Blender 2.8 development thread

A better comparison point would be comparing the leap from 2.79 to 2.8 to the leap from 2.49 to 2.5 (another time where there was a long period without a stable and final release, though with alphas and betas along the way). 2.5 also became production ready through an Open Movie (Sintel).

This one however looks to be even bigger, it also seems to be coming together at a faster pace due to a higher amount of resources. The move from 2.69 to 2.70 meanwhile was just a normal release cycle with more UI changes than usual.

All of the blender foundation movies have served as testbeds for blender development:
The production of Tears of Steel was integral to the development of the motion tracker.
Fur development was definitely improved by Big Buck Bunny.
Sintel was initially a test case for an overhaul of the old Internal renderer, and the production of that animation helped to reveal the difficulties in building on the existing engine. This was one of the motivations for the development of Cycles.
Cosmos laundromat brought smoke sim improvements (and ran into issues with motion blur + hair rendering)
The agent 397 teaser development improved the hair bvh, enabling a 10x increase in rendering speed for some scenes with motion blur and hair. (Could they do a re-render of gooseberry now?)

The tight integration between the BI and the BF helps to iron out bugs and develop features in a way that is very user centric, which helps to improve the experience for all users. On the other hand, some people think that the priorities of the Blender Institute (making open movies) can end up steering the development of the Blender Foundation (making open software). All of those features above are great, if you are making animated movies. If you are into architectural visualization, or 3d printing, or creating content for games, then a lot of those developments really don’t affect you, and can seem like wasted effort.

I tend to agree…I know for myself I see no point in rendering out in RT…I mean it is a realtime 3D engine…optimize it for RT!..but for me it feels like it is being optimized for rendering out animation??? I get it…a lot of people are doing that…that does not mean you sacrifice on speed…it is just one of my pet peeves as a game developer(self proclaimed)…

We last had a huge jump from the 2.4 up to the 2.5 series. There was a huge change in the UI. Since then there have been amazing developments such as Cycles, new modelling tools, etc… The general look and feel of Blender remained the same.

The jump up to 2.8 is going to be another of those ‘big’ jumps from the past.

Blender development has been tied to the open movies for quite a while now. It allows developers and artists to get together and improve Blender’s capabilities in a collaborative fashion while working on a real project. Artists get to try out the new features in a full production environment and can directly feed back to the programmers on a daily basis. This insures that new features are both easy to use and are ‘battle tested’.

It has been known for things to go wrong - nobody talks about Sintel’s render engine any more. Putting the odd example of that nature to one side, this has been a very successful development strategy for Blender.

For me, one part of the strategy is that they are directly involved in the painful bits. They push certain things to their technical limits and like that, they are getting a very good understanding of those. The consequence may be that certain parts need to be refactored or in the case of Blender Internal, it became clear that a restart in the form of Cycles would be better. The same is going to happen for the particle system, also because it became clear in the open movie projects that it is most likely less painful to start from scratch.
There are a lot of people who judge those kinds of situations a failure, but I believe that is worth to try to continue using the old systems and to get a clear understanding of the limits. If the limits can be pushed further with some work on the code, that’s perfect. If not, a lot of insight was gained as to why the limits are reached, which is going to help to design a new solution which is not going to have the same weaknesses thanks to the gathered insight.

Everyone has to go through this kind of tricky situations. Due to the open nature of Blender, we are actually seeing it, but every software company is dealing with those sorts of issues too.

I don’t know where is the better place for write this, and I write hire. Sorry if it is off topic.
I think developers can do it without my help, but I’ll write it anyway :slight_smile:

I saw these proposals about Icons.
And I don"t like it, because this icons look very similar to each other, difference between them is very small, about a few pixels. It’s a little hard to recognise.
For example:

I propose use 2 “guidelines”

  1. Shapes and formes (roundish, rectangular, squarish, triangular, cross, symbol etc). For example rectangular, it"s not a naked geometrical shape, but rectangular looks like.

  2. Color scheme. for example
    green - for creation
    yellow - kind of editing
    red- for remove
    or use other colors

For Delete-menu and Merge-menu Icons also can be used colors for vertex, edges and faces functions.

For example
vertex - blue
edge - yellow
face - purple
When you will look on this menue, you will know about blue scheme - is for vertex, yellow scheme - for edges etc.
And shapes maybe:
circular - for vertex
square - for face

sorry for my terrible English

@so3Datel, did you already reply in the linked proposal pages?

really cool icons!!

There is more of the collection icons right here :

I think I really like the blue folders, it differentiate easily the collection from the other type of data

No, I probably cant, am not from UI Team.

I have an account and I am not from the UI team. I could post a comment…

Create an account, is like any other webpage.

I am pretty sure they were looking for someone to do some icons anyway :)…those look really nice…maybe a little busy for my taste…but those are some good ideas…I really like the idea of implementing different colors for vert, edges and faces…it would be much easier ‘at a glance’ to know what you are selecting.

I have an account and I post messages sometimes. But I decide write here because maybe I’m wrong.
I created this ugly concept. Shapes and colors maybe look unpleasantly, but this concept ony for demonstrate principles.

value of buttons:
New Collection
Add selected object to Collection
Remove selected object frome Collection
Remove Collection
Create Group
Remove Group

And Delete\Merge-menu

The collection icons are confusing imho. I understand that colors are temporal, but it don’t help.

Also I think that your original proposal for delete menus is better. Is really clear.

maybe I’m wrong, but I think nobody can’t put in icon enough information for pure and pleasant reading without interpretation what’s pictured in icon. In the end you still need to be learn what each icon mean. My idea - separate icons to the groups through shapes and colors, to they did not merge visually with each other.

my personal opinion
something less stressful… less strain for the eyes & vision - less contrast and less complex design… try staring at those for 10+ years

Simply ask yourself:"How well versed am I in helping others? Am I Experienced & Confident to design for the masses?!?

& IMO best would be if it’s made possible for user to customize icons, buttons & texts - then leave it to the market, users to evolve the visual design

Well… your original icons were perfectly understood.

THe icons only needs few things. to be readable, easy to understand and don’t break the flow of the user.

I find this really confusing: the green V is universally known as a checkmark for confirmation, it doesn’t tell me “new”. Also, in the context of collections I understand + and - as “add” and “remove” collection, consistently with Blender interface where you find + and - to add and remove materials, particles systems, vertex groups etc…

I agree those icons burns eyes, modern times icons are mostly mild monochrome type icons that are really soft for eyes. Icons needs to be easy to understand from icon design. I myself are more visual user than keyboard shortcuts so it would be especially critical for me.