Announcement

Collapse
No announcement yet.

PathScale Open-Sources The EKOPath 4 Compiler Suite

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Originally posted by allquixotic View Post
    Unless you define open source differently than I do, a BSD-licensed project is every bit as open source as EkoPath. But, if the conclusions that I and other posters in this thread have derived are at all true, then we can not distribute our project's compiled binaries if we build them with EkoPath, even if we deliberately include exactly the same source code used to produce the binaries. So it seems that you speak in terms of "open source", and allude to being "open source-friendly", but a project that is above-board and openly willing to remain open source indefinitely, is unable to take full advantage of your software. That's my understanding, though I could be wrong.
    I'm not sure i understand what you're saying here. I don't think there would be any problem with a BSD-licensed project. You would compile it with EkoPath and then you are correct the binary would be licensed GPL, while all the source code would remain BSD. But then shouldn't you be able to distribute those binaries? They'd just require you to make the source available, etc. like the GPL requires. You'd almost certainly want your project to maintain compatibility with another compiler in order to allow avoiding the GPL, but I don't think it would actually stop you from doing anything in Ekopath if you wanted to add another option.

    Comment


    • To help clarify the license question

      1) GCC injects small pieces of code in your application which are GPLv3 licensed (Our crt* code, which is *very* small, isn't GPL)

      2) GCC runtimes are GPLv3 in part (I assume some (all?) may be LGPL. In any event ours are not)

      Once again I am not a lawyer and don't take this as legal advice.


      My new signature
      +
      +
      v
      The opinions expressed here are not approved, reviewed, acknowledged nor do they represent my current employer.

      Comment


      • Originally posted by codestr0m View Post
        To help clarify the license question

        1) GCC injects small pieces of code in your application which are GPLv3 licensed (Our crt* code, which is *very* small, isn't GPL)

        2) GCC runtimes are GPLv3 in part (I assume some (all?) may be LGPL. In any event ours are not)

        Once again I am not a lawyer and don't take this as legal advice.


        My new signature
        +
        +
        v
        The opinions expressed here are not approved, reviewed, acknowledged nor do they represent my current employer.
        That's seriously flawed.

        Comment


        • Originally posted by bug! View Post
          That's seriously flawed.
          the real question is - if it isn't GPL or LGPL, technically what are those bits code licensed under??

          ..and if it turns out to be totally proprietary, then this compiler suite isn't going to be used much i dont think. maybe phoronix should do some investigative reporting on the subject...?

          Comment


          • Originally posted by bug! View Post
            That's seriously flawed.
            only if you want to compile closed source software while being too cheap to buy a licence.

            Comment


            • Originally posted by energyman View Post
              only if you want to compile closed source software while being too cheap to buy a licence.
              erm, no... From what i can tell, the real problem here is potentially not being able to redistribute (opensource) binaries compiled with EKOPath's compiler. making this compiler pretty much useless to most distributions (not source-based distro's or Arch though), and also useless 4 FLOSS developers. Which is very similar to Intel XE compiler restrictions... Intel's compiler is free for personal/non-commercial use, but needs to be licensed for commercial use (or even redistribution of compiled code any kind). it would seem EKOPath is almost the same thing, except (most) of the code is licensed under free licenses, it's opensource and the compiler itself is re-distributable. (unlike XE/icc).

              on top of that if you do require a license for commercial use - to be able to redistribute code compiled with it - you're probably looking at paying the license fee listed on the PathSCale website for the suite a whopping $1750 US, that may be cheap for a decent size company, but not for an individual developer in free software, it's not by any stretch of the imagination...

              so far, XE compiler seems to be way more compatible with GCC than EXOPath, as in i can't get all sorts of crap to compile with EKOPath than i can easily compile with XE and judging by benchmarks of binaries produced with both, it would seem that for most stuff ICC gets similar performance... a commercial license from Intel also costs much less than what 'potentially' a PathScale license would cost.

              are you willing to pay almost $2000 Us for a compiler?!?!?

              im not.

              codestrom might be able to verify, whether this is true or not - but it sounds like i might not be too far off-target on what i think the potential problems with EKOPath might be.

              Comment


              • Originally posted by ninez View Post
                erm, no... From what i can tell, the real problem here is potentially not being able to redistribute (opensource) binaries compiled with EKOPath's compiler. making this compiler pretty much useless to most distributions (not source-based distro's or Arch though), and also useless 4 FLOSS developers. Which is very similar to Intel XE compiler restrictions... Intel's compiler is free for personal/non-commercial use, but needs to be licensed for commercial use (or even redistribution of compiled code any kind). it would seem EKOPath is almost the same thing, except (most) of the code is licensed under free licenses, it's opensource and the compiler itself is re-distributable. (unlike XE/icc).

                on top of that if you do require a license for commercial use - to be able to redistribute code compiled with it - you're probably looking at paying the license fee listed on the PathSCale website for the suite a whopping $1750 US, that may be cheap for a decent size company, but not for an individual developer in free software, it's not by any stretch of the imagination...

                so far, XE compiler seems to be way more compatible with GCC than EXOPath, as in i can't get all sorts of crap to compile with EKOPath than i can easily compile with XE and judging by benchmarks of binaries produced with both, it would seem that for most stuff ICC gets similar performance... a commercial license from Intel also costs much less than what 'potentially' a PathScale license would cost.

                are you willing to pay almost $2000 Us for a compiler?!?!?

                im not.

                codestrom might be able to verify, whether this is true or not - but it sounds like i might not be too far off-target on what i think the potential problems with EKOPath might be.
                More useful than codestr0m or anyone from PathScale would be to get the Free Software Foundation to evaluate the license landscape around EKOPath and provide an opinion to the general public. The FSF has real lawyers (or more accurately, the Software Freedom Law Center) who can make a reliable determination. Any public opinion they release won't constitute personal legal advice, of course, but generally their findings are reliable enough that the free/open source community developers can take their word for it without needing to pay for a personal lawyer. Just look at the number of people who take their advice regarding projects that are less ambiguously licensed under the licenses in the FSF's license directory. I'm not saying they are incapable of error, just that their opinion gives me about the same level of confidence as I'd get from hiring a lawyer for personal legal advice. And since this subject matter pertains to software licensed under one or more free software licenses, they have an intrinsic reason to be interested.

                codestr0m already said that (1) he's not a lawyer and doesn't like to talk about licenses and (2) the information he *can* provide is extremely standoffish, and from what I can tell it boils down to "release your software under the GPLv3 or else contact your lawyer and/or buy a commercial license". This may, in fact, be what the decision boils down to, but I'd rather hear it from the SFLC than from him.

                I'll send them an email and inquire about it.

                Comment


                • Originally posted by allquixotic View Post
                  More useful than codestr0m or anyone from PathScale would be to get the Free Software Foundation to evaluate the license landscape around EKOPath and provide an opinion to the general public. The FSF has real lawyers (or more accurately, the Software Freedom Law Center) who can make a reliable determination. Any public opinion they release won't constitute personal legal advice, of course, but generally their findings are reliable enough that the free/open source community developers can take their word for it without needing to pay for a personal lawyer. Just look at the number of people who take their advice regarding projects that are less ambiguously licensed under the licenses in the FSF's license directory. I'm not saying they are incapable of error, just that their opinion gives me about the same level of confidence as I'd get from hiring a lawyer for personal legal advice. And since this subject matter pertains to software licensed under one or more free software licenses, they have an intrinsic reason to be interested.

                  codestr0m already said that (1) he's not a lawyer and doesn't like to talk about licenses and (2) the information he *can* provide is extremely standoffish, and from what I can tell it boils down to "release your software under the GPLv3 or else contact your lawyer and/or buy a commercial license". This may, in fact, be what the decision boils down to, but I'd rather hear it from the SFLC than from him.

                  I'll send them an email and inquire about it.
                  that's a really great idea, you should totally do that if it pans out, post a link so everyone else can find out what the deal really is.

                  Comment


                  • Originally posted by ninez View Post
                    that's a really great idea, you should totally do that if it pans out, post a link so everyone else can find out what the deal really is.
                    OK, I sent them an email with a directed set of questions and a link to the repository and the press release. Hopefully they'll release an opinion on the SFLC Blog. Check frequently at http://softwarefreedom.org for the result. I will contact Michael at his email address as soon as I hear back from the SFLC.

                    Comment


                    • Originally posted by allquixotic View Post
                      OK, I sent them an email with a directed set of questions and a link to the repository and the press release. Hopefully they'll release an opinion on the SFLC Blog. Check frequently at http://softwarefreedom.org for the result. I will contact Michael at his email address as soon as I hear back from the SFLC.
                      awesome stuff, I love your pro-active nature - it's sweet

                      i'll keep an eye out for tan update.

                      thanks again, and cheerz!
                      Last edited by ninez; 07-09-2011, 02:20 PM.

                      Comment

                      Working...
                      X