Just as a reference point blender first shipped python 3.x in a non beta version in 2.57 (April 2011)
VFXPlatform first draft seems to be from 2014 (2-3 years later) which has python 2.7.3 on it, and then held on to it for dear life until 2020 when the clock literally ran out, like there was a literal clock, and they ran it out, sorry, but I’m committing to this one, that’s dragging your heels. I wouldn’t rephrase a thing.
Think it would have been a different story if we did not have a few crashers in an LTS release on our hands.
That being said, I still think selecting 3.7 again for 2021 was a mistake on their end, as was selecting an openexr version that did not come out until may of 2021, and an openvdb version that barely made it out on time (dec 24 2020), and a TBB version that didn’t build with the MSVC compiler they selected (which they didn’t correct until april 2021)
2021 was just a really really strange year for the VFX Platform, at least from the outside looking in, I’m all for following a standard, but it be real nice if said standard at least tried to be somewhat cohesive.
I’m however delighted to see they repeated none of those mistakes for the (preliminary) 2022 platform, sure python will lose bug fixes againhalf way through the year, but it’s not like they could have made another choice. ( I do feel for them there, their schedule and pythons are just never gonna line up )
So it is uppon Blender to be (as far as I know) the first DCC app to do the change to 3.9?
I love the work being done in blender, and I’m thankful for the talented and committed team behind it but right now, at this point in time, nobody is trying to catch up with blender, this is yet another way for blender to play alone and be isolated, good for blender to try to foresee technical challenges, but there is no guarantee that this solid year that it will have with 3.9 will be of any kind of relevance, if nobody cares, and nobody updates, we will not be able to use the new versions of blender in production, i feel it shows a bit of hubris to try to at this point basically expect for the industry to catch up to us or it may be (as it has happened before) that the case scenarios that are considered have way more to do with the blendersphere than the larger FX and animation ecosystem.
But it’s still there for every non LTS release, and as I mentioned, it is scheduled, but there is absolutely no guarantee, that the rest of the players in the industry will adopt, so, why be the first if we have no priority in the decision making process for third party app developers that are very important for having a well rounded pipeline, I remember there was a SideFX presentation where they announced they were switching finally to 3.x and “sorry for holding the industry on 2.7 for so long”, and yes they did hold the industry, because they are an indispensable player and all the other apps recognized it and didn’t update, so if there is no acknowledgment that we can’t be alone and we have to collaborate to be taken seriously in an industry that is more collaborative by the day, and if we do stuff like moving away from what is used by everybody just because we felt it is cleaner or better for us as if there is no need to take anything else into account, it is going to be very hard for us to be taken seriously, it’s not only about the features, how you can integrate also matters, It’s very optimistic to assume that people are going to change their pipelines for 2022 if the history of the industry with python is exactly the oposite, it would have been nice to at least wait for somebody with more leverage to take the plunge
Hate to sound like a broken record but like I mentioned previously multiple times, there were crashing bugs that were only resolved in newer python versions.
So choice was do we annoy one group of users because of a different python version, or another group of users having to deal with crashes for the next 2 years. Make a call! doesn’t matter what you choose, someone is gonna be unhappy.
We think we did the right thing for “most” of our users there, and you are absolutely free to disagree with that, and neither of us well ever be “right”
However to frame it as “blender is just arrogant, didn’t pay attention what everyone else was doing and expects the rest of the industry to follow them” is a gross misrepresentation of how this decision was made, and honestly a little insulting.
Yes you are absolutely right and that was never my intention, I have nothing but respect for the people that work hard in the development of the software, I don’t envy the position of having to make these decisions, it must be a huge burden, I also think you are right, most users are for sure going to benefit, I do think there is a disconnect between blender and collaboration with the industry that goes beyond what python version is being used, but again maybe this is by design, anyway, I’m truly sorry if I was offensive in any way, and thank you for all you do
It would be nice to have a policy of announcing important changes like this that might break 3rd party add-ons though. Something posted on the c.b.o blog and twitter to give a heads-up would go a long way to avoid frustration like this. The beauty of blender being open source is everyone can follow the development – but almost no one has the time to follow long discussions on #blender-coders . Pablo does a great job of communicating to end users and hobbyists; something for the studios / people earning money from blender is also needed
I don’t think there’s enough content there for a code.blender.org post, but it was mentioned in February in the weekly notes on devtalk here, bf-committers mailing list here, release notes here also nightly test builds have been available with this python version since that announcement early this year.
There was a bit of a stir in various places (including twitter) around that time since py 3.9 dropped windows 7 support (and thus we dropped win7 support), which did not sit super well with some users.
Yes, that does not mean we should not be constrained by the ethos of the mainstream industry, I actually get most of the reasons and I get that for the majority of the users this is not even going to be an issue, only for people that have invested in a pipeline and have blender working with others, It is clear to me that there is no malice or bad intent on the part of the blender team, it’s just a matter of perspectives and not being able to please everyone, I do think that there are some decisions that are taken in different areas that end up affecting established pipelines or making more difficult the effort to establish a unified pipeline, but I still get that this may be part of a bigger plan.
Having said that, saying something like
I think is exactly the wrong way to approach things, is like saying i don’t care, I’ll do things my way just because I can.
Just to clarify, I don’t think this is the approach that the team is taking, I’m pointing out that we should keep away from it as much as possible.
I do think that it is unfair for me to complain without having being more involved, and I think that goes for everyone.
Having said that I think that the 25+ sized studio community is still very small, so we wouldn’t have had the votes, my own perspective drives me to think that this business side of things should matter more but I’m probably very biased.
@Tonreached out around the end of 2020 (the only year we afaik committed confirm to VFX Platform) there was a recent thread on bf-committers about the 2022 platform as well, there are plenty of opportunities to have gotten involved here
You may not have had “the votes”, but unless people reach out, it’s very hard to gauge interest from the community in certain issues, so when there’s an opportunity and @Ton asks directly about a subject you care about, I’d say jump on it! You still may not get the result you would have desired in the end, but I guarantee you your voice will be heard.
Yes, for sure, my point exactly, we are not doing enough as part of the community, and that’s on us, this is my first time writing here I think, usually it’s hard to get involved because of time constraints, as teams get bigger this becomes a bigger issue, I’m actually commenting here because something broke, I don’t think I would have if this was not the case, since feature wise blender accomplish almost all we need it for, and it was kind of a call to start paying more attention, having said that, there are some trends in the industry and some conventions that I think (and may be completely off thanks to my bias) are not being given enough consideration, we as small, mid sized and big studios should be way more involved probably, and probably we are not used to having this level of input on the software but we do know what is important to operate our businesses and should be more vocal about it, we have no disagreement there.
Yes, I can pay and I do, that’s the point of having an ecosystem, I think if we are going that route and having established that I like Blender a lot and is our tool of choice for many tasks, maybe might be more productive to ask, are you willing to contribute to a developer yourself or maybe be part of a pull of small mid sized studios that pay a dev to make sure we can keep up with the Industry bugs wise?, I would say that this is something that can be discussed for sure
About the alternative option, please don’t take this the wrong way because I’m really not here to fight I if I’m rude I apologize before hand, If I’ve stated that I like the software, that I’ve developed a pipeline around it and invested money on it, That I have people trained in this software, that I do use other softwares, don’t you think it might be a bit dismissive to say basically if you don’t like it you don’t have to be here?, again I just want to raise the point and maybe I’m reading too much into it, it’s hard to read intent on text so if it not the case please ignore it, but if it is the case, it presents a complete lack of comprehension about how running a business and making long term decisions work.
Look, at the end of the day I hope blender becomes more cooperative and reliable as an ecosystem tool, if this is going to be the case that’s cool and all these problems are bumps in the road, if not, probably it won’t get integrated into more complex pipelines, and that’s ok too, nobody says it has to be that tool, happily a lot of signs point to the former hypothesis so keep up the good work, for my part I’ll try to be more involved
I see this conflict, as much as it is, as more about the difference in power balance of the relationship.
We are used to big software companies being opaque but conservative, with users voting with their wallets to keep things going along a path they like. But that is all the power they have.
Blender as a project is transparent, can move faster, and is more like a cooperative of like-minded users and developers. The difference is huge. One day you are having your groceries delivered to your door, but then you move to a hippie commune that is a collective farm. The results are better but your extra involvement is expected.
By selecting to use Blender you really need to consider that as becoming a member of the project. You are expected to follow what’s happening, to pitch in, to volunteer, to help make it better than when you found it. Blender is free, but you are not just accepting a gift made by hundreds of volunteers, you are expected to repay that gift in kind by helping make it better.
So follow developments and announcements, and the next time you see something you think is troubling, step up and say something. Because it is your project too. And it won’t get better for studios without studios pitching in, making their needs known, getting involved, contributing code, and otherwise helping any way they can.
From blender, nothing, the software is as stable as always, sorry if this wasn’t clear, from our end, we lost compatibility with our pipeline Manager, our render farm submitter, our bridge to Houdini and we can’t use Renderman on .93 which didn’t actually brake but we were really excited about it lol, as for fixing it, I haven’t had the time to dig into it so right now we are still on .83
I’m more looking on details why those things broke, and what we can perhaps do in the future to prevent such breakage, ie are these things directly linking to the python binaries, are these bpy api changes we did, etc. understanding where and how we inflict pain on our users helps making more informed decisions in the future.
We can ignore renderman here, we already understand pretty well why that doesn’t work (yet, i hear they are working on it)
