Re: oaklisp: contains 500kB binary in source
Marco Franzen <Marco.Franzen@bigfoot.com> writes:
> Barak Pearlmutter wrote:
>> source package; the source includes a interpreter and it would be a
>> relatively small matter to translate it from Oaklisp into RnRS Scheme.
> Correct me if I am wrong, but AIUI if someone wants to package a GPLed
> Java program that, as it is, currently runs only on a non-free VM, then
> even its source package is (at most) contrib. Even if everyone agrees
> that it would be a small matter to port it to a free VM, this is no
> consideration until someone actually does so.
Yes, this is so.
> The only difference I can see here is that you might claim that Oaklisp
> is already free. But that argument would be circular until someone
> actually does what you say is possible.
Well, clearly somebody has done it -- where do you think Oaklisp came
from? Where do you think GCC came from? It was first built with Sun
and Intel compilers, or a really awful C interpreter was used to run a
mostly awful C compiler which built itself, and then that was used to
compile GCC, and then the fully optimizing GCC built a fully optimized
GCC. This has been walked through in this thread already, but you
seem to have forgotten about it.
That bootstrapping requires either tools not packaged in Debian or
non-free tools. Those unpackaged tools are a bit useful to someone who
wants to port a compiler to a new machine, but mostly as documentation
Currently, nobody in Debian is interested in packaging those tools,
and they are not necessary to keep Debian running or to bootstrap
Debian onto a supported architecture -- you start with all the Debian
source and another free OS, like Red Hat Linux or NetBSD, and build
The question of what you do if you have *no* free infrastructure
doesn't make sense: the non-answer is "build some."
>> All source is available: if you have any doubts at all you are ideally
>> situated to verify the system's integrity.
> All source may be available or not; this is hard to tell until someone
> actually tries what you say is possible. Bootstrapping might fail
> because an Oaklisp-specific feature of the target system is subtly
> implemented by the same feature in the host system, accidentally or not.
Somebody has done it: That's where Oaklisp came from.
> Stepping back a bit, maybe the question is, which side has the burden of
> proof? The side arguing something is (or might be) free or the side
> arguing something is not (or might not be)? In the case of Java
> programs, main contains only what is free, not what might be free.
> (Again, kindly correct me if I am wrong.)
You are applying a very dissimilar example.
Brian Sniffen email@example.com