Blender 3.0

General discussion about the development of the open source Blender

Moderators: jesterKing, stiv

Money_YaY!
Posts: 442
Joined: Wed Oct 23, 2002 2:47 pm

Blender 3.0

Post by Money_YaY! » Mon Jul 19, 2004 12:07 am

Ok I admit I am just looking for a fight :P

No really, what plans are there in the next Blender Masterpiece ? When Siggrapth comes and goes will OpenGL 2.0 be incorporated into it ?

And has Ton and crew started the work flow on 3.0? I understand it will be a while still for this attack, but when does it start? Being Open Source is a new endeavor of sorts, but with that logic it would seem that it could be started tomorrow.

From what I hear a total code rewrite takes a while. And gets really boring. Should there be some sort of funding to get some coders to do basic boring rewrite stuff?

Just curious for my cherished blender.

I for one want a blender that is just a game engine like app. Something to make fast interactive eye-candy.

Rigger
Posts: 0
Joined: Wed Feb 26, 2003 1:15 am

Post by Rigger » Mon Jul 19, 2004 12:49 am

Yes Some news would be good... :D
I for one would love to help with the programming side of things but find the current source to hard to understand with its mix of C,C++ and python.

I for one would love to see a version that is 100% C++, I think with this we could do something like the Netbeans guys have and have a core Blender that can have moduals loaded as needed.

This would I think allow everyone who wants a chance to help build a better Blender.



P.s. I would be like a more Animation/Modeling studio version...

gabio
Posts: 0
Joined: Thu Jan 15, 2004 6:41 am
Location: Canada - Québec - Sherbrooke
Contact:

Post by gabio » Mon Jul 19, 2004 1:00 am

Yes it's open source. You could start a CVS reposity right there and ask every coders to start putting piece of code, draw architecture,write doc,draw map... though you need to know where you'r going. So you need all mentionned: draw architecture,write doc,draw map... Then when all choice are done start coding, adding stuff while some pp still work on 2.x release. the more it goes the more activity switch to 3.x tree: porting code to work with new core, rewrite in a more easy way etc... This is what i though might append if only all was already planned. Maybe discussed at siggraph and blender conf. i don't know.
This is a rewrite context.
But maybe blender 3.x will come after 2.99 so in this case. the stuff in blender 3.x will deppend on what appenned between 2.0 and 2.99. In short: Blender 3 might be the result of a long evolution you are now seeing.

matos
Posts: 0
Joined: Wed Feb 11, 2004 4:58 pm

Post by matos » Mon Jul 19, 2004 1:06 am

im no programmer but i agree with Rigger.i dont know if its possible, but a blender with a struture like lightwave (lightwave works like a box) could be interesting to the programmers.

Money_YaY!
Posts: 442
Joined: Wed Oct 23, 2002 2:47 pm

Post by Money_YaY! » Mon Jul 19, 2004 1:09 am

gabio wrote:Yes it's open source. You could start a CVS reposity right there and ask every coders to start putting piece of code, draw architecture,write doc,draw map... though you need to know where you'r going. So you need all mentionned: draw architecture,write doc,draw map... Then when all choice are done start coding, adding stuff while some pp still work on 2.x release. the more it goes the more activity switch to 3.x tree: porting code to work with new core, rewrite in a more easy way etc... This is what i though might append if only all was already planned. Maybe discussed at siggraph and blender conf. i don't know.
This is a rewrite context.
But maybe blender 3.x will come after 2.99 so in this case. the stuff in blender 3.x will deppend on what appenned between 2.0 and 2.99. In short: Blender 3 might be the result of a long evolution you are now seeing.
Can we start this now ? Ton seems to not be doing it so someone has to organize it. So far Blender just seems to be geting a lot of needed new features, but it would help if it was more plug in coder friendly, or super python charged.

sten
Posts: 177
Joined: Sun Oct 13, 2002 7:47 pm

Post by sten » Mon Jul 19, 2004 5:41 am

right now we are in the 2.3* series...so I guess 2.5* series will be out first...

and I belive it is little to early to start coding 3.* , from what I guess some time next year maybe...just a guess...

and since OpenGL 2.0 isn't ready yet, why start coding something that will miss things?

if I were you I would cool down a bit ;) but ideas can be drawned if you have the correct papers for the development...but only then :P
Money_YaY! wrote: Can we start this now ? Ton seems to not be doing it so someone has to organize it. So far Blender just seems to be geting a lot of needed new features, but it would help if it was more plug in coder friendly, or super python charged.
I guess he has a lot more to do than just make stuff up for Blender 3.* ....don't you think that ;) ?

jesterKing
Site Admin
Posts: 207
Joined: Fri Oct 18, 2002 12:48 pm
Location: Finland

Re: Blender 3.0

Post by jesterKing » Mon Jul 19, 2004 7:20 am

Money_YaY! wrote:No really, what plans are there in the next Blender Masterpiece ? When Siggrapth comes and goes will OpenGL 2.0 be incorporated into it ?

And has Ton and crew started the work flow on 3.0? I understand it will be a while still for this attack, but when does it start? Being Open Source is a new endeavor of sorts, but with that logic it would seem that it could be started tomorrow.

From what I hear a total code rewrite takes a while. And gets really boring. Should there be some sort of funding to get some coders to do basic boring rewrite stuff?
The keyword is patience. Blender as an Open Source project is still very young, even though it's been a while for us users and coders since the freeing of the sources.

Just lately we have seen new people coming in who show ability to work with the sources in a way or another. We are already working on modularising sources and creating documentation (see the first architecture documentation), and if you've been paying attention documentation is being done as much already for new features (read), together with new ways to communicate current progress to you users (have you noticed the B-Con status icon on the frontpage of blender.org? Have you read the first Development Digest?). Now, don't forget that the majority of the developers are people who work on Blender as a hobby, and they do not have the same obligation as a normal employee (which is both an advantage and a disadvantage), and it's amazing that Blender even is where it is right now.

The fact that Blender is Open Source does not mean that changes will happen overnight. As said, as an Open Source project Blender is still pretty much getting organised, people learning the sources and so on. Documentation of even current Blender is still scarce, but being worked on. I would hate to see Blender left half-way where it is just now when slowly efforts get better organised in favor of a Blender 3.0 that would take a huge amount of time to design and document, let alone implement. Organising around Blender still is pretty much in progress, and will be for quite some time.

Looking at current progress it would be madness to concentrate now on Blender 3.0. Let the people first learn Blender from the inside, learn the concepts and learn to appreciate its ways and see for ways to improve it as it is now, there's still much to be done.

And just to compare to some other Open Source projects who had a similar start (ie. from a commercial basis), look at Mozilla and OpenOffice. For both projects it took a long time to really start getting somewhere, and to be where they are now. Getting sources doesn't mean you have the day after everything fixed. You have the code, but you need also developers who understand it, and to keep wildgrow from happening you need to manage and organise, or you will loose focus.

Before you start pushing more, stand still for a while by the fact that developers donate much spare time, and are willing to go through the growing pains of learning the sources, and gradually get ready for really working on it. Without those dedicated people there wouldn't have been so many releases. Without those people you wouldn't be looking at a gameengine, without those people you wouldn't be looking at noise textures and ambient occlusion, without those people you wouldn't be seeing bugs getting fixed and testing builds released, without those people you wouldn't be seeing the LSCM unwrapper and weighted creases in the upcoming 2.34 release, without... well, I could go on for hours on end. You get the drift. Oh, and in a day or two you'll be looking at the second installment of the Development Digest. All at your service.

Patience. Eventually there will be a Blender 3.

/jesterKing

bobthevirus
Posts: 0
Joined: Wed Jun 25, 2003 7:11 am

Post by bobthevirus » Mon Jul 19, 2004 10:56 am

Probably not a very useful suggestion... but:

Maybe now is a good time to start to create a roadmap for Blender 3.0, with plans to start coding 2 years down the track, or whatever. The reason I say this is that there are a lot of feature requests that are popping up at the moment asking for things like a stronger node-based architecture,rendering in passes, winged edge base type (like nurbs/subsurf/polygon/curve), material tree thingys, etc, etc. All these things would, if added, be dirtier and dirtier hacks as time goes on, making the code more and more messy, and harder to develop base features for...

Therefore, I propose that we set a timeframe for getting designwork for 3.0 done, and create a mailing list and wiki so that people interested in this can put forth ideas/ decide on c vs. c++ vs. c# vs. LISP or whatever. The timeframe should be suitibly long so that the blender 2.X series has time to mature, but not so long that people feel that they must add all the features that will be part of 3.0 to 2.X also.

If feature requests are made for things that would muck up the code/slow down/whatever the 2.X series, they can be put into a 3.0 feature requests page, and the designers can, when there is only 6 months to go, realise that they haven't done anything and have all the features that they need to design for in one place. Rules should be made like "absolutly no coding until 14/05/2006" and things like that, therefore it wouldn't interfere with the 2.X work, but would still provide an outlet for all the "must-have" features that can't/shouldn't be implemented in 2.X.

So, first things:
1. Make a mailing list/wiki etc.
2. Decide on a time limit for not coding until (6 months to XX years)
3. Gather up all the feature requests that won't be implemented in the 2.X series
4. Do design work - modular/object-orientated/language/game-engine design/etc/etc
5. When the above date comes around, If enough design work has been done, start coding!!

(I just want to prevent the code getting any more messy than it already is, and this way coders don't have to worry about it till it's actually time to start coding, although they should have a say in a lot of the thing: perhaps the mailing-list could create a series of working drafts for the design, and release them for comment?)

jesterKing
Site Admin
Posts: 207
Joined: Fri Oct 18, 2002 12:48 pm
Location: Finland

Post by jesterKing » Mon Jul 19, 2004 11:19 am

bobthevirus wrote:Maybe now is a good time to start to create a roadmap for Blender 3.0, with plans to start coding 2 years down the track, or whatever.
We first need to get b2.x organised better. If we can't do that, then there's not much hope for b3.x and after. A roadmap for b2.x is in the makes, and on architecture documents of the current Blender is being worked. But until we get the current group of people and work organised it's kinda useless to start organising beyond current scope.
bobthevirus wrote:So, first things:
...
4. Do design work - ...
...
Regarding the designing, atm it's already difficult to get people to design and document (and I don't mean think up features, we all can do that, but also technical design etc.)

In short, looking at current situation it's still too premature to fully concentrate on b3 as such. With the upcoming architecture documents there'll be pointers for b3, with regards to current codebase, the reuseability of existing code and beyond that. With that it's clear that all work being done now (architecture docs, etc.) can be seen as preparation for b3 design.

On the organisational part is being worked hard (at least this replies writer is committed to that), but for those who are not closely involved it's still patience.

/jesterKing

ps. I can imagine that many are eager to hear about blender 3 *right now*

levon
Posts: 0
Joined: Thu Jul 31, 2003 6:06 am
Location: adelaide

Post by levon » Mon Jul 19, 2004 1:09 pm

blender 3??? bah thats old news im waiting for blender 4 :P

seriosly people, i think all the coders are working hard enough as it is not to stiring up comotion about something that isnt due for quite some time.

the longer it takes the better planed it will be, which will be a bonus in the long run.

kid_tripod
Posts: 35
Joined: Sun Dec 15, 2002 10:52 pm

Post by kid_tripod » Mon Jul 19, 2004 3:16 pm

I must admit most of what I'd want to be in Blender 3 is in Blender now already! I just don't know how to use all these new features . . .

Anyway, the only real monster addition I'd like to see is multithreading throughout. This is going to become a real issue over the next 2/3 years, and is a fundamental architecture thing. I'm certainly in the camp of "let's see if we can get a whole new core going now, whilst work on 2.x series continues".

Of course, my vision of 3.x series would also probably be optimised for 64-bit etc. and with real colour management!! (I spent all weekend trying to get the image to load into Photoshop and look the same as the Blender window next to it . . . curses!) And procedural modelling, arbitrary colour precision . . .

But then you look at the GIMP development, and what an utter mess that is these days. Not most encouraging.

lukep
Posts: 0
Joined: Sun Apr 04, 2004 1:39 pm

Post by lukep » Mon Jul 19, 2004 7:23 pm

Rigger wrote:Yes Some news would be good... :D
I for one would love to help with the programming side of things but find the current source to hard to understand with its mix of C,C++ and python.

I for one would love to see a version that is 100% C++, I think with this we could do something like the Netbeans guys have and have a core Blender that can have moduals loaded as needed.

This would I think allow everyone who wants a chance to help build a better Blender.
Only C++ code would be a nightmare IMHO. C++ is very good for some things, but lack seriously flexibility. I know it is fashionable to say C++ is better, but this is simply untrue. You can write bad code in any langage, and C++, if of industrial strength when correctly used, is per itself hard to use with elegance (and elegance == speed).

more important than the langage is the architecture design. Blender one is very good if quite wicked and obfuscated by numerous layers of code which differs in quality and organisation.

Ton has started writing architecture documentation, that and modularizing the code will allow newcommers to more easily understand the way things work. This will also permit a much needed cleaning of some parts which really dont smell like roses.

Then, considering the 330 000 lines codebase, a major rewriting ex-nihilo is impossible. That would need more than 20 persons full-time during 1 1/2 year.

So better than pushing for a 3.0, please let the coders work improving what already exist.

Rigger
Posts: 0
Joined: Wed Feb 26, 2003 1:15 am

Post by Rigger » Mon Jul 19, 2004 8:21 pm

Only C++ code would be a nightmare IMHO. C++ is very good for some things, but lack seriously flexibility. I know it is fashionable to say C++ is better, but this is simply untrue. You can write bad code in any langage, and C++, if of industrial strength when correctly used, is per itself hard to use with elegance (and elegance == speed).
Yes you are right, C++ can be a nightmare as I think any language can be. I did not suggest it to be "fashinable" but to suggest an OO design my mistake.

more important than the langage is the architecture design. Blender one is very good if quite wicked and obfuscated by numerous layers of code which differs in quality and organisation.

Ton has started writing architecture documentation, that and modularizing the code will allow newcommers to more easily understand the way things work. This will also permit a much needed cleaning of some parts which really dont smell like roses.
Yes I have read these docs and do find them very useful, and yes arch is more important and this is my point I can't seem to find a clear Blender Architecture and this makes it very hard to help with the coding. I believe if blender keeps being developed this way fewer and fewer pepole will be able to under stand it much less code it.
Then, considering the 330 000 lines codebase, a major rewriting ex-nihilo is impossible. That would need more than 20 persons full-time during 1 1/2 year.
Right again maybe we should look at a fork of the project, that can bring over code as needed??? No real Idea this is the last I will think/talk about it, I guess if I don't like how blender is coded I should just shut up or write my own....

:wink:

jesterKing
Site Admin
Posts: 207
Joined: Fri Oct 18, 2002 12:48 pm
Location: Finland

Post by jesterKing » Mon Jul 19, 2004 8:28 pm

Rigger wrote:Yes I have read these docs and do find them very useful, and yes arch is more important and this is my point I can't seem to find a clear Blender Architecture and this makes it very hard to help with the coding. I believe if blender keeps being developed this way fewer and fewer pepole will be able to under stand it much less code it.
I've tried to make clear that the architecture documents for current Blender are being written.

/jesterKing

Money_YaY!
Posts: 442
Joined: Wed Oct 23, 2002 2:47 pm

Post by Money_YaY! » Tue Jul 20, 2004 7:07 am

I see no harm in trying to gather paper and ideas towards the next version. As it has to start soon so it can just begin the rewrites within a year.
It will take a while but it has to get eaiser for the ones that spend a great deal of time in the code. Unless they like it messy and geeky, I don't know I just know as a user I love playing with new things to see what effects it can create.

Post Reply