Please refer to the new mailman archive!
[Date Prev][Date Next]
[Chronological]
[Thread]
[Top]
Dear McStas users, What is going on between the ILL and Risoe is becoming very sad, and you probably noticed that. I can give you some details about what and how we came to the present situation: 1- In January 2001, the McStas workshop gave users the opportunity to set a useful 'to-do' list. This list is available at <http://www.ill.fr/tas/mcstas/workshop0101.html>. The workshop was really a success. McStas was version 1.4.0. 2- Following that workshop versions 1.4.1 and 1.4.2 were released nearly at the same time, essentially including new components (no kernel modification). 3- In summer 2001, I discussed with Per-Olof on how to implement new features in the McStas kernel. He said these ideas were nice. Indeed, the simulations I needed became very big (1.5 Mo source code), and difficult to compile. I had to find a solution. 3- In August 2001, Per-Olof came at ILL to see me. I had an experiment at the same time, but we managed to set a list of things to do for the next release (planned for ICNS2001). Collaboration was fair. Per-Olof made the few modifications now included in the kernel 1.5 (new macros). He proposed me to enter in the McStas project development team (Kim, Per-Olof and me). That was nice. 4- At ICNS2001, Phil Seeger noticed a bug in McStas sources. They did not focus the simulation correctly, and thus the illumination was not isotrope. 5- Coming back from ICNS2001, I knew what to do to correct this bug, and thought about how to implement most of the January workshop 'to-do' list. It was clear in my mind that these new features would be implemented in version 1.5, so that all of you would benefit from them. 6- I made all the new feature implementations within 1 week. I must say that I realized how Kristian Nielsen made a perfect work: clear, portable, deep thinking. I was glad to send my contribution to Per-Olof and Kim (end of september). I think I also realized that Per-Olof could not have done that work, as his time schedule is very heavy, and he does not know the McStas project structure as I do. I made lots of check (most components, kernel, etc...) and found bugs, probably because they have always been there. 7- My contribution was completely rejected. I do not understand why. Both the new kernel features, and the bug corrections in components were entirely ignored by my collaborators, Kim and Per-Olof. 8- The McStas version 1.5 was released (Oct 10th) by Risoe, without previous discussion with me. Getting it, it was obvious it had not been checked, and published in a hurry. There were bugs in the generated C code (could not be compiled), in the components (were buggy, specially all sources), and the documentation was incomplete, and without any component list. You probably experienced that if you upgraded to 1.5. 9- I sent e-mails immediately to Per-Olof and Kim, to tell them about the bugs I had seen (in the kernel, the components, etc.). I also told them to inform the community about the identified bugs, and that they would be corrected as soon as possible. I also proposed them to look at what I had done (I use my version 1.6-ill daily, without bugs, and corrected components) as it was there. My feeling then was that McStas users should keep their version 1.4.2 or switch to the 1.6-ill 'test' version, that I decided to put on the ILL McStas page. 10- All my e-mails were ignored. There was a temptative to correct the source inhomogeneity problem (Kim), but it then resulted in an even worse bug (neutrons with 1 m/s), probably because this was not tested. The compilation problem, that prevented any usage of McStas 1.5 was corrected at last, because there were users experiencing problems. 11- Per-Olof sent e-mails to the neutron -mc list saying that "(...) The parallel development of McStas at ILL is very unfortunate (...)". 12- Today (three weeks after release), the version 1.5 that you can find on the web is still not usable. If you finally manage to compile your instrument, there are chances that results will be wrong. Some components simply do not compile, and none of them was tested correctly. The list of bugs that I sent to Kim and Per-Olof was ignored, and bugs were hidden to the community (or rejecting the fault on the ILL if necessary), and not corrected. 13- The version 1.6-ill works fine since its release, and most components available on the ILL Mctas page have been tested. You will also find there the corrected version 1.5 components. Conclusion: I was invited to participate in the development of McStas, and finally, my work was ignored (and I was even insulted by Per-Olof). A parallel version was released, without collaboration with me. I tried my best to inform people, identify and correct bugs, but these are still there. Thus I brought an alternative solution, so that McStas users do not get angry with McStas. For diplomatic reasons, I said it was a test version. As a matter of fact, it works perfectly, and thus it is not a test version. If you want to continue using McStas, either keep version 1.4.2, or switch to the version 1.6-ill. The version 1.5 is a failure. Be careful if you use it. Hopefully, McStas is a really nice project, and is very strong. (I hope) It will not suffer from these misunderstandings. And I would like to tell all of you: keep on using McStas. It's great, and I hope you like it as I do. Also, I would like to have your opinion on the current situation and on the future of McStas. If you encounter problems with the version 1.5, please talk about it ! And ask Kim and Per-Olof to correct these bugs. I told them to, but it appears that all my words are now ignored, or returned to me with offense. What did I do wrong ? Emmanuel Farhi. Thanks in advance to all of you. I like McStas, this project is very valuable, and should continue to have your support and approval. -- What's up Doc ? -------------------------------------------- Emmanuel FARHI, http://www.ill.fr/tas/people/Farhi.html \|/ ____ \|/ CS-Group ILL4/156, Institut Laue-Langevin (ILL) Grenoble ~@-/ oO \-@~ 6 rue J. Horowitz, BP 156, 38042 Grenoble Cedex 9,France /_( \__/ )_\ Work :Tel (33/0) 4 76 20 71 35. Fax (33/0) 4 76 48 39 06 \__U_/