Hi Christoph,
I feel like I'm in a bit of an awkward spot, as I seem to have some power to decide what should and should not be put into the main Radiance distribution, and yet, not having tested these variations myself, I have very little information to go on. I know what criteria I would apply to such a decision, namely:
1) The modification should not adversely affect calculations it has nothing to do with.
2) The modification should be deemed an improvement by a majority of users.
3) The modification should compile readily on all supported platforms without requiring third-party libraries.
4) The modification should do what it claims to do, and should converge to the correct solution.
5) The modification should have minimal impact on standard Radiance file formats.
In the past, I put the things into Radiance based on user requests and my own aesthetics. Now that we are opening up the source tree to development, we need to come up with a process for including enhancements. One idea is to offer modified versions for download, and solicit feedback from users who try it as to what worked and what didn't work about it. This is sort of what we have now, but without a formal feedback channel. Getting others to validate new methods would be best of all, but this usually takes time. Provided the changes aren't too extensive or obnoxious, we could release new versions with the modifications in as conditional-compiles, and take away each condition once there is general agreement that the new calculation is doing the right thing.
I'm posting this to the Radiance developer's group, since that seems to be the proper forum for it.
-Greg
···
From: "Reinhart, Christoph" <[email protected]>
Date: Thu Feb 13, 2003 9:08:59 AM US/Pacific
To: "'Greg Ward'" <[email protected]>
Cc: "'[email protected]'" <[email protected]>, "'[email protected]'" <[email protected]>, "'[email protected]'" <[email protected]>, "'[email protected]'" <[email protected]>
Subject: RE: Perez and gendaylit compared to actual data
Hi Greg,
Thanks for the information. The new RADIANCE license agreement from LBNL is
a really positive development and I am confident that it will trigger a wave
of activities and different RADIANCE front ends. That is good. As mentioned
by numerous of our colleagues, we will all have to discipline ourselves,
that there won't be multiple RADIANCEs floating around. The Fraunhofer ISE
is planning to marry Roland Schregle's RADIANCE forward raytracer with the
few changes I implemented for the DAYSIM model in rtrace to calculate
daylight coefficients.
I think that it would make sense if we could bring our contributions somehow
into the mainstream RADIANCE. Maybe with an option in the makefile to switch
it on or off. The ISE has actually hired a real programmer to do the trick
(not a beginner like me). He could probably implement the changes so that
they are not disruptive to the rest of the program. And of course Peter
would always be around:)
This is just a suggestion. Love it or leave it. I do not want to make any
changes to RADIANCE any more as I am now working on lighting control models
etc.. I am happy with the modified rtrace as it is.
Greetings from Ottawa,
Christoph
Greg,
Has there been any more chatter about this topic?
-Chas
Greg Ward <[email protected]> wrote:Hi Christoph,
I feel like I'm in a bit of an awkward spot, as I seem to have some
power to decide what should and should not be put into the main
Radiance distribution, and yet, not having tested these variations
myself, I have very little information to go on. I know what criteria
I would apply to such a decision, namely:
1) The modification should not adversely affect calculations it has
nothing to do with.
2) The modification should be deemed an improvement by a majority of
users.
3) The modification should compile readily on all supported platforms
without requiring third-party libraries.
4) The modification should do what it claims to do, and should converge
to the correct solution.
5) The modification should have minimal impact on standard Radiance
file formats.
In the past, I put the things into Radiance based on user requests and
my own aesthetics. Now that we are opening up the source tree to
development, we need to come up with a process for including
enhancements. One idea is to offer modified versions for download, and
solicit feedback from users who try it as to what worked and what
didn't work about it. This is sort of what we have now, but without a
formal feedback channel. Getting others to validate new methods would
be best of all, but this usually takes time. Provided the changes
aren't too extensive or obnoxious, we could release new versions with
the modifications in as conditional-compiles, and take away each
condition once there is general agreement that the new calculation is
doing the right thing.
I'm posting this to the Radiance developer's group, since that seems to
be the proper forum for it.
-Greg
···
From: "Reinhart, Christoph"
Date: Thu Feb 13, 2003 9:08:59 AM US/Pacific
To: "'Greg Ward'"
Cc: "'[email protected]'" , "'[email protected]'"
, "'[email protected]'" ,
"'[email protected]'"
Subject: RE: Perez and gendaylit compared to actual data
Hi Greg,
Thanks for the information. The new RADIANCE license agreement from
LBNL is
a really positive development and I am confident that it will trigger
a wave
of activities and different RADIANCE front ends. That is good. As
mentioned
by numerous of our colleagues, we will all have to discipline
ourselves,
that there won't be multiple RADIANCEs floating around. The Fraunhofer
ISE
is planning to marry Roland Schregle's RADIANCE forward raytracer with
the
few changes I implemented for the DAYSIM model in rtrace to calculate
daylight coefficients.
I think that it would make sense if we could bring our contributions
somehow
into the mainstream RADIANCE. Maybe with an option in the makefile to
switch
it on or off. The ISE has actually hired a real programmer to do the
trick
(not a beginner like me). He could probably implement the changes so
that
they are not disruptive to the rest of the program. And of course Peter
would always be around:)
This is just a suggestion. Love it or leave it. I do not want to make
any
changes to RADIANCE any more as I am now working on lighting control
models
etc.. I am happy with the modified rtrace as it is.
Greetings from Ottawa,
Christoph
_______________________________________________
Radiance-dev mailing list
[email protected]
http://www.radiance-online.org/mailman/listinfo/radiance-dev