Well besides learning code from books or all alone by your self. Why not create an area that small diddys can develop blender code together for small thing to be submittd as a patch? everyone could learn a little from it and pick up enough to help out, Of course we have python forum, add in C I guess ???
things could be commented and fixed via simple quote boxes ther by keeping a history and preventing someone adding in bad or weak code to blenders cvs…
just an idea and one I woul;d be willing to help out in/…
Reason why that will fail:
*Not enough support
*There is enough support in the mailing forums for what your asking
*A Massive rampage of newbie C-coders will more like dissapate within a fe months because they will be tried of ripping out hair tring to find support, aka MMORPGism
Yeah, have to say I agree with Sutabi I’m afraid. Blender is a large project consisting of many files, anyone who doesn’t have lots of experience in development would just be overwhelmed. The python forum is usually just dealing with pretty small scripts.
mailing lists dont work for many end users and are bad for searching, plus this was geared towards blender c coding not codeing in general…[/quote]
What I’m saying is that a newby programmer isn’t going to be able to understand or even know a little about how Blender works, both because he doesn’t know about the Blender source, and because he hasn’t really made any programs him/herself. Sorry if you already knew all this stuff, I have a hunch you already code…
It doesn’t need to be focused on newbie programmers, it could be biased more towards experienced programmers that have not dug in the Blender code yet… but doesn’t the development website has a forum for it?
From my point of view, this could be done with the project of integrating POVray with Blender. It should be nice to have both RCRuiz and JMS’s work online and every body could code (Python), and finally have a completly powerfull render engine working.( internal and external ways)
…or have specialized CVS for tools that leak, like SNAP to vertices (I think there are a lot of architects here, that relly need this feature )
I actually agree with the idea behind this thread.
Yes C is quite incredibly difficult… and the blender source is equally complex.
Also true that a lot of C coders for some silly reason don’t seem too helpfull too n00bs… which in a way is understandable since most programming n00bs like myself completely underestimate what’s involved.
Still it would seem fitting for an open source community to encourage new coders… Especially since complaints have been heard about a lack of skilled ones involved (the ammount not the quality of the current group that do most of the work… cause they’re all bloody amazing).
There are a few basic tutorials on the blender.org site that help starting coders find their way around the code.
It would be lovely if a more structured approach was implemented… like a n00b section… but I do think that that should be on the blender.org forum… not here… though perhaps this is a more “friendly” environment for n00bs… we’re all used to stupid questions here LOL… mostly because I ask them naturally hehe
Like SOC had people submit their ideas to a commitie and be aprovved then they started… This has more to do with some one coming up with a small idea and finding out from a higher up weather it is impossible to create or not…
Kay looks like a great list was created. http://mediawiki.blender.org/index.php/Competitive_Analysis/Goals_for_Siggraph
Would be nice to see if they can all be started on.
Then someone submitting ideas where the areas that this would/could go into aka files.c and so forth…
Lots of details, but the ones that are needed to start
One thing I am being told is that coders are possessive of their code or area of coding, can this be prevented? If not can we work around this? new kind of coder? Teamwork can be a very good thing
Hmm, right now I’m learning C++. And I mean I’m really learning. I just made a program that will loop to 10,000 and if you press a it will do it again and if you press b it will exit. But I don’t understand a Blender specific C++ forum. Are you wanting this because you like this community better?
I think that it may not be necessary for people to learn C/C++ to co-operate. Small workshops could be formed around feature suggestions. These workgroups would define these features, think how these features could fit the UI, how it should be implemented, etc.
You probably need quite a bit of coding experience under your belt before you can start adding features.
The goals for siggraph, pretty much all of those are beyond someone without a fair bit of experience both on the C side and fairly technically capable (ie can read and implement papers from previous siggraph, or can do algorithmic design - the porting of UV mapping to editmesh is probably the only item not requiring algorithmic sophistication).
If you want to start programming on Blender and are not already an experienced programmer, I’d recommend starting out trying to do a python script. Python is much easier to program in and has a far nicer API and documentation of the API.
Or alternatively find something that has already been done as a script, and turn it into a C feature.
There are minor features that can be added that are close to the skill level of a complete beginner, ie merging multiple bones into a single bone.
Alas I can’t think of many of them offhand.
Another task that is on the list that is doable is working on improving the python import and export scripts. Most of the ones we have are missing some features, and usually a similar method to implement something will already exist in one of the other scripts.
I don’t see why not these processes could not be applied to open source development as well in a larger scale.
The main advantage of a process is that you get well defined feature which to code in to the software. This could be done, if possible, first with Python and later with C.
Coding is only small part of the job. (No offense to the programmers. ) Specifying what to code is the most time consuming part.
we already have specificatios for most things in terms of existing functionality in other programs, unless you mean algorithmic design. In which case, yes designing the algorithm to accomplish the goal can be extremely time consuming.
I’m just thinking out loud here, but would it be a good idea to start a class or taskforce or whatever it is called to help harvest developers?
There are many people willing to develop for Blender, but some doesn’t even know C. Some are very good in C but are’nt familiar with the codebase.
I’ve had some communication with university professors that published in Siggraph, and are willing to help, but they can’t because they have no knowledge of the code base.
How many of such golden opportunities should we keep missing?
I think the experienced developers should start a group that takes a group of experienced C developers under their wings/ tutolage, these C developers should on their turn help out others trying to step in. Maybe it will slow development, but hey, I’m still learning stuff from 2.36. But the reward will be great if you suddenly can harvest entire crops of hardcore developers for Blender.
Just an idea.
) Specifying what to code is the most time consuming part.