On 2011-02-15, David Kastrup <dak(a)gnu.org> wrote:
Julian Bradfield <jcb+xeb(a)jcbradfield.org> writes:
But that's not what "viral" or "contaminate" implies. Those
some action _not_ under control and will of the coder. They also imply
some unexpected restriction. But the _expected_ outcome using
"non-viral" "non-contaminating" code is that you may neither
nor even run the resulting code. What is more of a contamination: that
You must be buying the wrong licences for your commercial code.
> If you include GPL code, and distribute your software, it must
> under the GPL. And it's not "based on". The GPL defines a modified
> work as anything that involves copying in a sense controlled by
> copyright law.
The GPL does not define copyright law.
I don't understand your point.
Sure, but the reason is not that the GPL is viral or contaminating,
that copyright law is an abomination. He'd have to be careful about any
I'm quite happy to agree that copyright law is a mess, and personally
I think we could do without it entirely, and I'm happy to campaign for
its abolition. "Abomination" suggests a quasi-religious moralistic
mentality which I don't share.
> Such as distributed closed-source software and charging for
Oh, that's normal with proprietary software licenses? Interesting. So
Yes. You want to use a widget kit for your proprietary software, you
go and buy a developers' licence for a proprietary widget set. It will
allow you to distribute and sell your software under your own terms,
as long as you don't infringe on their rights, and may or may not
allow you to include a users' licence for the kit in your
product. Motif was a well known example in the Unix world until it
show me, say, the modified versions and forks of Microsoft Word that
distributed legally like XEmacs, a fork of Emacs.
As far as I'm aware, Microsoft doesn't sell licences that allow
this. They're a very possessive company. In the good old days, people
used to distribute modifications for IBM mainframe operating systems -
admittedly more because IBM hadn't yet realized they could charge
money for software than hardware.
Sorry, but you don't just "uninfect" or
"uncontaminate" during genetic
engineering if you feel like it.
Possibly the analogy is getting a bit stretched. I'm sure that within
a few years you will - that is, there will viruses engineered with an
easy kill-switch in them. It's already a DARPA project.
> But the only reason I can do that is that I write all my own code
> the projects in question), and don't re-use anybody else's, and only
> rely on LGPL libraries.
But you place parts into XEmacs. According to your word use, that
should infect and contaminate them.
Any code I submit to XEmacs is of course GPLed. It's all so trivial
that I wouldn't even bother to claim ownership. I *could* release the
code fragments under some other licence, but what sense would that make?
I'm talking about my own projects, not XEmacs.
So what point is there in throwing around bad insinuation which
know to be inappropriate?
It's called "bonding". Mutual disparagement is a common technique,
at least among males (*are* there any women on this list?). Listen
in any playground, locker-room or club. Disparaging outsiders is an even
commoner one, especially the "outsiders" from the next village -
that's why the XEmacs source is stuffed full of obscene insults to the
FSF. I think the current developers are all too nice and too
professional to do that, but I'm not so nice, and I've been reading an
awful lot of that code recently.
You'll find that this is not the normal BSD proponent's
are not bothered about proprietary rereleases, but are quite annoyed to
see parts of their software crop up in GPLed software.
With the BSD licence satisfied, or not?
Yes, this is a nuisance for XEmacs development. But it is not like
It would be much more of a nuisance for me. If I wanted to include a
single piece of GPL3 code, I would have to do all the work on the 21.4
codebase that Stephen and co. do on the 21.5 codebase in order to release
my fork (which I do intend to do once it has all the features I
consider essential). And then SXEmacs would be unable to use it if
they wanted to.
This is, as you say, because of the incompatibility, but that
incompatibility is contaminatory in nature because the GPL is required
to apply at the level of program, rather than at the level of module
or source file. Again, a deliberate decision to ensure that the GPL
spreads anywhere it is planted (to use Stallman's preferred analogy).
He said "spider plant"; others might use "giant hogweed" or "cane
The more you repeat the propaganda of proprietary software
the more harm you do to Stephen and the XEmacs community for arriving at
a version of XEmacs that does everything you can expect it to do given
the available material upstream.
I'm not really interested in upstream. I'm not a heavy emacs user, I
just use it for editing (multilingual) text and mail. As I have said,
21.4 did everything I need apart from Unicode.
It doesn't appear to me that Emacs is going anywhere fast, either.
They are mature pieces of software.
XEmacs-Beta mailing list