2010-06-01T04:47:36  *** dsulgrov has joined #opticks
2010-06-01T04:47:40  *** goffena has quit IRC
2010-06-01T04:47:45  *** goffena has joined #opticks
2010-06-01T05:01:14  *** goffena has quit IRC
2010-06-01T05:04:36  *** goffena has joined #opticks
2010-06-01T05:36:13  *** tclarke has joined #opticks
2010-06-01T05:36:13  *** ChanServ sets mode: +o tclarke
2010-06-01T05:45:16  *** kstreith has joined #opticks
2010-06-01T05:45:16  *** ChanServ sets mode: +o kstreith
2010-06-01T05:53:30  *** yuguess has joined #opticks
2010-06-01T06:01:09  <yuguess> hello , sir
2010-06-01T06:03:21  <kstreith> hello
2010-06-01T06:03:37  <yuguess> it seems I made a big mistake last week : (
2010-06-01T06:03:47  <kstreith> yes, but I noted you deleted the GitHub repo
2010-06-01T06:03:54  <kstreith> so, that is good
2010-06-01T06:04:10  <yuguess> en...should I write something like declaration paste that at the begining of my code
2010-06-01T06:05:18  <yuguess> just like most of code in Opticks
2010-06-01T06:05:28  <kstreith> when you use external code, you should always determine the license of that external code
2010-06-01T06:05:37  <kstreith> and then follow its license
2010-06-01T06:05:49  <yuguess> en, I will be very careful about that later on
2010-06-01T06:06:08  <kstreith> just give Michael and I a heads up on which matrix libraries you might use
2010-06-01T06:06:15  <kstreith> we might have some experience with said libraries
2010-06-01T06:07:35  <kstreith> and can offer some additional insight
2010-06-01T06:07:41  <yuguess> ok...I am just about to ask you about that
2010-06-01T06:08:12  <yuguess> I planed to use a math library which contains some matrix operations
2010-06-01T06:08:22  <kstreith> which one?
2010-06-01T06:08:23  <yuguess> I think maybe lapack would be OK for me
2010-06-01T06:08:26  <yuguess> lapack++
2010-06-01T06:10:03  <kstreith> those can sometimes be difficult to build
2010-06-01T06:10:26  <kstreith> you might look at uBLAS
2010-06-01T06:10:30  <kstreith> it's part of Boost, one of our existing dependencies
2010-06-01T06:10:38  <kstreith> http://www.boost.org/doc/libs/1_33_1/libs/numeric/ublas/doc/index.htm
2010-06-01T06:11:20  <tclarke> it's not as efficient as other blas/lapack libs but it is easy to use
2010-06-01T06:11:32  <tclarke> it's also fairly easy to swap these out for other implementations as they use similar APIs
2010-06-01T06:11:38  <yuguess> en...blas and lapack provide more function
2010-06-01T06:13:59  <yuguess> like, I will use a function to inverse the matrix
2010-06-01T06:16:23  <kstreith> downloading 2.5.3 of lapack++
2010-06-01T06:16:23  <kstreith> http://www.boost.org/doc/libs/1_33_1/libs/numeric/ublas/doc/index.htm
2010-06-01T06:16:27  <kstreith> sorry, wrong link
2010-06-01T06:16:27  <kstreith> http://sourceforge.net/projects/lapackpp/files/
2010-06-01T06:16:29  <tclarke> GSL is good but is GPL which would add a GPL license
2010-06-01T06:16:38  <kstreith> Lapack++ is LGPL v2.1, so that is good
2010-06-01T06:16:51  <kstreith> but to compile Lapack++ looks like you need your own "blas", "lapack" (without the ++) and a Fortran compiler
2010-06-01T06:18:23  <tclarke> my favorite is ATLAS but it is a bear to build on windows
2010-06-01T06:18:33  <yuguess> I used lapack before
2010-06-01T06:18:45  <kstreith> the largest problem I've noted with these can be building them
2010-06-01T06:18:56  <tclarke> in that case, use what you are familiar with
2010-06-01T06:19:02  <kstreith> if you can get them built and working this week, go ahead and use lapack++ or regular lapack
2010-06-01T06:20:19  <kstreith> if you spend more than a week building, I would recommend abadoning for a different library
2010-06-01T06:22:33  <yuguess> Actually, I just need to use a matrix inversion algorithm in my code and I have implemented my own version of that matrix algorithm
2010-06-01T06:22:43  <yuguess> I just worried about the performance
2010-06-01T06:24:35  <tclarke> it's generally best not to prematurely optimize
2010-06-01T06:24:46  <tclarke> it often leads to code which is difficult to maintain and often you won't know what parts need optimizing until you can do timing tests
2010-06-01T06:26:49  <mconsidi> Opticks uses ossim for matrix calculations.
2010-06-01T06:31:14  <yuguess> is that Open Source Software Image Map ?
2010-06-01T06:33:46  <tclarke> yes
2010-06-01T06:33:48  <tclarke> http://www.osgeo.org/ossim
2010-06-01T06:33:50  <tclarke> we use some of their libraries
2010-06-01T06:33:53  <mconsidi> It is Open Souce and is currently an Opticks dependency
2010-06-01T06:33:54  <tclarke> specifically some of their matrix operations and their NITF code
2010-06-01T06:44:08  <kstreith> michael is looking for an Opticks plug-in that does matrix inversion that uses Ossim to do so
2010-06-01T06:46:59  <yuguess> ok, thank you very much, that would be great help : )
2010-06-01T06:55:01  <mconsidi> yuguess: Take a look at Mnf.cpp in the spectral library
2010-06-01T06:55:10  <mconsidi> Look for invertSquareMatrix2D.
2010-06-01T06:55:36  <mconsidi> Opticks has this method in matrixFunctions.h
2010-06-01T07:03:50  <kstreith> you can Mnf.cpp here:
2010-06-01T07:03:50  <kstreith> https://extras.ballforge.net/source/browse/extras/Spectral/trunk/1.0.X/Code/Mnf/Mnf.cpp?revision=406&view=markup
2010-06-01T07:05:20  <kstreith> and here is the Mnf plug-in as part of the Spectral extension
2010-06-01T07:05:20  <kstreith> http://opticks.org/confluence/display/opticksExt/Spectral+Processing
2010-06-01T07:56:21  *** gmartin_cn has joined #opticks
2010-06-01T09:48:24  *** rforehan has joined #opticks
2010-06-01T09:48:37  *** rforehan is now known as Guest25600
2010-06-01T09:52:48  *** Guest25600 has left #opticks
2010-06-01T10:05:34  *** _rforehan has joined #opticks
2010-06-01T11:20:37  *** dsulgrov has quit IRC
2010-06-01T12:18:35  <okibot> New news from opticksspectrum: Trevor Clarke Blog: Importing multiband BSQ data from multiple files || Optick Illusions: Opticks Extension Development - Get Started In Less Than 15 Minutes || Optick Illusions: Visual Studio 2010 || Optick Illusions: Google Summer of Code || Trevor Clarke Blog: Google Summer of Code || Trevor Clarke Blog: Opticks programming contest || Optick Illusions: Mil-OSS Re-cap || Optick Illusions: Opticks and IDL - Helpful Hints || Optick Illusions: New Mouse Pan Feature || Optick Illusions: GEOINT 2009 - Come See Opticks || Optick Illusions: New Opticks Website || Optick Illusions: Presenting at Mil-OSS Working Group || Optick Illusions: Opticks 4.3.1 - It's Coming... || Optick Illusions: Mil-OSS Working Group || Trevor Clarke Blog: Putting the LT back in ELT || Optick Illusions: Opticks 101: Mastering the Import Dialog || Optick Illusions: Why is this blog called Optick Illusions? || Optick Illusions: Opticks 101: Loading Data || Trevor Clarke Blog: COAN Released for 4.3.0 || Optick Illusions: Hopes for this blog
2010-06-01T13:36:21  *** tclarke has quit IRC
2010-06-01T17:25:15  *** epifanio has left #opticks
2010-06-01T17:27:16  *** _rforehan has left #opticks
2010-06-01T17:27:34  *** rforehan has joined #opticks
2010-06-01T17:27:39  *** rforehan has left #opticks
2010-06-01T18:45:37  *** gmartin_cn has quit IRC
2010-06-01T18:55:37  <yuguess> hello
2010-06-01T20:35:33  *** yuguess has quit IRC