
But even as a tool for the single artist Blender has limitations. It so happens that a more robust tool for a team is also a more robust tool for a single artist.
Additionaly if you look at it more broadly, part of what makes Maya the tool of choice and why it won out over XSI is the artist freindly part. I used XSI first and then Maya. Night and day.
Part of the ecosystem of Maya is the ability of artists to learn it at school and work as a freelancer as well as at a studio and hop back and forth as needed.
I will not derail the thread on a comparison, but it was ubundantly clear to me that XSI did not have a chance against Maya
So now you take Blender and while many people are happy to use it as a freelancer tool, the job opportunities are quite limited if you can’t quickly jump on board with a large team in need during crunch time. Many Maya artists do this frequently.
Then there are the people who have built a business around Blender and want the tool to scale up to a larger and larger team. You don’t want to find out that your tool of choice craps out when under pressure.
Blender as a tool used by larger teams is already a reality. It is attracting that crowd.
Blender as a tool for individual artists alone is a milestone long since past, in my experience.
What would be the best way to get the devs involved with medium to large studios? (Also define the headcount)
Does the blender foundation, just need to be more ambitious with the movies they are making?
Or should they organise a close partnership with a studio for a year or two, and just work to get the tools polished and address the concerns raised by the studio?
I think the code base, data structure or whatever you technically call it would have to be refactored from the ground up.
I don’t think it is a trivial task. And it would have to be done at a studio who is making a feature. Be that BI or other. And this would give them the experience.
I worked on feature film with XSI. The studio had a special technical relation with XSI. Not sure if it was part of the support or what, but if we ran into a problem our Technical guys could not fix they could call XSI and tell them about it and get a solution within a day or two. Sometimes they would call and XSI would send a couple of tech guys to meet with our people to hash out the problem.
Pretty sure the Blender foundation doesn’t have the time or resource for a larger movie. But if they had a small group of programers that a studio could call on for support that may help. They may have that already.
One thing I want to mention, and this affects all software companies. A majority do not have the staff, or resources for proper bug testing. The problem is you need to have people that know the software well enough to test it an know what it is supposed to do. And then they need enough time to test it. Most bug testing, and I have literally seen this, “Were releasing this tomorrow, take a run through it and tell me if it works” or the Dev tests out out on his machine and “Hey it works here”. Users also come up with uses and things that developers, who are usually not artists, have never tried. The reason Lightwave lights went Above 100% is because someone working on Star Trek asked for it. And the Developers response was “Why would you need that?”
And you cannot rely on the user base for bug testing. It’s ok for finding issues but not for reliable testing. Everyone is focused on their one little bug that they want fixed. Not the larger problems or how fixing it may affect something else. Also users will not know a new feature well enough to test it, or test it consistently.
Production ready… I’ll bite…
It will be different for a freelancer, a small team or a studio, but basic functionality makes it easier for all of them.
What does make a application “production ready” is at least a set of basics features that deal with things like project management, scene & internal asset management, render management, and a way to extend the application’s functionality regarding startup procedures.
This will help any user, and a\ll of the above are underwhelming or non-existent in Blender for now.
Houdini, Maya and even XSI had these from the start and were very easy to implement in existing pipelines. (I’m still mourning like @Mash3d Mash3d and @Richard_Culver…)
Extending Blender’s ability outside the Preferences>Paths option is hard (or impossible) unless you know Python, but is very easy to do in Maya, Houdini or XSI by changing the startup variables.
Setting up, or assigning a project to work in is non-existent in Blender, and no addons available atm that mimic this. Hopefully now with the asset manager this will be finally looked into.
Scene management. There’s no simple overall manager to see what files are used inside a scene, change file paths on selected files, or copy external files into a running project. Basic stuff, but no addons to be found as well.
Rendering. No batch render functionality, or any way to render out multiple RenderLayers at once. And have different cameras per RenderLayer, or have different timeline frame ranges. There’s addons out there, but they still lack the fluidity of other solutions.
Rendering in general also lack certain features requested for years, like the ability to include or exclude objects from lights. Used all over the place, but not even a on the radar for the new Cycles version.
These are some of the gripes I have with Blender, besides others (not going into my animation annoyances…) and unfortunately I don’t see these fixed or added ay time soon.
I bow my head to people like @BlenderBob to have all this working in his Studio, but it took a lot of work and effort with some very skillful programmers.
But peace out all, it’s time to start the weekend over here…
This might be the whole problem here. Different people understand it differently and there is no hard definition.
For me (and probably some others) it is “production ready”. For some others not.
I have used it without problems “in production” in “commercial products”, it was “robust enough” and there was “money at stake”. And I see no reason why it would not be so in the future. So it even appears to fit your definition.
I took a while to get our pipeline fully functional. We started with Nuke, then Blender and now we are working on getting Houdini in the same work environment but that will take a good two months to setup. But as I mentioned in my clip, this is the best pipeline I ever had the chance to work with.
Why do you feel like the code base needs to be refactored? Are there feature in particular that you think Blender needs but it is not possible to implement with the current code base?
I guess you technically don’t want to apply a band aid on top of a band aid because that is just saving pain and hardship for later on.
