Hello all,
I received a reply from Ton Roosendaal who runs the Blender Foundation. He shared his experiences with grants, and said I could share his thoughts with you.
I agree with Ton that if we are going to fund projects that a professional atmosphere, proper administration and management, will help to achieve better results. We might consider funding research projects administered by a University, or managed by a company.
Ron Teitelbaum
From: Ton Roosendaal
Hi,
From Blender Foundation experience:
We've participated twice in the Google Summer of Code, grants for students to work for two months on a coding project. Results of this is very mixed; it mostly depends on the professional attitude of a student. A downside of this approach is that it divides volunteers/hobbiests a bit... students get paid for what others do for free. That's why it is perceived like a lottery; some people just got the luck to be granted.
We've also hired a couple of times active volunteers to do servicing (website, development support). In almost all cases, the contributions they did while getting paid was less (quantitive as well as in quality) than what they did for free.
In 2004, a student of the Amsterdam University graduated on a research on this topic, the results of a survey she did in the Blender community is summarized here: http://download.blender.org/documentation/bc2004/Martine_Aalbers/ results-summary.pdf I've asked her to also look at how financial rewards would work in our projects. Her conclusion was that this has the danger of diminishing motivation. In scientific research on other communities, this is called "crowding out". Her entire paper is unfortunately only available in Dutch: http://download.blender.org/documentation/bc2004/Martine_Aalbers/ MartineAalbers.pdf
As an alternative, I then decided to experiment with another approach. This became the "Orange Open Movie" project, which has resulted in the 3D animation short "Elephants Dream". The target was to establish a temporal but highly professional studio in Amsterdam, and invite key members of the community to come over to work for half a year on realizing a movie short. That project worked out great in all aspects. It helped Blender development, it helped our 'brand', it increased commitment from the active volunteers as well as from professionals.
This leads to a separation of two groups of contributors to Blender; - volunteers: people who contribute to Blender without getting paid. - professionals: people who contribute to Blender as part of their daytime job.
It's important to realize that 'volunteers' still can be highly professional in their contributions. For example a 3D developer working for company XXX can contribute to Blender in his spare time still. This seperation is not about quality of work, but about differencing ways for how to support contributors.
My current conclusion is that - when money gets involved - it's important to participate in an existing professional environment, or to create one yourself (like we did for studio Orange), or to help people to setup a business to become 'professional'. We didn't give out large grants yet, but if we will do I would look at sponsoring companies/organizations to hire Blender developers/artists for projects.
For volunteers, what works quite well is more incidental support: - sponsored hardware (we got boards from ATI and Nvidia for example) - small grants for creating documentation (or helping creating books) - organize events, and give volunteers expenses coverage, free access, drinks/dinners - art/movie festivals with prizes
-Ton-
Ron Teitelbaum wrote:
Hello all,
I received a reply from Ton Roosendaal who runs the Blender Foundation. He shared his experiences with grants, and said I could share his thoughts with you.
I agree with Ton that if we are going to fund projects that a professional atmosphere, proper administration and management, will help to achieve better results. We might consider funding research projects administered by a University, or managed by a company.
Ron Teitelbaum
From: Ton Roosendaal
Hi,
From Blender Foundation experience:
We've participated twice in the Google Summer of Code, grants for students to work for two months on a coding project. Results of this is very mixed; it mostly depends on the professional attitude of a student. A downside of this approach is that it divides volunteers/hobbiests a bit... students get paid for what others do for free. That's why it is perceived like a lottery; some people just got the luck to be granted.
We've also hired a couple of times active volunteers to do servicing (website, development support). In almost all cases, the contributions they did while getting paid was less (quantitive as well as in quality) than what they did for free.
In 2004, a student of the Amsterdam University graduated on a research on this topic, the results of a survey she did in the Blender community is summarized here: http://download.blender.org/documentation/bc2004/Martine_Aalbers/ results-summary.pdf I've asked her to also look at how financial rewards would work in our projects. Her conclusion was that this has the danger of diminishing motivation. In scientific research on other communities, this is called "crowding out". Her entire paper is unfortunately only available in Dutch: http://download.blender.org/documentation/bc2004/Martine_Aalbers/ MartineAalbers.pdf
As an alternative, I then decided to experiment with another approach. This became the "Orange Open Movie" project, which has resulted in the 3D animation short "Elephants Dream". The target was to establish a temporal but highly professional studio in Amsterdam, and invite key members of the community to come over to work for half a year on realizing a movie short. That project worked out great in all aspects. It helped Blender development, it helped our 'brand', it increased commitment from the active volunteers as well as from professionals.
This leads to a separation of two groups of contributors to Blender;
- volunteers: people who contribute to Blender without getting paid.
- professionals: people who contribute to Blender as part of their
daytime job.
It's important to realize that 'volunteers' still can be highly professional in their contributions. For example a 3D developer working for company XXX can contribute to Blender in his spare time still. This seperation is not about quality of work, but about differencing ways for how to support contributors.
My current conclusion is that - when money gets involved - it's important to participate in an existing professional environment, or to create one yourself (like we did for studio Orange), or to help people to setup a business to become 'professional'. We didn't give out large grants yet, but if we will do I would look at sponsoring companies/organizations to hire Blender developers/artists for projects.
For volunteers, what works quite well is more incidental support:
- sponsored hardware (we got boards from ATI and Nvidia for example)
- small grants for creating documentation (or helping creating books)
- organize events, and give volunteers expenses coverage, free access,
drinks/dinners
- art/movie festivals with prizes
-Ton-
Sorry for the last email.
Ron, that was very interesting. Thanks for asking the question. Too bad Martine's full paper isn't available in English.
I wonder how the "crowding out" effect differs between working on one's pet project, and doing something that is not really fun, but is necessary. For example, I'm not intrinsically motivated to write tools to support harvesting, so there would probably be no negative effect from receiving an extrinsic reward.
I think the idea of funding a production using Blender is a great idea. I'm not sure what the analogue is for Squeak, though. I can think of some things (eg: building tools for collaborative development in Croquet), but everyone will have their own idea, and there seems nothing so clear-cut as using Blender to make a movie.
Josh
I think we should try to figure out what is going to upset people, and will funding projects alienate people in this community and keep them from participating.
I believe that Ton has hit on something that I think will help. It would seem to me that what will upset people more then anything else is spending donated money and not receiving back proper value for that money. If we do receive back proper value for the investment then everyone benefits since we all use the software.
There could be other problems like a sense of favoritism, jealousy or unfairness. For example someone could think they could do a better job then the person selected for a grant or they are working harder then the person getting paid without receiving anything for their contribution so payment in general is not fair.
I believe that by raising the bar, by insisting on a professional management type atmosphere, that has real structure and accountability, most people would feel more comfortable about this type of grant.
The three major things that should happen are. We should agree on what task or event we want, and what the value of that task or event is to the community. Then we should partner with someone that is willing to administer the project, either a business or educational institution that may or may not have a financial stake in the project and may or may not receive some benefit themselves. At that point we need to have highly qualified community members screen applicants so that we can all be assured that the candidate can provide the value we are looking for. The process should be open and transparent with frequent updates as to status.
I think we could all benefit from more movement and development. For me it's a good idea.
Ron Teitelbaum
From: Joshua Gargus Sent: Tuesday, August 29, 2006 11:38 AM
I wonder how the "crowding out" effect differs between working on one's pet project, and doing something that is not really fun, but is necessary. For example, I'm not intrinsically motivated to write tools to support harvesting, so there would probably be no negative effect from receiving an extrinsic reward.
Josh
On Aug 29, 2006, at 10:10 AM, Ron Teitelbaum wrote:
The three major things that should happen are. We should agree on what task or event we want, and what the value of that task or event is to the community. Then we should partner with someone that is willing to administer the project, either a business or educational institution that may or may not have a financial stake in the project and may or may not receive some benefit themselves. At that point we need to have highly qualified community members screen applicants so that we can all be assured that the candidate can provide the value we are looking for. The process should be open and transparent with frequent updates as to status.
So, here's a quick experiment - maybe more along the lines of a bounty than a grant. I would love for an open source libCURL plugin to exist. If there's anyone out there who would like some funding to do that work, contact me and I'll arrange it.
Cheers, Avi
Avi, I'd like to participate in the experiment (not that i believe that this an easy way to get beer, but i think squeakers should start doing such kind of stuff :)). So, here is the initial version of the libcurl binding for the start: http://minnow.cc.gatech.edu/squeak/5865 I've compiled it on Windows, I think that it should compile normally on other platorms (one needs to have a working vwmaker setup and libcurl and openssl libs&headers installed, obviously). Making it robust and feature-complete will require additional effort, i may do it myself or give it away to anybody who wishes
regards, Danil
On Aug 29, 2006, at 10:10 AM, Ron Teitelbaum wrote:
The three major things that should happen are. We should agree on what task or event we want, and what the value of that task or event is to the community. Then we should partner with someone that is willing to administer the project, either a business or educational institution that may or may not have a financial stake in the project and may or may not receive some benefit themselves. At that point we need to have highly qualified community members screen applicants so that we can all be assured that the candidate can provide the value we are looking for. The process should be open and transparent with frequent updates as to status.
So, here's a quick experiment - maybe more along the lines of a bounty than a grant. I would love for an open source libCURL plugin to exist. If there's anyone out there who would like some funding to do that work, contact me and I'll arrange it.
Cheers, Avi
--
That's terrific!!!
Ron Teitelbaum
-----Original Message----- From: squeak-dev-bounces@lists.squeakfoundation.org [mailto:squeak-dev- bounces@lists.squeakfoundation.org] On Behalf Of danil osipchuk Sent: Monday, September 04, 2006 7:52 AM To: The general-purpose Squeak developers list Subject: Re: libCURL binding (was Re: The success of Grants)
Avi, I'd like to participate in the experiment (not that i believe that this an easy way to get beer, but i think squeakers should start doing such kind of stuff :)). So, here is the initial version of the libcurl binding for the start: http://minnow.cc.gatech.edu/squeak/5865 I've compiled it on Windows, I think that it should compile normally on other platorms (one needs to have a working vwmaker setup and libcurl and openssl libs&headers installed, obviously). Making it robust and feature-complete will require additional effort, i may do it myself or give it away to anybody who wishes
regards, Danil
On Aug 29, 2006, at 10:10 AM, Ron Teitelbaum wrote:
The three major things that should happen are. We should agree on what task or event we want, and what the value of that task or event is to the community. Then we should partner with someone that is willing to administer the project, either a business or educational institution that may or may not have a financial stake in the project and may or may not receive some benefit themselves. At that point we need to have highly qualified community members screen applicants so that we can all be assured that the candidate can provide the value we are looking for. The process should be open and transparent with frequent updates as to status.
So, here's a quick experiment - maybe more along the lines of a bounty than a grant. I would love for an open source libCURL plugin to exist. If there's anyone out there who would like some funding to do that work, contact me and I'll arrange it.
Cheers, Avi
--
On Sep 4, 2006, at 4:51 AM, danil osipchuk wrote:
Avi, I'd like to participate in the experiment (not that i believe that this an easy way to get beer, but i think squeakers should start doing such kind of stuff :)). So, here is the initial version of the libcurl binding for the start: http://minnow.cc.gatech.edu/squeak/5865'
Awesome. It looks like this blocks the VM for now, right? It would of course be most-useful to get the non-blocking API working... but in the meantime, where do I send the beer?
I've compiled it on Windows, I think that it should compile normally on other platorms (one needs to have a working vwmaker setup and libcurl and openssl libs&headers installed, obviously). Making it robust and feature-complete will require additional effort, i may do it myself or give it away to anybody who wishes
Well, to flesh out the experiment: I'll put up a $1500 USD bounty for the finished version (including compiled plugins for Mac/Windows/ Linux). Danil, if others come forward to help, I nominate you to do the split.
Cheers, Avi
I'll gladly help with the Mac plugin compile once it's a bit further along.
On 5-Sep-06, at 3:11 PM, Avi Bryant wrote:
On Sep 4, 2006, at 4:51 AM, danil osipchuk wrote:
Avi, I'd like to participate in the experiment (not that i believe that this an easy way to get beer, but i think squeakers should start doing such kind of stuff :)). So, here is the initial version of the libcurl binding for the start: http://minnow.cc.gatech.edu/squeak/5865'
Awesome. It looks like this blocks the VM for now, right? It would of course be most-useful to get the non-blocking API working... but in the meantime, where do I send the beer?
I've compiled it on Windows, I think that it should compile normally on other platorms (one needs to have a working vwmaker setup and libcurl and openssl libs&headers installed, obviously). Making it robust and feature-complete will require additional effort, i may do it myself or give it away to anybody who wishes
Well, to flesh out the experiment: I'll put up a $1500 USD bounty for the finished version (including compiled plugins for Mac/ Windows/Linux). Danil, if others come forward to help, I nominate you to do the split.
Cheers, Avi
-- ======================================================================== === John M. McIntosh johnmci@smalltalkconsulting.com Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ======================================================================== ===
Ok, even a part of that is certainly enough for writing on my back as a subject for contemplation on week-ends for my wife :) And sure - non-blocking version is a requirement for the release of "non-beer" version. I think this thing will not be abandoned ;)
Danil
I'll gladly help with the Mac plugin compile once it's a bit further along.
On 5-Sep-06, at 3:11 PM, Avi Bryant wrote:
On Sep 4, 2006, at 4:51 AM, danil osipchuk wrote:
Avi, I'd like to participate in the experiment (not that i believe that this an easy way to get beer, but i think squeakers should start doing such kind of stuff :)). So, here is the initial version of the libcurl binding for the start: http://minnow.cc.gatech.edu/squeak/5865'
Awesome. It looks like this blocks the VM for now, right? It would of course be most-useful to get the non-blocking API working... but in the meantime, where do I send the beer?
I've compiled it on Windows, I think that it should compile normally on other platorms (one needs to have a working vwmaker setup and libcurl and openssl libs&headers installed, obviously). Making it robust and feature-complete will require additional effort, i may do it myself or give it away to anybody who wishes
Well, to flesh out the experiment: I'll put up a $1500 USD bounty for the finished version (including compiled plugins for Mac/ Windows/Linux). Danil, if others come forward to help, I nominate you to do the split.
Cheers, Avi
--
=== John M. McIntosh johnmci@smalltalkconsulting.com Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ======================================================================== ===
Cool! I applaude loudy (in silence).
Stef
I've compiled it on Windows, I think that it should compile normally on other platorms (one needs to have a working vwmaker setup and libcurl and openssl libs&headers installed, obviously). Making it robust and feature-complete will require additional effort, i may do it myself or give it away to anybody who wishes
Well, to flesh out the experiment: I'll put up a $1500 USD bounty for the finished version (including compiled plugins for Mac/ Windows/Linux). Danil, if others come forward to help, I nominate you to do the split.
Cheers, Avi
squeak-dev@lists.squeakfoundation.org
-
Avi Bryant -
danil osipchuk -
John M McIntosh -
Joshua Gargus -
Ron Teitelbaum -
stéphane ducasse