Re: copyright, license, document, standard, code, trademark, etall

[Philip Hands <phil@hands.com>]

> The You Can Do Anything License is infectious, where the GPL is not. A
> proprietary vendor can incorporate this code into proprietary licensed
> code, because there is nothing restricting him from doing so (except
> possibly his own intelligence). As soon as the proprietary author tries to
> inforce his claim on the whole package as "his", some smart lawyer will
> notice the YCDAL buried in the bowels of "his" code, and say, but, you
> don't have the right to say that about this code. There is an even chance
> the court would decide the only valid license was the YCDAL, and presto
> chango the infection is complete.

Interesting :-)  I like that.  Although I have a feeling that YCDAL is 
probably equivalent to saying ``This is in the Public Domain'', but of course 
IANAL

> One more point and we are ready to move on...
> 
> I really want to use the nested while loops Phil presented, but they
> present a quandry which can only be resolved by using the GPL.
> 
> After his posting suggesting copyright infringement for providing the
> suggestion by cutting and pasting my original post, I asked for the rights
> to use this code. Now, while Phil did finaly say that the YCDAL applied,
> he _did_ first apply the GPL. His later statement to the contrary, we are
> now suck with that GPL code until Phil can release a second, upgrade to
> the original GPL code, which it is my understanding he has the power to
> do but probably not the inclination.

I'm not sure where you got that idea --- The GPLing of that code was (as I 
said later) a joke to make a point.  If you want to interpret that to mean 
that I released the code under _both_ the GPL and the YCDAL feel free.

If I were able to assign my copyright over to you, I would be happy so to do, 
but I don't think that's allowed under IP law.

> As a result, I will shortly be releasing the new "version" of the example
> Standards Test Plan (STP) under the GPL, not only because I must, if I
> want to accept Phil's new design, but because there is no longer a reason
> not to.

I'm glad you're doing this, but I'm not sure I deserve the credit for forcing 
your hand ;-)

I did mean that the YCDAL should be taken to apply to the double loops, and
you can interpret that as whatever you like, including:

  It's now Public Domain

or 

  That the licensee must strip naked, paint themselves in woad, and 
  dance widdershins around their computer every Summer Solstice at dawn

if you like :-)

> The nagging problem that is bothering me at the moment is the library
> testing. We are looking for so links of particular names, and potentially
> validating the existance and performance of specific routines in them, but
> I am concerned that this is not enough. How do I validate that the
> "standard" library interfaces are the ones that are served up by the
> loader when programs are loaded?

Sounds like a job for a load of C code mini-test programs, like those used in 
Configure to check out what functions are available.

Unfortunately, I assume that a compiler is not a required part of the LCS 
standard system, so some sort of binary that is capable of checking out what 
dynamic library functions are available would seem to be required.

nm would tell you the symbol names in the libraries, which might get you some 
of the way there (assuming you trust the local nm)

Also, the current approach seems to be assuming that the /bin/sh is POSIX 
compliant.  Presumably, a non-POSIX /bin/sh would be grounds for rejection, 
but who knows what the shell script might do when run by a non-POSIX /bin/sh.

This seems like a good reason to make the test suite be a statically linked 
program binary.

Cheers, Phil.