Page 1 of 1

Per Project Proposal Ways to Support Foundation

Posted: Tue Jul 15, 2003 5:29 am
by thorax
Users/Programmers can suppor tthe effort of blender
by puting in time into development, but I can't help but look and see what is happening. We are essentially proposing another NaN style
support model, when the books/cd's are bought to be put forward to
getting blender to the next year despite the milestones..
The problem is its a yearly goal and not a itemized goal, milestones
could with the proper architecture become parallel developments.
The purpose for the commercial license with blender is to allow
companies to adopt blender and resell it without releasing the sources,
I'm not sure how well this is working either.. The model I'm
proposing will go a step further, blender is developed as if it were a commercial project but the results are put into open source, so
that what is produced is useful on the level of the user and the developer.

What I propose is that when we have the feature lists and archiecture
suggestions, we essentially put up money, as users/companies,
to hire programmers to develop the features needed for blender.
This will give the developers peace of mind, also will help us find
programmers who will be willing to do the work. I can see on possibly
three levels the support can come in.

One the administrative
costs, for Ton and B@rt and others that are managing the site,
communications, docs, etc.

Two the massive code rewrite,
the eventual swich to an object oriented system, or something that
is more manageable.

Three the feature requests..

The way I see it, it could be a little bit like Ebay, users either
implement the features, or put up a proposal to pay a certain amount of money, as one or many users.

Then the developers bid on the work and the time frame that it can get done.. It may involve splitting this part into two, one for an architect
who can plan out the work, and the second for the programmer to code it.

Out of the money put forward to a feature request, some of the money, say 10% goes to the administrative costs to manage documents and
other things with this project.

The purpose of this is to allow the users to help out in the development without having to become coders, and to allow the coders to get paid back for the hard work, in wasy other than vocal praise which only
goes so far.

I am just opening this up for discussion, others can think about it.

----
PS-
This work can lead into the development of documentation,
tutorials and other sorts of stuff.. For instance, the users could
propose a tutorial on IK that covers all the features, we get
each tutorial designer to submit the first page or so of their tutorial,
along with their required costs to do the tutorial, and then the users
vote, the highest paid project, leads to the development of the IK tutorial. The result is put into public domain. I'm giving this just as an example..
But we raise the money the same way as last time. Then the users
are left to raise the money to reach the target necessary to do the work,
we pay the programmers the minute they complete the work, or
by some payment cycle, like 30%-50% initially, and the rest at the end..

For a tutorial though the amoutn might be something like 1K, 2K, etc.
But programming projects tend to run in the 10-20K region because
they tend to take longer.. I think with 1000 people reading messages
every day, if everyone chipped in a buck for a tutorial or 10 dollars apiece for better IK, and such, this would work at more nicely than
puting forth 100K every year.. And may end up being more effective because we can see where the money is put, and the cause and effect of
the model will encourage more people to fund the effort, as they will be rewarded for doing the work and for getting the features.

This is similar to paying for Lightwave or Maya, but in this case, its
treated a little like a nonprofit organization, and the results are put into open source and made available. Also the developers who tend to develop good additions and refactorizations of the sources will be able to make a living off the development of blender like independent contractors.

Another idea.. Considering Money-Yay's below, note he doesn't quite understand where all this free software comes from..

Using the method I'm suggesting, only what needs to get developed
is developed.. And the results of all developments, contracted or not, are available in the sources forever. Its a combination of the commercial model of software development with open source, with everything copy-lefted and the resulting product forever free.. If the big companies develop blender's sources, which they can with the commercial license, they must first pay the cost of the license up front, then they have a choice of either releasing the sources free or not..

So what I'm prooposing is a second method to develop the sources for money that has the benefits of copylefted open source development
with the speed and focus of a commercial development project. This is just an option.. Not something that would compete with the source development, CVS allows branches of source to be merged, so any changes made by the users and by the paid for work, can be merged
into the same source distribution, allowing both the free developers (users) the contract developers (professionals/experienced-blender-developers) to benefit the users, ultimately..

This won't compete with the commercial licenses, as the paid for
developers agree to copyleft their work, they may be able to purchase a
commercial license to continue development of the sources, which they
were always allowed to do, but this allows them to build up the capital to start their own companies without having to have money in the first place to purchase the commercial license of blender, hire employees, etc.

The purpose is to start businesses based on blender's sources
and produce a kernel of open source code that everyone
benefits from. Of course the users will have complete control on the
open source development of blender, always, as blender is copylefted,
but it also allows those users who grow up and go onto better things to
make the better thing "blender" and not something mundane like database design.

Posted: Tue Jul 15, 2003 5:43 am
by Money_YaY!
Um, Money is the bIggest and best thing. But This should NOT be on the agenda for the little users to have to think about.
As Sooooo many will shun it right away. To many complaints will happen. And Chaos will reigh supreme with it's L33t self.

I say NO! This is for Big companys to play with not the little peeps.

LOCK THREAD please >

:evil: :twisted:

Posted: Tue Jul 15, 2003 8:16 am
by kflich
so what you are proposing is actually pay-per-feature. sounds very productive and i for one will be more tha happy to donate, but a few problms may rise from this :
1) uneqal development : the number of people who use the game engine is far less then the one that use blender for rendering\animation. they will probably have more influence and cause the development to go stray away from the realtime functions (they will be able to offer more money).
2) coders will see possible profit from developing blender and with time (it always happens when money is at stake) cost will rise, eventually slowing the development.
3) finding an agreeable cost will be hard, due to the fact that development pay in the real world is very high.

it's a very problamatic system but still may work. i have another idea that may be possible - we all seen what finnancial power the blender user base has when everbody donated to free the source. maybe we could gather a solid list of features that everbody wants to see in the next blender, and put out an offer for an independent software company to develop it. the pay will come fro another donation rally from the community. this way, we get development that is not dependant on the abilities of the coders who tag along, rather on a real world company obligated to a certain standart and time limit.

after all, blender is ours to do what we want. if we can't write code (most of us are artist i belive) why not find someone to do it for us ? that way we keep the free community effort from the users who can code (maybe channel it to less hard-to-implement features) and advance blender as a whole to the next level.

/me wonders if it will ever work ? ...

Koby

Posted: Tue Jul 15, 2003 12:56 pm
by ton
Thorax addresses a few problems we need to solve, that's right. Here's what I digest from his post:

1- Focus, milestones, planning
2- Getting companies to participate in development
3- Devise a Foundation business model that's related to development, not to 'community' products.

These are all targets I can agree with. However, from an organizational perspective, we have to get ready for these targets as well. We have to find the natural and feasible growth path. So aspects I'd like to focus on first are:

1. Getting involvement of sufficient volunteers for most (if not all) development tasks, to prove Blender is alive and can be supported in an open source model.

2. Organize the volunteers efficiently, establish review boards, decision levels, try to get the focus and roadmap agreed on.

3. Get buy-in from universities and institutes to participate in development. Find students who like to do graduation projects with Blender.

4. Organize a "back office" (the daily Foundation business) with minimal means, but stable and efficient to facilitate the projects. Get involvement from other funds, donators or sponsors.

Don't forget that organized Blender development didn't start before februari this year. We're just 6 months in development, and the previous four topics are getting realized now.

For me, the next point is an important step:

5. Get full-time paid employees (in the BF) who can further coordinate and facilitate the development, at a professional level with true milestones and planning.

The amount of volunteers and the complexity of development already demands this. We can't wait for - nor expect from - a volunteer to pick this up...
The BF then can not only proof Blender is commercially viable, but also get ready for a convincing offering to companies (Partner program, commercial licensing, support).
I've already had a few tentative contacts with companies who were interested, but I sensed a reluctance with them, they don't really know what goes on, nor what Blender's future will be. They prefer to see actual evidence; a lively well organized open source project which delivers exciting opportunites for them.

6. establish the Foundation business model.
My vision is to keep the Foundation as small as possible, it should remain idealistic, non-commercial, community oriented.
Around the Foundation a network of companies can be established, with all potential commercial Blender activities located there, and not in the Foundation. The Foundation then only acts as 'facilitator' and 'coordinator' here.
Opportunites for companies in this network:
- Publishing business (manuals, cdroms, magazines)
- Education business (training programs, on-site, on-line or with DVD)
- Content studios (being part of the network creates opportunities)
- Technology support (companies or individuals who assist on integration projects)
- Product development (special Blender versions, plug-ins, add-ons)

I challenge Thorax, and everyone here, to seriously consider this. Don't expect the Foundation conducting this business... but grab the opportunites when they arise. Expect from me (and the BF) to be extremely supportive for such activities. I would love to see developers starting an e-bay style feature-request system!

How exactly the BF would get a (commercial) share of this business is not so relevant now... most crucial is that people should be stimulated to do business with Blender. If this happens succesfully, we will benefit from it anyway. Even when it "just" gives a better Blender version... :)

Posted: Tue Jul 15, 2003 1:02 pm
by thorax
kflich wrote:so what you are proposing is actually pay-per-feature. sounds very productive and i for one will be more tha happy to donate, but a few problms may rise from this :
1) uneqal development : the number of people who use the game engine is far less then the one that use blender for rendering\animation. they will probably have more influence and cause the development to go stray away from the realtime functions (they will be able to offer more money).
Read up on CVS and logfile systems.. There is no way to lose the previous developments.. They can coexist.. CVS is a concurrent revision control system.. What a revision controll system does is it stores only eth changes ont he source, unless the changes are remove from the source tree (the tree of changes), they are never lost, and changes that are nonoverlapping are additive.. ITs the open source developer's construction history, essentially.. Although its possible the construction history is more multi-dimensional than Maya's construction history as
every modification is based on some branch of a previous release,
and branches from splits in the codebase can be grafted elsewhere if there is no conflicts in the sources.. This is why random developments can work, but it helps to have focus..

I see gameblender is a seperate development, but blender ultimately will
benefit from object oriented design, which is why the two are not compatible.. But the two can be developed without each having to be too much involved with each other, game blender could be developed off of old sources of blender, by virtue of the CVS mechanism..
2) coders will see possible profit from developing blender and with time (it always happens when money is at stake) cost will rise, eventually slowing the development.
What causes slow developement:

1. lack of knowledge..
2. lack of interest..
3. speed of the developer's ability.
4. release of features according to a schedule to make the biggest return on an investment..

(note Intel uses 4, when competition is in a sort of stasis they will
develop ahead of the competition and release chips maybe every year or so.. I know IBM also does this too).
3) finding an agreeable cost will be hard, due to the fact that development pay in the real world is very high.
Not in India or China.. to write software you only need to have a brain and a computer and someone to pay the electric bill. Its why there is so much competition in the software field.. Its also why a lot of software developers apply for jobs with the government (there is a problem with trusting foriegn nationals to write government systems, considering the
current US security situtation. I'm sure this is the case for other countries as well.. I know in Holland that it is not possible to hire out non-dutch citizens out for work unless the work can be done by a dutch citizen for the amount of money, but if you plan to pay like 3000 a year, it opens up the market to Indian and Chinese developers.. I know all you developer guys are looking at me "evil" :twisted: :twisted: Yes I know.. But you can't deny their existence.. There are other ways to coexist, such as those I suggested..

Posted: Tue Jul 15, 2003 1:31 pm
by thorax
Ton's number 3 point is a good one.. I was going to comment in my last message on that one, but I erased a lot of my message as it wasn't
relevant I thought to kflich's message..

I would say, when I was at UNM, I was wanting the CS department there
to develop blender as a part of a education process.. But it would
interest the students and grad students to be offered funding to
develop the sources, and in turn use their developments to
support their thesis in getting masters or PHD.. There is so much graduate work that goes to programming stuff that never really profits anyone anything, its so much better when it does..

Also in a educational environment, the considerations tend to be less monitarily controlled and more in the interest of choosing better approaches and pushing the envelope..

I'm glad he likes the ebay idea.. I guess I mentioned www.hotdispatch.com ... I would not suggest this site be used, but
it serves as a model of sorts.. I know the idea their site is not patentable
because I was one of the first ten employees for the guy that created the "information sales" concept. I've used hotdispatch to get actual work
done, I used it to get a reference to source code for doing low level
infared port communications with palm pilots.. A neat
option of hotdispatch is you can choose (they may have changed this)
to have the result of your question, the answer or solution, freely
available to others.. <--- This is what I'm proposing for the blender development in a nutshell..

Posted: Tue Jul 15, 2003 4:22 pm
by Money_YaY!
heh, so it is about money.

Not enough interest I would assume is why there are so few developers. More developers is needed. I guess cash would solve that. But can you ever get enough, cause it always runs out .

^o^

Posted: Wed Jul 16, 2003 1:10 am
by thorax
Money_YaY! wrote:heh, so it is about money.

Not enough interest I would assume is why there are so few developers. More developers is needed. I guess cash would solve that. But can you ever get enough, cause it always runs out .

^o^
If its well architected, it will be possible to have it done on time
and in the budget. The problem with most projects is they
are paid to be hacked to some goal, not architected in the first place..
Its like building a house given wood, nails and people and no clue what to do. If you have blueprints you can assess the completeness and progress of the development.. That's the purpose for having a plan aside from
knowing what to do..

The process may work like this:

1. users agree on features that are needed.
2. users report how much they would spend to get features.
3. developers bid, giving their plans and proposal and possibly a first installment on the development (so we can judge their code).
4. users decide, to a vote (consult others, like Ton).
5. select the party to develop it.
6. get them to sign a contract.
7. pay the party money to cover their costs for initial development,
this will cover the first third of the development.
8. the party hired to develop will post their progress and users will approve. If development doesn't proceed as expected or if there is some problem with the developer {chemistry?}, users can back out of the deal or nullify the contract.
9. developer is free to ask the clients (us) to test out their development over time and give feedback.. Its important to keep in mind the plans for the development the developers progress, just you would want the developer to check back often and give reports on his progress, possibly in the form of checking in sources to the CVS repository.
10. users get features requested..

In any case concerning development for money, the emphasis should be on giving the the users more control over the process because
they are paying for the work, the developer is given less control in such a deal, they have the choice whether or not to take the project.
But its harder to get cheated if there is more validation of the
code quality and agreement on the terms of the development.
Also we would like the developers to feel welcome, so there needs to be some amount of trust, but the developer has a feduciary responsibility to us to develop according to our needs..

Its always possible the work may not get finished on budget, but this will look bad on the developers record, this is bad business for them,
another reason we should use something like hotdispatch, we can find the good developers that have satisfied their clients well.

However a price can't be put on good volunteers.. Developers that
work on blender for the pleasure of doing it won't be pressured..
They can come or go, as we know.. But it helps to know what to
do with the sources, and to prefer those who have worked on the sources.. This could be used to consider hiring those closest to
us than outsourcing to to hotdispatch, but say its something nobody
wants to do, we can surely find some outfit on hotdispatch, to outsource the work to.. That's what those firms are for.. They do the work nobody else wants to do..

Posted: Wed Jul 16, 2003 4:44 am
by kflich
hmm ... in detail it sounds better (or too many words just confuse me). so lets give it a go somehow. open a thread named "payed feature request" take sugestions from people who are willing to pay for it. then collect the money, make a bid, and see which of the coders are interested in taking the offer. all that's realy needed to get it moving is to create a general contract between users and coders, to be used in the deal. i have a feature request i'm willing to pay for (and it has a coder working on it), maybe i'm not alone. worth a try.

Project management is where it's at ...

Posted: Sat Jul 26, 2003 11:30 am
by purpleblenderuser
Niether, Open Source is free (develop it adhoc)
Not really much of an option is it - 'adhoc'. Open source development can be very organised, directed and fast. Should the Foundation take more of a development role - quite possibly. Management of the project, especially one this size, is the most important thing.

I would pay money to support the blender project but I can not agree with paying developers. Better still I would give my own time for development and project planning.
to write software you only need to have a brain and a computer and someone to pay the electric bill
Yes to instruct (and motivate!) that somebody to produce something meaningful is the important thing. Who is going to pay to handle the contracts, etc.
What causes slow developement:
1. lack of knowledge..
2. lack of interest..
3. speed of the developer's ability.
4. release of features according to a schedule to make the biggest return on an investment..
Oh! Developers may need more than a brain and a computer. And, yes, to do (4) you have to rely not on the developer but on the planning. Regarding points (1) and (2) how is getting independant contractors (from anywhere) going to increase the knowlesge of the product and the motivation.

Mr Ton said,
5. Get full-time paid employees (in the BF) who can further coordinate and facilitate the development, at a professional level with true milestones and planning.
Yes!

Re: Project management is where it's at ...

Posted: Sat Jul 26, 2003 11:02 pm
by thorax
purpleblenderuser wrote:
Niether, Open Source is free (develop it adhoc)
Not really much of an option is it - 'adhoc'. Open source development can be very organised, directed and fast. Should the Foundation take more of a development role - quite possibly. Management of the project, especially one this size, is the most important thing.
Well the management is intertwined with the ability to
know what needs to change architecturally.. But the reason I suggest
OOP and rewrite is the possibility that having the code and data
localized to objects will allow more parallel development rather
than developers stepping on each other's toes because one
is affecting the state of variables in another's domain. The CVS
can't track the usage of global variables, and have a localized
scope using objects will help parallelize the development (you know
from parallel processing that you you can only parallize code that is
not dependent on other information elsewhere, same problem)..

To streamline the development there will need to be a consideration
about dependencies across the source code and when kernel code
is being modified what to tell which developer not to modify until the
kernel code is fixed because everything id dependent on it.. Otherwise
a developer may have to make the same change twice.

Its also possible there will be developers changing code that is
dependent on a particular state of assumptions, if model changes
the assumptions are proven false the code continues not to work,
this is due to not correctly understanding what the state is doing..
This will also need to be determined by the foundation or whoever
chooses to oversee the project, but its more complex than simple
project management (one reason I tend to can't stand managers without
any programming experience).

Managers entering the area of managing programmers from some other
management area probably have to get new degrees and experience
because the problem is more complex because the pieces are not
independent of each other unless using a OOP design..

I would pay money to support the blender project but I can not agree with paying developers. Better still I would give my own time for development and project planning.


That would be welcome, what are your credentials? Have you done this
kind of work.. I haven't but I can imagine what would be needed,
particularly deconstruction of the sources to their parts and
determining the proper organization, then determining which
parts need to be reworked in which order to meet the design needs
given the suggestions of the users and developers, then to
do test designs to project the scalability of the solution, and then to
determine a paln for the redevelopment. I think Ton is doing
this but still there is a interest of doing this in C and haven't
been given significant proof that not using C++ is going to keep
the design quick and efficient. I mean all an object is is a
database interface that can manage instances of itself, the functions
that manage the database are the methods, and the data
are the state of each instances and the data the instances contain.
This would be possible in C, but for every kind of object you
would need a new database interface, possibly overlayed on the
same database interface. In C++ its just easier to manage
without thinking about the problem so much. Note a function call
takes the time necessary to preserve the state of execution (storing registers on a stack) and accessing another part of memory (which
could cause a virtual memory page to be swapped in from disk)
and executing the code, then popping off the stack the register
to continue executing where it left off.. Recursion just causes the
state stack to grow, that's why a computer will eventually crash if
you recurse too much, or unless the compiler supports the unrolling
of recursions as interations. But in C its a function call with a pointer
to some data (a passed parameter via a register not stored on the
stack).. In C++ its a pointer to some data with a function call,
may require at least some pointer arithmetic to locate the function by
reference to the particular type of object.. Also if we are doing a method
search, that will cost some time, but I don't see why C++ would
offer any different runtime costs than using C, its just a different
way of organizing information..

to write software you only need to have a brain and a computer and someone to pay the electric bill
Yes to instruct (and motivate!) that somebody to produce something meaningful is the important thing. Who is going to pay to handle the contracts, etc.

Yeah that would also be needed.. What I was saying about "the brain and the electric bill" is the fact that with software development your
developers can exist anywhere because it doesn't require
chemicals and a lab, it just requires electricity and a brain (and a computer of course). Its fairly low cost, you are really just
paying for the brain power and the ability to think.. If you are a
mathimatician its even easier, all you need is food and a brain..
There is the joke that you insert coffee and out comes equations..
That must be why math departments tend to be underfunded in schools and
electrical engineering and chemistry departments are more funded..
hard technologies cost more money to produce and are harder to replicate. But good software and math can make lots of money
because it changes the way things are done and can be easily replicated.
Hmmm I guess..

Anyhow..
What causes slow developement:
1. lack of knowledge..
2. lack of interest..
3. speed of the developer's ability.
4. release of features according to a schedule to make the biggest return on an investment..
Oh! Developers may need more than a brain and a computer. And, yes, to do (4) you have to rely not on the developer but on the planning. Regarding points (1) and (2) how is getting independant contractors (from anywhere) going to increase the knowlesge of the product and the motivation.
Well food, and to know how to do stuff, but I'm speaking tangibly
here, not accurately.. You pay to support the guy but you don't
have to pay for the tools he uses, especially if all the tools needed are
already in open source.. Ton said that only a few of the developers
at NaN actually used Rational Rose.. This is no big deal but it
shows you how much mind was paid to architecting plans in the
code in a standard way.. One thing that will face this project
is having a common language and way to describe ideas.. It seems
that writing code is the preferred method, then why not pseudocode
which is not part of an implementation but can be understood by
several coders.. The other approach is to use something UML like.
That will be desperately needed if there is going to be any organization
of the plans.

Well as for your point on 1 and 2, I think that they could derive the knowledge from the net.. And for 2, I figure we find someone who is interested.. I know from hanging out with the chinese grad students
in my school, some of them had used 3DsMax before.. They said that in china the school had purchased a whole bunch of SGI workstations,
but there was no need to do graphics work on them because there wasn't much of a market for it, so the machines were used for programming work, not graphics work.. And we think of the Chinese of not having
the resources to compete with us.. Actually something like one in a thousand chinese peopel make it to college, the rest end up tending farms the rest of their lives.. So what we see of competition in these markets are the top percentile of the best engineers.. And the other thing that happens is that they take tests that determine their aptitude for a particular area for which they are good at (even if they hate it)..
So they are bunched with others who are also learning the field,
practically living in the same room, eating in the same room.. If they
have a girldfriend or boyfriend or get married they only meet them on
holidays like Xmas.. It sounds very cruel but they are bred due to the
few resources to be the best.. Who we eventually see in America are
the best of the best.. And way make our assumptions that they are
all like this, no.. Sitting back on the farm is someone who
probably could have made it to college and didn't.. But in America
we take it for granted that we have all this access to knowledge that
some people don't, because we have the Internet, we have a computer,
we have libraries, people around us who know something..

So I guess what you say is correct, we would need developers at
the blender foundation..

Mr Ton said,
5. Get full-time paid employees (in the BF) who can further coordinate and facilitate the development, at a professional level with true milestones and planning.
Yes!
Am I replying to the wrong stuff?
Sorry..

Oh you are saying a project manager.. Yeah a very smart project
manager, preferably someone who is very easy to deal with
from the executive side of things.. Wait there is no executives..
Hmm could this project manager be Ton himself?

Re: Project management is where it's at ...

Posted: Mon Jul 28, 2003 2:45 pm
by purpleblenderuser
Wow how many words a minute can you type! How much coffee you drinkin'

Ok - My position in a nutshell,
- I will _not_ pay for developers - not good for the open source model.
- I will support universities and institutes that participate in development with (very) specific graduation projects. But - I think these projects should be pushing the boundaries as much as possible, i.e. are more academic in nature. Sponsorship at best.
- I will pay for professional project management. In the long term I don't see a need to pay for this once an infrastructure has been put in place.

* The very nature of open source is that all effort no matter how small can be put forward to improve the 'product', i.e. need for good and appropriate management. How can you expect anyone to volunteer their effort when you are paying others.
* Paying for development would be expensive, i.e. not cheap. Fact. Regardless of where the developers happen to be.

The main issues are all to do with communication. A project manager does not manage the programmers - they manage the communication.

Me? I have been a developer for 10 years (C/C++, Perl, PHP, et al) within, chip design, telco, games and internet companies. Everything from parsers and CAD to helpdesk and HR systems. The last 2 years have been cantered more around project management and business process improvement. I don't have masses of experience with open source projects but do participate.

OOP (a little off topic I think) - just because you use C doesn't rule out OOP.

P.S. I am glad that in your world developers eat :wink:

Posted: Tue Jul 29, 2003 12:15 pm
by thorax
I would think that open source also changes the roles of the
project manager.. Unless source code was developed for money..
What relationship would a developer desire of a project manager
outside the confines of a commercial development headed by investors and executives.. Who's moral is the project manager trying to build, what
ideas are they trying to breakup into bite-size pieces for the developer.. I think in a open source developement the role of the project manager can't exist because its about free speech and open communication.. Project managers exist to control developers not help development along.. But a good architect is worth his/her weight in gold..

Posted: Tue Jul 29, 2003 7:14 pm
by dreamerv3
Thorax:

I am strongly opposed to the ebay style of paying per feature.
Open source is supposed to free application projects from the imposing political interests of deep pockets.

How cost effective for a corporation to decide to use blender and then bid only $40,000 for a team of developers in china to write the features they want regardless of the communitys' interests.

No, this should be a democratic system, users should vote for features and there should be lock out mechanisms to prevent employees from company "X" to come in register and rock the vote based on management orders.

Remember, we paid to free the source code, should we just stand beack and feed it to corporate america with terms which are ridiculously palettable?

You'll tell me that CVS will keep old features, but old features stagnate when money talks to developers about what money wants.
"Oh" you say the existing developer base will maintain them!

If the developers can get paid to develop on blender regardless of special interest what makes you think they'll toss the cash and work for free?

Should we really be exposing our development team to such corporate interests.

You'll tell me that this is for the greater good of blender and the community but, this can't be the best way to go about with development. If the system is based on pay per feature the door is open for big money to hijack the development and you know this.

That door should be closed, if a company wants a feature they should download the source code add it and then release the code per GPL rules.

Then the community can decide if they want to integrate this change with thier "pristine" branch.

Pay per feature sounds good on the surface and I even gave it a long hard look but it's just too dangerous, the potential for forced development changes by outside parties is just too great.

And just so you know, 3d software companies with big budgets and big clients don't always make the best software for the individual, they usually make the best software for thier clients needs based on thier clients feedback.

I think the best thing is to hold a yearly fundraiser for the Blender Foundation, they can run the servers and foundation and organization of the code with this annual drive and still be removed enough as I understand it now to not hire an bunch of coders to do some big contibutors bidding.

We should hold a vote every year to evaluate the blender foundation based on the communitys' version of a "job approval rating" and if the rating falls below a certain level people will be turned on the fact that thier annual pledges should not be going to a foundation which might not be perfroming the communitys' wishes.

I'm not insinuating that any of the foundation members are prone to such sudden influence, but the model is prone and we've got to safeguard against this.

I don't mean to sound like I'm accusing Ton or anybodyelse of corruption, I'm out @ Siggraph right now and have met up with Ton and he's just great, very altruistic as far as I think. But You have to operate on the lowest common demoninator when it comes to money and political power in dewvelopment, and that denominator says pay per feature could open a pandoras box of trouble.

A company can just as easily hire a coder to work on thier feature the important thing is GPL compliance.

Pay per feature just doesn't feel right, theres something wrong here...