Blender Developer Notes: November 1, 2015
Here are the notes from today’s meeting in irc.freenode.net #blendercoders.
Ton Roosendaal writes:
Hi all,
Here are the notes from today’s meeting in irc.freenode.net #blendercoders
1) Blender 2.76 update
- The build we made Friday for 2.76a had an important bug… loop select in Mesh Editmode would crash on edges with seams. Bowing heads in shame It was discovered after more than a day. Decided was to not promote this ‘a’ \prominently and make a ‘b’ on Monday or Tuesday. After even more careful testing! :)
- Meeting then spent a while on discussing various testing ideas we should improve on. Agreed on was:
- More auto-tests – there’s already a “make test” in our code which runs 100s of small tests. It’s not complete at all, very good to keep working on it.
- Make the module teams more aware and responsible for testing before releases too. Notify module owners if you make a commit of which you think might need testing.
- Implement a permanently running test cycle on a system (like the buildbot) that runs a large number of more complex tests (renders, anims, games, screenies). And put the results of this on a public web page in a grid with stats and history. (Would be useful to know if a posed Llama still has hair and eyes, for example).
2) Blender 2.77 targets
- The current shortlist (remember, next release only gets decided on in December).
- Proposal: upgrade to Python 3.5.
3) The 2.8 project
- Gaia Clary heard rumours that people want to ditch Collada. To my knowledge that comes from the conference coders session. No decisions were made, it’s open.
- Martin Felke sent a proposal to the list about supporting plugins in Blender.
- Martijn Berger wrote a blog post with notes from a coders meeting at the conference.
- In general: it’s good to start collecting design ideas and proposals from everyone. We agreed however to give it sufficient time to mature. Core team members are still inworking on the release and bugfixing…
Thanks,
-Ton-
Hey development continues after 2.76 a 2.77 ??
Thought development would halt, and 2.8 was next in line, well i dont mind, i like reading about Blenders development.
Ditch Collada?? Noooo, PLEASE!!
Collada is the only reliable bridge that I know of,
between DAZ Studio and Blender,
when it comes to fully rigged and morphed and animated figures!!
I agree with Tom – PLEASE, do not ditch Collada! It is also the only reliable bridge between our company’s CAD software and Blender.
Love to read these kind of developement notes. Also a good choice to set course to a new release as known as v2.8. Blender is a great application but to see the possibilities these days, Blender and there development team has a explosive future to go. After all thinking that blender is modeling without a wireframe is really strange but logical these days.
I think that after a good proposal document and a sort of timeline a kickstarter program can absolutly bring wonderfull development needs to life for blender. The community is big and strong…….!
When I heard Ton at BCon15 talking about removing wireframe, my jaw hit the floor. I use wireframe every day to align objects in my technical animations. Often, objects are physically placed inside other objects, and it would be impossible to align them correctly if I could not instantly ‘see through’ the whole structure. Wireframe MUST remain as an option, or I will not be able to use Blender any more.
The Blender team must remember that Blender has a large following of very technical users! We are not all ‘artists’ making squidgy sculpted cartoon characters, but technical / CAD people who use Blender to visualise industrial hard-body models.
wait, OF COURSE he doesnt mean that, wireframe is the base for polymodeling etc.
I think He is just sayng that there are other way to have a nice visualisation of the scene in viewport, he probably refered to colored wireframe option.
I am concerned because Ton said:
“And of course, no more wireframes! Shocked? And why not? Well, wireframes are the stupidest things from the ’90s, the 80s… wireframes are stupid!”
He then showed an AutoDesk program as an example of what you could have instead, and pointing to a mechanism said “If you would have wireframes for that kind of stuff you would not see anything.”
WRONG! That’s just the sort of mechanism I need to look through with wireframes to check the components are aligned. If there were no more wireframes, Blender would need an alternative x-ray / transparency setting for each object together with overlaid edges and vertices.
Yes, Ton said that wire-frames are stupid, turns out he is stupid..
Also he said that saving is retarded and blender should save like google docs, turns out he is retarded and being able to not save is a must have.
There was a lot of other nonsense in his speech..
He should retire and leave the job for someone capable.
He is just thinking outside the box, out loud, on stage. Someone has to do it or we never get anything new or interesting. Is there a better way to work visually with a million poly model than wireframe view? Maybe, lets try it out. Can blender keep a realtime copy of my project in a temp file until I save it? Sounds great, right now it does exactly that but its only every two minutes by default.
I don’t think Ton is stupid – far from it – but perhaps some of his statements go too far. I certainly would not want Blender saving for me Google Docs style, that would have to be a user selectable option.
forgive the possible double post – I posted the following yesterday on Blender Code but it never showed up and I’d like to get the thoughts out into the ether – here it is again:
adding my two cents worth: the VSE is an excellent no-frills tool and should be retained, at the very least maintained, if for no other reason than to provide an open option instead of Premiere Pro or Resolve or similar – of course Premiere Pro and Resolve offer all the bells and whistles but, as a professional filmmaker, I can say that Blender’s simple VSE is a breath of fresh air, fast, pain-free and unencumbered by bloat and corporate sociopathy – I keep a custom version of Blender set up as a standalone VSE and use it for my day-to-day previz, rough cut and transcoding work – if the Compositor gets a few upgrades I can see the VSE and Compositor working as a powerful no-nonsense option for finaling too –
another VSE thought: Blender and open source offer digital artists a level of security that corporate packages do not – it’s hard to be motivated to build a small studio and commit to long-term career building around a specific toolset if a package can be whisked away at any moment –
Adobe Creative Cloud is guilty of this – focusing on a specific toolset is important if one wishes to become proficient – take Photoshop: after years of use one becomes proficient but, at least in my case, the bells and whistles are incidental to the core functionality – v7, v12, CS5, CS5.5, CS6, my basic use of Photoshop hasn’t really changed for decades – and then along comes Creative Cloud and removes accessibility for a whole class of users –
this is not to say one can’t take a left turn into open source but, fingers crossed, open source will not also remove accessibility – for me, the sense that all the tools comprising Blender will be there tomorrow makes it possible to lay out long-term projects with confidence and commitment –
the VSE may be simple and no-frills but it is reliable and, I believe, contributes (along with the rest of the package) to independent artists having the confidence to invest time and heart as they build their projects and studios and workflows –
the idea is that, through repeated use and familiarity, digital art tools get out of the way and become transparent – if those tools also come loaded with the potential for inaccessibility, mounting unease builds as one invests years and heart in their use (and the developer becomes demonized!) – this can harm confidence and motivation and perhaps affect goals and aspirations and goodwill –
maintaining the VSE (and perhaps adding formats as things develop) would be an ongoing gift to independent studios and artists such as myself – preserving it as a dependable rock solid option would be a wonderful outcome – thanks for listening – SN
Hmm. Interesting times with Blender Foundation’s direction. Can’t say I’m surprised, though.
Trying to clear things up:
Ton tries thinking outside the box, out loud, in front of people. We need that. And we also need the disagreement, he is right, it is a very old feature and for highpoly stuff it looks awful very quickly. So: Let’s start the discussion about the problem at hand. What is good with wireframe, why do we need it. Can we satisfy that very need without the problems wireframes have? Maybe we cant, then we keep it. Simple as that. But if nobody dares to provoke a discussion about what exactly it is we like and need in certain areas, those areas never evolve.
So chill, he doesn’t mean to delete them, he means to start discussions about the future. A very difficult topic.
Also: read Sergey’s answers under the 2.8 kickoff post clarify some about that confusion.
Dropping the VSE means drop ancient code and replace it with a rewrite.
Im very optimistic with 2.8, 2.5 inspired at least as much anxieties and it turned out to be the best thing that ever happened to blender.
Peace!
well, yes, but: in the intervening years between 2.5 and 2.7 Blender has experienced huge mainstream adoption which has amplified the effect any major changes will have on thousands of professional users – also, just because a feature or piece of code is ancient should not mark it for death – if it works (and more importantly, if many users depend upon it) the radical option might be to retain/maintain it in the background, perhaps add new functions, perhaps put it in a legacy mode or split it off as an add-on or similar (easy for me to say – I’m not a developer and maybe this ancient code is holding the program back as a whole?) –
I characterize ‟retain/maintain in the background” as radical because it is something software companies rarely do – Blender could advertise this as a desirable feature: absolutely dependable maintenance – look, I get it, software development is a creative art, developers look forward, not back, it’s not sexy to maintain old code, it’s boring, it’s static – but when a philosophically healthy program like Blender transcends its original envelope, when the user base explodes and builds careers and pipelines and studios and projects around it, a weight of responsibility amasses that Blender has admirably respected for years (while Adobe/Autodesk sociopathically take freedoms away) – it’s kind of like Star Wars, originally a small heartfelt space film that turned into a cultural phenomenon owned by the fans (somewhat to Lucas’ chagrin – although I’m sure he likes the cash!) – no one can argue that he shouldn’t have been free to tweak and improve the movie as he saw fit – the problem was that he also chose to make the original unavailable.
I think what’s being expressed here and on Blender Code is fear, not of change (2.7 is a significant improvement over 2.5) but of a loss of security toward the shifting sands of deprecation – Sergey says the developers “don’t remove anything without bringing something totally new and fresh to artists” – ‟new and fresh” is awesome but not at the expense of ‟old and dependable” – old and dependable is as important as new and fresh – both sum to a unique package, forward-looking and legacy-respectful – this is something Blender is brilliantly positioned to offer with 2.8 – cheers, SN