Re: Webpage discussing GPL x BSD licensing...

Es schrieb Roland:
> 
> http://www.softpanorama.org/Copyright/License_classification/index.shtml

interesting. some new POV mentioned. takes a strong side for bsd-is-better.
it only mentions L-GPL in rare occasions, let me quote from part 2:

[... that GPL can not quite prevent reengineering by commercial 
  entities... and in the kerberos case elaborates:]

The value of the reference codebase for Microsoft is unclear 
and that means that the question about the license for this
implementation is largely irrelevant. Moreover the mere fact 
of existence of GPLed Kerberos implementation may convince
Microsoft that that there is not much commercial advantage in 
implementing the protocol and thus Kerberos would be
languishing in obscurity. Even LGPL is problematic here. Some 
people argue that is Kerberos was released under LGPL
Microsoft would have been able to use the code just as well. 
But the resulting KERB.DLL would have been available to other
parties. That is not an attractive for Microsoft solution 
taking into account the FSF origin of LGPL and Microsoft resources.
[and goes about gpl-is-scary to MS to the conclusion that MS might
have not taken kerberos in the first place but start its own project
being only similar but then different, so] the situation  might be 
even worse that it exists today. 

[later on he says]
All this lead to the fragmentation of free software developers 
efforts and essentially endorses GPL isolationism.  All those
considerations suggest that the newly introduced term  "GPL-
compatible" looks as problematic as "Windows-compatible" and that's
not that  funny. Actually a good appetite for discussions about 
"GPL compatibility" from some members of FSF  make use even any
LGPL component in a commercial software very problematic. That 
also creates uneasiness among free software developers that use
other licenses.  [interestingly about this paragraph: he does not
bother to justify when he extends GPL-problemisms on to LGPL and 
just quotes it additionally, it just gets personal about "persons
at the FSF "for which I have to counter that there is no need to
crosslicense with the FSF or endorse their POV]

[and finally in the closing section]
The closer the product is to a reusable component (for example, 
library) the less attractive GPL became. For example the approach 
advocated in the paper suggests that GPL serves very well for gcc, 
reasonably well for Linux kernel and bad for Gnome. A well known 
version of GPL called LGPL was created specifically to fix this 
issue for the libraries and might be useful as a compromise license 
for intermediate stages of the development of the project. [where
the latter refers back to the "dynamic licensing" principle that
is actually presented in this chapter - it is the thing that this
paper actually talks about, and justifies it by looking at the
problems that exist - which is why this paper goes into some
depth with GPL x BSD licensing phenomena and therefore good 
reading for interested parties. Me just says that dyna-licensing
and dual-licensing constitutes other problems not addressed
throughly in this paper perhaps since it could torpedo THE idea]

thanks for the link, guido
(sorry for not reading the ML, it's just that I have lots of other
things to do, you known, licesiningdiscussions are not that alike
compared to some good deal of coding....)