I say extend the schedule further. There will be a lot of press attention on Blender 4.
Either do it right, or just call it 3.9.
At this point, I think if they give themselves even an extra 3-6 months it would be better than having a ho-hum 4.0 release and then a 4.1 release that feels like it should really be much more major than a .1 update.
Calling it 3.9 would be even worse in my opinion, because there are breaking changes. Sure, 4.1 is not supposed to have them according to the definition of their versioning, but it is way easier to communicate that, rather than going with 3.9 in my opinion.
I don’t think this would help to be honest. 4.0 has so many changes anyways, the bug tracker is going to be flooded. The earlier you are getting bug reports for features that were supposed to work, the better.
Sure, PR and marketing people don’t like that (and users who believe there is more those numbers, other than them being snapshots of the main tree).
What is the rush for an open source project with no financial incentives to get a big point release out the door, when apparently the most significant changes will come in 4.1?
EDIT: Oh, saw this now…
Ok… so releasing 4.0 will most likely generate press which will point out the missing features, and then with 4.1 which will apparently break things, there might be more negative press due to that?
Those of us who have used Blender for a long time know by now that the BF has never managed to pull off a big compatibility breaking release with all targets met. They do not care about what the press would think, the only thing that matters is that the missed targets will likely come in follow-up releases. It is not about generating good marketing and viral buzz, but all about making sure the tools in their final state are robust. Blender 4.0 alpha has enough bugs already and relatively more so than in previous alpha cycles, if the release cycle is extended, it should be for the purpose of getting the bug count down.
A little disappointment at first is simply a part and parcel expectation when using FOSS software. Those who do not want that should remove all FOSS from their machine and strictly use closed-source paid software only. As for me, I am simply grateful that FOSS these days is actually capable of producing such powerful solutions without paywalls and with no catches.
About delaying 4.0, there are upsides and downsides to it. One upside is that the release will have bigger splash than releasing features for 4.1. Features that are close to being done will also get the chance to get in before things are closed. On an emotional level I think delaying 4.0 would be preferable.
One downside is that user testing in the Betas will be more limited, where more bugs may slip through the cracks because of so many massive changes being introduced at the same time. Not to mention that the scope for 4.0 will increase significantly. A good general rule for development is to ship a working product within a set time schedule. It sucks having to cut stuff, but at the end of the day there will always be cuts at some point. Delaying 4.0 will also throw away any development schedules the devs would have had for years to come. A delayed 4.0 will delay 4.1, 4.2, and so on.
The way I see it, the people who actually want to use these cut features can just jump into the Alphas and keep working. Been personally using Alpha builds for the last few years with only jumping back to older versions when things become completely broken. Haven’t really waited for a new release since 2.80 pretty much.
I wonder what is proportions of LTS and non-LTS users and their demographic.
I personally havent used non-LTS version (except new features testing) since 2.83, because there is always add-on that won’t work properly and have to be updated or even fixed manually.
Well I live dangerously and have been using the “latest” builds for the last few years, but I do not use external plugins or do professional work so for me some minor bugs do not really matter. I have not run into any “critical” bugs yet.
I do tend to keep hold of the latest stable version just in case and have a (healthy) habit of making backup files.
There are probably only few users that uses LTS, but it’s probably the safest choice for long standing projects, like movies or series that spans over a year or more.
Updating to a latest release in the middle of a project is too risky when you have a lot of people using blender on your project everyday, that leave studio only with LTS in case they got some bugs during the project.
LTS is then probably only a fraction of blender usage, but that helps to secure adoption from studios and this put a bit of pressure on schools to teach blender and makes more and more general adoption.
And since there are LTS, blender can evolve quickly between releases too. That way most users could use the latest features and studios can play safe.
What is the reason that version 4.0 should stand out more than previous versions? Is it a promotional move? Which sounds a bit strange because Blender Foundation is not driven by competition?
(I’m not criticizing, I’m just genuinely curious).
I would say, it’s in human nature to prefer x.0 releases ^^ (2.80, 2.90, 3.0, 4.0, etc…)
“today is a special day, because it marks the release of the 1.0 version of our software !! yay wohoo, a few weeks/months later… this is version 1.1 with just a few hotfixes… yeah whatever…”
how many times have you read such release notes ^^ ?
For the moment, there are only few breaking changes documented in release notes.
7 for animation and rigging. (hotkeys removal, slider range of 3 tools changed from 0/1 to -1/1, behavior change of one modifier, merge of two operators into one, 2 NLA changes,1 Python RNA change for Fcurve)
1 for Core and VFX/video page (Depth Pass removal from OpenEXR multilayer)
22 Python API breaking changes.
Honestly, if they can revert and postpone some of breaking changes to future release, they could rename it 3.7.
Nothing spectacular seems to have passed the cut.
There is no UI fundamental change.
In release notes, new default keymap seems more like a fixed one than a new one.
Most of what is in master looks like completion of work started in previous releases and fixes.
To me, it seems that they could release a 3.7 and a 3.8 as non LTS releases with some breakng changes, mainly fixes and little improvements to gain 4 months on big projects.
That does not really make sense to release a Blender 4.0 release, just to have a 3.7 with just partial python API of 4.x series.
To me, that makes more sense to wait to an EEVEE next ready, to make name of series meaningful.
This tends to be a big thing for addon developers. Along with them, their users.
Now you think of it like corporations would, where those naming conventions are basically for marketing. Unfortunately, when you follow that marketing centric route, you easily end up with development issues, where it often takes longer for features being released, leading to all sorts of issues and way more complicated internal communication (unnecessary waste of time, if you don’t put a lot of emphasis on the marketing aspect).
What I have seen in the past years is that they seem to try hit some releases for marketing purposes, like conferences and have more shiny functionality for “big” releases. Thankfully, they seem to take the reasonable route of pulling the out of the release, if it doesn’t work out.
Of course Blender is driven by competition. The number of people that uses Blender are very important to Blender itself and to us users. The more people the more plugins addons, the more tutorials, the more cooperative and work options, the more resources and enthusiasm exists.
Pretty much everything in the world is competition. Starting with competition for your time.
It has many forms and it is not most of the time about money directly, it can be about social status, power, defence of communities you belong - the Blender Community that you invested time and resources- and different levels of intensity and importance. But competition it is.
I liked before the post about 3.9 but i am not sure now. Even if some changes will only work latter if it has significant changes it should be called 4.0.
There were breaking changes of python API, in each release of 3.x series.
A dozen in 3.4.
22 is the double. But if a good portion of them can be postponed to next release ; that would probably not change perception of 3.x series, from addon developers.
And if python API can be more complete at first release of 4.x series, that would probably simplify work of addon writers.
But communications to users is important, too.
When people, passing once a while, here, ask when will EEVEE next will be released ; answer will be Blender 4.
Not Blender 4.0 or Blender 4.1.
Seriously, that was not dramatic for development, that they released Cycles X and Asset Browser for Blender 3.0.
That is useful when trying to help somebody to easily remember big leaps, and easily identify them to a series rather than a release.
And I am just saying that if status of master, at release date, is an hybrid one. That would not hurt to make a release corresponding to that status and name it 3.7, with an odd number that does not evoke an LTS.
Developers are used to work with features in experimental status and postpone them.
That is what they are planning to do, anyways by postponing features to 4.1.
So, choosing to keep the main stuff in experimental status and push as much QoL improvements to a 3.7, should not be a disastrous announcement, one trimester before scheduled release.
That is not like the name of release was more important than presence or absence of feature in master.
Some people don’t understand that a X.11 version is more recent than an X.9 version because they think it is X.1.1…
I also found something about the programmers of 3ds max where there was a discusion about majar and minor version… and the said… it’s just the next version… so no need to wait fro something… ( but of course i can’t find it anymore )
This is maybe also the reason why other vendors switch to the 2023.something versioning…
So 4.0 will break for example with some backward compatible file format reading (you have to use at least 3.6)… which is an awesome feature by the way this is not always the case…
…so now the changed it ( even if some addon developer aren;t happy for every API change)… but remember ther was also the mutliple rewrites of Cycles and Eevee because people want more and more feturres… and not evertime the actual software architecture / model is able to support those features…
And there are also somes app which do have a 1.what.ever version and then people say they are abandoned… even if the last version is only one month old…
( And then there are version numbers wich add another position after the decimal point of pi or e … )
Some other people in the community, however, disagree. I went in with their mindset that Blender Foundation is indeed not competitive, in order to find a logical reason to why version 4.0 should stand out.
To express my very own opinion: I do think that Blender Foundation is actually very keen on being competitive against Maya and 3ds max. Because of the very reasons you mentioned.
But I am open to the possibility that I am wrong, I guess we’d have to ask Ton Roosendaal himself to find a more certain answer.
I am one of the people who disagrees, that Blender is being driven by competition.
In which way do you think Blender is driven by competition? Would it not make progress, it there weren’t competitors?
Sure there are parts of Blender which are directly driven by competition, like companies improving Blender for their hardware.
But when it comes to geometry nodes for instance, I don’t see how it is driven by competition. And I don’t understand why it has to be.
Edit: To clarify: I am aware you can look at everything through a competitive lens. But that doesn’t mean it is driven like that. Anything can be improved based on the drive to make it better.
Among some other minor things, I think the main event that gave me the impression that they are trying to be competitive is the UI revamp in version 2.8. If Blender Foundation is not trying to compete with Maya and 3ds max, why would they standardize the UI to be more familiar to people using other 3D software? If they had no competitiveness in mind, they could just stick with their own unique way of working with 3D.
I’ve started playing with Blender in 2.79b (?) version, no relevant experience with other 3D software and… UI felt pretty awful back then (imo) The way I see it, the change was needed even if other software didn’t exist.