| Index | Recent Threads | Unanswered Threads | Who's Active | Guidelines | Search |
 |
World Community Grid Forums
Category: Support
Forum: Suggestions / Feedback
Thread: Sources / licensing |
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 13
|
|
| Author |
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
I wanted to take part in the project and was trying to find the source code for the WCG software on the website but can't find it anywhere, I'm assuming that the project is closed source?
<my obligatory (but serious) rant> That's a real shame for a project that calls itself open or scientific. I know there are more BOINC-based projects that are closed-source and it's something I really fail to understand. All the software running on my computer has been compiled on this same computer and this is rarely a limitation anymore, we're in 2008/2009 and closed-source licensing is something from the past to me and most of my friends, it's "so 90s". Yet you hit the "source" barrier when you want to take part in a "research" project. Also, looking at other posts where people running 64-bit OSes like me find themselves excluded from the project and get an answer that it's not yet really popular enough for WCG to maintain another target is just really a hint for anyone who is mildly familiar with cross-platform programming, that the software is of quality that cannot be trusted. It's thaught at schools today that properly written software is portable and independent of the target, i.e. there's no cost involved in "porting" it. Same kind of feeling (that of the authors of the software being incompetent) is provoked when the authors try to hide the sources. </rant> Please don't count me in as one asking for a 64-bit port, posting the sources like any normal project, would surely be enough. I also tried to find a mention of what the status was abut source code in the license agreement before registering to WCG, and later in the FAQ.. and found nothing. I imagine this *is* a frequently asked question so may be worth adding. While I'm at it, does the download page comply with BOINCs license? I think it's one of those that require a copy of the license text and a link to sources download to be provided (I may be wrong, am just wondering). |
|||
|
||||
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
WCG are NOT responsible for the Applications they run....the Applications belong to the Universities etc....
|
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi balrog-kun,
Our application programs are university software run (mostly) by university scientists. The exceptions, such as TSRI (Scripps) are still pretty academic. Most university software is licensed to run for free by other university research programs (or very low cost) but charges for corporate use. Scripps is exceptional in that it has released an open-source version of AutoDock (which is not the version we are using). This sort of thng may become more common in the future. Oh yes, cross-platform compiling is not very difficult. The difficulty is in the verification function. The simple way using multiple copies of work units requires a great deal of work to maintain each separate work queue on the servers. knreed keeps track of what sort of processors are asking for work. So far, he does not seem excited at the thought of maintaining 64-bit work queues. The 64-bit systems can run 32-bit application programs. Just not as fast as they could if the applications were compiled in 64-bits. Lawrence |
||
|
|
twilyth
Master Cruncher US Joined: Mar 30, 2007 Post Count: 2130 Status: Offline Project Badges: 
|
Just out of curiosity, and I'm sure this has been answered before someplace, does IBM require the source for the project apps and compile the executables itself or does it accept execs from the projects with no source. Or something in between?
----------------------------------------Is there a chance that WCG would require projects to be open source at some point in the future? I think a lot of people who are proficient programmers might want to review the code they run to see if it is using host resources as efficiently as possible. Any improvements would benefit the project as well as WCG and its users.   |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
They require the source, and a license that allows them to modify it and use it on the grid. This usually involves lawyers and complicated multi-party agreements.
Actually requiring open source software would place too heavy a restriction on the projects World Community Grid could accept. Remember, often the researchers don't own the software. Instead, open source activists need to convince the writers of the software. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Also, looking at other posts where people running 64-bit OSes like me find themselves excluded from the project and get an answer that it's not yet really popular enough for WCG to maintain another target is just really a hint for anyone who is mildly familiar with cross-platform programming, that the software is of quality that cannot be trusted. It's thaught at schools today that properly written software is portable and independent of the target, i.e. there's no cost involved in "porting" it. Same kind of feeling (that of the authors of the software being incompetent) is provoked when the authors try to hide the sources./quote] Back the truck up a minute. 1. If the code is open source, then go find it. It is not up to WCG to distribute it. The role of WCG is to crunch scientific applications, not distribute software code. 2. If the code is not open source, then your demand that it be, is irrelevant. 3. If you want a 64 port then you volunteer a year or two of your time to create the port. Don't sit on the outside and toss grenades at the staff. 4. Just because you did not get the answer to your 64 port question, does not make the answer any less relevant. 5. Just because you can't find the code does not mean that WCG is hiding the code. See #1 & #2 above. 6.The 85,000+ daily users on WCG are not using 64 OSs but I am sure that if you prove that you can more users to run WCG on one 64 OS than we already have on either Linux or Mac, then the priority of the 64 port will increase. 7. Just because you want something, does not mean that others are going to take steps to make it happen. Welcome to reality, and please let us know when you have it done. [Edit 1 times, last edit by Former Member at Jan 4, 2009 2:59:08 PM] |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi, balrog-kun
there's no cost involved in "porting" it This is true, there is almost no cost involved in simply recompiling for a different platform (although there may be unforeseen difficulties). The cost is in maintaining the port. The effort involved in supporting a platform is proportional to the number of projects that are running. Up to eight projects, all requiring an additional port to be maintained. And 64-bit doesn't just require a single extra build, it requires one for every 64-bit operating system. Simple arithmetic suddenly eats up a lot of the technicians' time - and they have no time to spare.But this isn't a final answer. 64-bit support is inevitable. As older platforms are dropped (Windows 98, and the United Devices software are gone already) space opens up for new platforms. I think your licensing question has been answered already, but I just want to remind you that the LGPL text is included in every download, which contains instructions for getting the source (it's available on the BOINC site). Finally, something maybe you didn't know: some of the research software we use is written in FORTRAN. All of it is old, with a long history of constant evolution. This is normal for scientific software. It doesn't make porting easy. |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3715 Status: Offline Project Badges: 
|
The 85,000+ daily users on WCG are not using 64 OSs That's plain wrong and I am surprised that you, a regular user of WCG forums, could write that. There are no specific 64-bit versions of the applications, but they are all running on 64-bit OSs with the appropriate libraries and many of us do it and produce up to 25-30 % more work depending on the projects they are crunching. What WCG is not ready to do is to support twice more platforms specifically for 64-bit mode, but machines running in 64-bit mode are not excluded at all. Cheers. Jean. |
||
|
|
JmBoullier
Former Community Advisor Normandy - France Joined: Jan 26, 2007 Post Count: 3715 Status: Offline Project Badges:
|
Regarding source code of applications WCG is not only requiring it, it is actually auditing it, i.e. checking that it does what is is supposed to do and that there are no other security issues. And this application auditing step accounts for much in the time it takes to launch a new project. But that is necessary for being seriously able to tell members and partners "you can trust us about these applications".
----------------------------------------Cheers. Jean. |
||
|
|
Former Member
Cruncher Joined: May 22, 2018 Post Count: 0 Status: Offline |
Hi,
there's no cost involved in "porting" it This is true, there is almost no cost involved in simply recompiling for a different platform (although there may be unforeseen difficulties). The cost is in maintaining the port. The effort involved in supporting a platform is proportional to the number of projects that are running. Up to eight projects, all requiring an additional port to be maintained. And 64-bit doesn't just require a single extra build, it requires one for every 64-bit operating system. Simple arithmetic suddenly eats up a lot of the technicians' time - and they have no time to spare. What I meant is that maintaining a port isn't costly.. as long the code is written portably (cf. correctly). You can often see software projects build their programs for a different platform (e.g. different CPU arch or a different system) and run it there, even if they don't intent the software to be used on it -- just because the different system/CPU is just another test case, and it helps a programmer find bugs in his program. The bugs might just not have surfaced on his favourite platform during his testing but might be quicker exposed on a different arch. Cross-compilation is just one of the many tools a coder has for debugging code. Ofcourse in some projects you have deadlines and you prefer to ship with potential bugs as long as they aren't too visible. However, considering what was said about IBM auditing all the code, I wouldn't expect any such problems. I suppose you know that programs written in C or in FORTRAN will run identically on any architecture as long as they don't rely on properties which are out of that language's specification (which usually is done unintentionally and regarded as a bug), (ruling out compiler bugs and cpu hardware bugs). Sorry for going off-topic. Finally, something maybe you didn't know: some of the research software we use is written in FORTRAN. All of it is old, with a long history of constant evolution. This is normal for scientific software. It doesn't make porting easy. Yes, I've talked to a person involved in a different BOINC-based project that is also closed-source. Their reasoning was that the application required quite accurate floating point arithmetics and that there existed a popular FORTRAN compiler which provided insufficient accuracy and they were afraid people might be using that compiler and be submitting bad results to them. So it was a kind of measure against having modified/miscompiled applications on people's computers. It's obviously not an ideal protection and I think this is similar to "security through obscurity". I don't know much about FORTRAN but I suppose there exists a formal specification telling you what accuracy you can rely on and beyond that your program isn't really a correct FORTRAN program. Cheers |
||
|
|
|