Hi,
On 8/15/10, Frank Heckenbach ih8mj@fjf.gnu.de wrote:
Rugxulo wrote:
How so? Only if GPLv3 compatible!
OK, indeed, to use in GPC it would also have to be GPL compatible. But most free licenses, and especially PD, are, unless the authors added explicit clauses to prevent this.
BTW, I assume you're still loyal to GPL (and maybe GNU) even if (some?) GCC devs really haven't been that helpful directly. Obviously if you were to (help) rewrite GPC, it doesn't have to be GPL anything. You could use BSD, ISC, WTFPL, or whatever.
I blindly assume it's not anybody at FSF or GNU proper who is suggesting that GPC needs a rewrite. So they don't care. So all your worries (as is typical for a GNU project, a la GCC or Emacs) are probably unfounded. Note that XEmacs never worried as much as GNU Emacs about explicit permission / licensing in writing (legalese).
OK, you don't have millions. Me neither. ;-) But I suppose some of the companies that use GPC do (for a medium-sized company, millions in assets is not really much, you know -- but losing millions is often fatal). Do you think we should put them at risk just because you personally don't worry?
I've e-mailed Pemberton. If he ever gets back to me (doubt it), he probably won't have any decent response or proof anyways. And that's IF he cares. And even then he could be wrong. So it's never going to be perfect, and it's pretty much never going to please you. It would be okay for me, but not for you (or GNU?). My joking suggestion about plugins is probably the only way to appease both of us, and I know that won't happen!
IP isn't open source and was never declared as "public domain".
Neither was P4 (by its authors/owners) for all we know so far.
Pemberton, who legally (I assume) wrote a book about P4, explicitly says on his website that it's "public domain". Not legally airtight, but that should be enough to not have a cow, man.
I'll admit that the term "public domain" has sometimes been misused, but I don't think so here. It (P4) is only like 4000+1500 lines (5500), nothing huge.
That's about 1000 times (wild guess) as large as required for copyright protection.
My silly Befunge interpreter in Pascal is approx. 150 lines (or 25 if manually obfuscated and stripped of unnecessary crud). It's *extremely* trivial, IMHO, and I wrote it! It's released to "public domain" because:
1). I don't want to even think about licenses, they are annoying 2). I want it to be useful (well, nobody cares, but in theory ...) 3). I know it's insanely trivial and I'm a moron, so it's not worth a hill of beans
You really want to tell me that copyright applies there? Oh, and yes, I did publicly publish it, with PD notice, on the Pascal newsgroups, mostly because there were no posts except spam, *grumble grumble*. But I guess I might "change my mind" and try to enforce it when I'm (31+40=) 71, esp. since "public domain" isn't even possible in some countries (muahahahaha, profit, damages, whee!!). ;-))
I'd rather you ask FPC about whether they are (or will adapt to) GPLv3 (compatible).
Huh? What's that got to do with it? (As Jonas wrote, it is compatible, but even if it weren't, what would this mean in this context? Are you again bringing up random examples of random things?)
Damn it, Frank, give me some credit here. I'm not THAT dumb! Sheesh.
Obviously if you don't think GPC is tenable (crufty C treenode bullcrap) and P5 is too murky, you're obvious choice is FPC (GPL'd), adding yet another dialect to it, e.g. ISO 10206. It's written in Pascal, which you already said GPC being written in C wasn't helpful to attract developers.
The point was that you hadn't really (openly) considered adapting FPC, at least from what I could tell. That's surely got to be easier than rewriting everything in C++. FPC isn't perfect, but it's good and equally as popular as GPC, maybe moreso.
Are the GCC devs just ultra busy or anti-Pascal or ...? (Doubt it, but ....) Why is it so hard?? Lemme guess, RMS doesn't use Pascal.
No, AFAIK he doesn't. And he's under no obligation (neither other backend developers) to care about Pascal, just as little as we are to care about Fortran.
Right. BUT if GNU and GCC are to have separate language frontends, *somebody* has to care. If not RMS, then who? Doesn't matter if he uses Pascal, he should have some say in who does what, how they help others, etc. Okay, maybe that's not his job, but it should be somebody's.
Each language frontend is developed by those who care about it. We can blame the backend developers for (sometimes unwarranted) major API changes and lack of cooperation, but not for not doing our work.
Well, yes, obviously, I mean, it would be easier if GPC was updated to work correctly with GCC 4.x instead of abandoning it completely. But why won't they help fix the bugs? Why don't they integrate your patches? Etc. etc. It just seems dumb (to me) that they would ignore you since GPC is quite good.
Bah, then make a plugin system, a la GCC 4.5.0, and use P5 in that. ;-))
This would still put the frontend (i.e., all of our work) at risk.
How so? (Well, I was mostly joking anyways.)
- Contact Steven Pemberton and ask him if he has a documented statment of its being PD
Legal papers?? From whom? Why would he?
Because he's stating that it's PD on his website and he wrote a book about it. (Not saying he's required to have such papers, but he might since he obviously spend some time of his life with it.)
You know, dare I call it a virtue, but most people don't live their life just to be legally pedantic and screw others over. So he might just "not think that way". Hence he probably didn't think it was a big deal! (Blindly guessing.)
What if he lost it?
Then try the second alternative. (It didn't say they were guaranteed to work, just possibilities.)
Note that this doesn't mean he's lying, just that he can't 100% prove it to you.
What if there is (or was) no "legal" way for it to be declared PD even if the authors' intended that??
Which I suppose. Then all we could hope for is release under some free (possibly all-permissive) license. In which case PD would be legally incorrect, but for practial purposes close enough.
Which is really dumb, BTW, that an author / copyright holder can't do what they want with their own work just because some weird law disagrees.
Gah, I guess I should just e-mail him, but as you know, that never gets anywhere. *sigh*
Such is the tedious work usually required in issues that go back many years. I'm afraid, I can't change it, but at least the occasional success feels sweet then.
The original (DOS, 386+ only) Befunge interpreter I used wasn't mine. I hacked it up, modified (shrank) 20% of it, fixed some bugs, ported to other assemblers from TASM. Neither of the two previous authors has a website or current e-mail now (from extensive searching, at least AFAICT), and I have no idea what licenses they used (if any, can't find mirrors of their work either [MCBC, specifically]). So I just call it PD. At least my changes are PD, and I kept the original for comparison.
Oh, and I joked with myself that aPACK isn't GPL friendly (after having read some erstwhile complaints years ago, it has no unpacker since it makes a custom stub for each binary), so I challenged myself to manually shrink it to below 1024 (from original 1280) without compression (mostly for fun, I wasn't really worried). And yes, then I rewrote it completely in REXX and Pascal "for fun" (and also "just in case" anybody complains, though the 982-byte one being forced to be GPL or whatever wouldn't hurt my feelings). Well, also DOSEMU x86-64 had a bug (now fixed in SVN) where the (self-modifying) 982-byte .COM wouldn't run, heh, which was annoying. (Also annoying that it was mostly 16-bit code but with 32-bit parts. So it was neither 8086 friendly nor fully 32-bit, which makes porting to other OSes harder.)
Moral of the story, it usually doesn't hurt to hack something, esp. if nobody complains. But it's safer to write your own. (And yes, that's tedious and error-prone and not always possible.)
Scott can't even get Jim Welsh to contact him about the "model implementation". Like I said, if you try and try and try REALLY hard to contact them and they don't respond, are you still a big bastard for still using it anyways??
Yes, if by that you (knowingly and unnecessarily) put your users in danger.
It's not knowingly and unnecessarily if you did your best to resolve it, is it?? It just seems dumb to be SO pessimistic and cynical that nothing good ever gets done just because "somebody somewhere might get mad at random".
What good is software that isn't ever used again? Somebody please tell me!
You can lobby your lawmakers to change this (unfortunately, this would need to happen globally, due to Berne), but until we get there, as John said, just ignoring laws that we consider wrong won't help.
Please tell me how I'm such a horrible criminal for assuming that the authors don't care after 40 years, esp. when Pemberton says it's "public domain" and sources are freely available.
"Hey, I found $10, I put it in the bank for 40 years, made $10000, is it yours???" (Wonder what the reply will be.)
That's why it's better to ask before you made $10000 (i.e., before we base a big project on it), to avoid unpleasant surprises later.
Except in this case, we both get $10000. Nobody loses! They get all our improvements. That's why software isn't really the same as a book or movie (which don't change and don't actually accomplish anything or solve any equations or whatever). Software is meant to be used, not to sit on a shelf (though I could vaguely imagine glancing at it like Beowulf, for historical curiosity, of course less so if I can't tell what it was actually meant to do).