Quo Vadis Radiance ... - dictators vs. volunteers

Carsten_Bauer2 · December 9, 2003, 2:28am

Greg Ward wrote

I am not a dictator ...

Ave, Greg! Why not take the chance and be one , this is no shame, others do it, too. ( remember that you live in the worlds most powerful dictatorship, Mr. Bush didn't win the elections but still somehow manoeuvered himself into the Oral - sorry, different story - Oval Office )

just some thoughts on it-
1) There has to be one - or a very small group - who sets the guidelines, otherwise things will fall apart sooner or later. But this can make it difficult for volunteers. I compare it with my artwork, I never got really big payment, but I only did what I wanted, so it was OK with me. But if I shall fulfill someone else's ideas just for free, hmm, well, that's close to slavery. Wasn't it Abraham Lincoln who forbade this?

2) The engineering/design firms mentioned make lots of bucks with their consulting projects (Jack, Rob's Boss or the guys from Arup probably will correct me here on the 'lots'...) but don't need to pay a cent for the software. This is NOT meant to blame anyone. Its the other way, generally businessmen are straightforward and realistic, so they might even prefer paying some acceptable price for a guaranteed stable version including support instead of anything free, but 'experimental', sometimes working and sometimes not. (OK, maybe I'm too optimistic ...) But I'm a realist, too. See point 1 ...

3) LBNL has losened its grip, the software is now open source, it doesn't sound as if they are willing to spend much time on it anymore. But to assure if any of the possible new developments eligible for integration doesn't taint the program and drags the name into the dirt is a tedious job and needs a considerable amount of time, too. Currently it seems that poor Greg has to do it for free, (so he's a dictator and a volunteer in one person)

.......

This are only some organisatorial side-topics. There's of course more to say, but I think I better not open another can of worms and discuss different ideas and possible strategies concerning the addon and its integration itself (separate/connected to the core, etc., etc.). There's only one solution available right now anyway.

-Carsten

Peter_Apian-Bennewi1 · December 9, 2003, 7:11am

Carsten Bauer wrote:

....

just some thoughts on it-
1) There has to be one - or a very small group - who sets the guidelines, otherwise things will fall apart sooner or later. But this can make it difficult for volunteers. ... 2) The engineering/design firms mentioned make lots of bucks with their consulting projects (Jack, Rob's Boss or the guys from Arup probably will correct me here on the 'lots'...) but don't need to pay a cent for the software.
....
3) LBNL has losened its grip, the software is now open source, it doesn't sound as if they are willing to spend much time on it anymore. But to assure if any of the possible new developments eligible for integration doesn't taint the program and drags the name into the dirt is a tedious job and needs a considerable amount of time, too. Currently it seems that poor Greg has to do it for free, (so he's a dictator and a volunteer in one person)

Currently, developpers are cross-financing their code work with using the code for their own projects, like Greg, Schorsch and yourself. Which results in very practical and efficient code. But both revenue generating project work and "strategic" programming tend to be full time jobs, so larger sw changes don't happen. Supporting other people's code is simply out of question for these folks too.
Institutions like LBNL, ISE, EPFL, IRC and DMU provide money if they happen to have funds, e.g. government/industry funded projects that run for months or a some (1-3) years. LBNL tried to get further funds for Radiance in vane for years, my experience at ISE was the same (apart from Jan Wienold's project that funded photon-map), Christoph Reinhart seems to have some luck at IRC, EPFL had apparently funded Greg to do considerable work for a year. Funds are sporadic and require considerable work to get them. Once it's there, some institutions are reluctant to give it to outsiders. Which is most efficient (hiring Greg gives the maximum ratio of Radiance results per cost) and ensures that their project is at the forefront of international research.
Volunteers are welcomed, but should be prepared to look after their code themselves to keep it alive. (In the current photon-map situation, I'd pledge for a CVS branch, but that's a different thread). At least Radiance gives them a platform that conveys programming ideas to real, professional and live usage, which is more than what happens to most university-level, academic software. If they're good and their code performs well, that is a solid reference for other industrial (and thereby better than ac paid) jobs for them. The draw back is that feedback, especially positive one, from the community is not frequent. A bit more motivation would be good for them.
Commercial users may want to fund small to medium software changes to streamline their daily work, but I don't know how often that happened. Whether one could set up a Radiance fund, that commercials contribute to voluntarily ("pay what it is worth to you" - tax detuctable) and that gives away yearly prizes for the most influential Radiance contribution, I don't know. With Radiance, commercial users can influence the software to fit their needs (try that with Autode$k).

It seems that no single way provides enough resources all by itself, so we very probably have to live with a mix, trying to coodinate the authors and communicating the current situation and new lines of thought to the wider user base. At least it is more on the move that it had been 5 yeas ago.

my 2 cents thoughts
-Peter

···

--
pab-opto, Freiburg, Germany, www.pab-opto.de

Jack_de_Valpine · December 9, 2003, 5:03pm

Hi All,

I have to bite on this one for what it is worth and probably offend lots in doing so. Let's be clear about a few things (at least insofar as I understand them).

   1. Radiance was originally developed with funds from LBNL as well as
      EPFL, along with some pieces at SGI
   2. Greg was/is the main developer of the software and has maintained
      his involvement as best possible over the years
   3. As work paid for by the government it had been released for use in
      industry with no fee for use except if the use was to build and
      sell an application with Radiance as the engine (I believe this
      required a separate license with fee from UC Berkeley/LBNL)
   4. With Greg's departure from LBNL, support for radiance fell off
      dramatically to the point where ultimately I would suggest the
      tool was unsupported (and I am sorry that I do not think
      support/development of Desktop Radiance equals support/development
      for the core radiance engine)
   5. Greg has more recently been in a position to be more involved with
      Radiance (for better or worse) now over the past few years
   6. Peter has setup and maintained radiance-online.org
   7. Radiance has been released under "open source" style license.

I had to support and do the best I could with radiance source code if we encountered problems (which we did). During these dark years, Greg and others such as Chas Ehrlich did the best they could to provide at least answers to questions when they had time as well as some patches. I am very grateful to them for their energy during this time. But again, I feel that during this time my company had made the decision to continue on with Radiance as best we could and support it as possible internally ourselves. In many ways, I think there are a lot of users who are using the tool out of dedication as well as the fact that for some kinds of applications/problems there is really nothing out there like it or with its robustness (and if there is such as Integra to name one, these tools are far too costly for the kinds of applications/services that they would be used to support).

I think most significantly with Peter's creation of radiance-online.org (6) there has been a tremendous upswing in Radiance development and use. The site has provided a central location for people to share information and exchange knowledge. This has been furthered by "open-sourcing" of radiance (7) and Greg's emergence (5) and ability to be involved in development. In addition to this, Georg Mischler's development of Rayfront has placed in him in a position to know enough about Radiance and its inner workings that he can actually have meaningful discussions with Greg about the evolution of the tools.

Although Carsten suggests he is not blaming anyone, his intimation that companies using radiance to support their service offerings "don't pay a cent for the software" does just that. I would suggest that there are many ways to return or transfer value in a community such as this one. I think that there are many of us who have contributed in our ways over the years. This has probably been as simple as asking a question that results in the discovery of a potential bug that then gets resolved, to supplying patches or even just having interesting images to show that have been generated with the software. I would gladly arrange contracts for development of specific things if and when I can get the work to support such contracts. Unfortunately the reality is that this has only happened on rare occasion.

I think given the current size of the community, it has to be up to the developers to determine how much time they are willing to contribute before saying in response to some question or idea "Hey this is great, <your name here>, but it is really beyond the scope of what I can do with donated time. Do you have any funding available to support what you are asking about? And if so are you willing to donate the resulting code development back to the community?"

-Jack de Valpine

Carsten Bauer wrote:

···

From my standpoint, there was a period, the dark years, after (4) where

Greg Ward wrote

I am not a dictator ...

Ave, Greg! Why not take the chance and be one , this is no shame, others do it, too. ( remember that you live in the worlds most powerful dictatorship, Mr. Bush didn't win the elections but still somehow manoeuvered himself into the Oral - sorry, different story - Oval Office )

just some thoughts on it-
1) There has to be one - or a very small group - who sets the guidelines, otherwise things will fall apart sooner or later. But this can make it difficult for volunteers. I compare it with my artwork, I never got really big payment, but I only did what I wanted, so it was OK with me. But if I shall fulfill someone else's ideas just for free, hmm, well, that's close to slavery. Wasn't it Abraham Lincoln who forbade this? 2) The engineering/design firms mentioned make lots of bucks with their consulting projects (Jack, Rob's Boss or the guys from Arup probably will correct me here on the 'lots'...) but don't need to pay a cent for the software. This is NOT meant to blame anyone. Its the other way, generally businessmen are straightforward and realistic, so they might even prefer paying some acceptable price for a guaranteed stable version including support instead of anything free, but 'experimental', sometimes working and sometimes not. (OK, maybe I'm too optimistic ...) But I'm a realist, too. See point 1 ...

3) LBNL has losened its grip, the software is now open source, it doesn't sound as if they are willing to spend much time on it anymore. But to assure if any of the possible new developments eligible for integration doesn't taint the program and drags the name into the dirt is a tedious job and needs a considerable amount of time, too. Currently it seems that poor Greg has to do it for free, (so he's a dictator and a volunteer in one person)

.......

This are only some organisatorial side-topics. There's of course more to say, but I think I better not open another can of worms and discuss different ideas and possible strategies concerning the addon and its integration itself (separate/connected to the core, etc., etc.). There's only one solution available right now anyway.

-Carsten

_______________________________________________
Radiance-dev mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-dev

--
# John E. de Valpine
# president
#
# visarc incorporated
# http://www.visarc.com
#
# channeling technology for superior design and construction