Not a bash, just the facts

Ewan Mac Mahon ewan at macmahon.me.uk
Sun Mar 26 22:27:51 UTC 2006


On Sun, Mar 26, 2006 at 08:02:54PM +0100, Daniel Carrera wrote:
> Ewan Mac Mahon wrote:
> >It works OK for the other Free languages like perl, python, ruby, and
> >indeed the gcc implementation of C/C++.
> 
> Microsoft doesn't feel threatened by any of those.
>
They don't feel threatened by Java either - once, maybe, but not any
more. Java is old enough now that the status quo is set. At the time MS
created their warped version of Java they had a hope of locking people
into their Java, and hence their platform; that hope is gone. While MS
could legally take a Free Java and make it incompatible it wouldn't help
them - the mass of existing Java software and expertise is too great.
 
> Also, the first three don't have the position that Java has in the
> market. You don't use Perl for very large mission-critical programs
> that must work just as well in 5 years.

Amazon's been going for more than five years, and lots of not most of it
is in perl. Judging by their job adverts quite a lot of the bbc.co.uk
infrastructure is in perl too. Perl was once the overwhelmingly dominant
language of cgi, and it remains extremely common.

> And Python and Ruby are largely unknown.
> 
> Your gcc example is flawed. The C specification is not GPL. GCC is to
> C what Kaffe is to Java. GCC is just a FOSS compiler that tries to
> adhere to the C spec just like Kaffe is a FOSS compiler that tries to
> adhere to the Java spec.
> 
> BTW, the reason why the gcc example was flawed by the first ones
> weren't is that Perl, Python and Ruby aren't specs. Their
> implementations are the spec, so you can say that the spec is
> forkable. And forkability is what Sun wants to avoid with Java. Java
> is not forkable in the same way that C is not forkable.
> 
There's something too what you say, but I think you're arguing against
yourself - if Java exists as a specification independent of the
implementation (which it does, AIUI) then Sun could free their
implementation while still maintaining the spec; it would be in the
interests of anyone hacking on the released java source to keep to the
spec in the same way it is in the interests of the GCC hackers to keep
to the C specification (and indeed as it is of the gcj hackers to keep
to the Java spec) . It would also allow the benefits of software freedom
such as porting to other architectures that are presently denied to Java
programmers.


> >No single distribution ships a weirdly broken implementation of any
> >of these languages[1]
> 
> But Microsoft might. It is not FOSS that Sun is afraid of regarding
> Java.
> 
They could, but if they broke it so that it couldn't run existing Java
code no-one would use it. If they broke it so that it incuded new
extensions then programmers would only use those extensions if they were
useful - if that happened Sun, Ubuntu and everyone else would be able to
pick up the extensions since MS would be obliged to release their
modifications. I'm assuming a strong copyleft licence like the GPL, of
course, rather than something BSD-like that would allow closed
modifications. Either way there's no prospect of vendor lock-in.

> >If someone comes up with an extension that actually seems useful then
> >everyone can adopt it;
> 
> FOSS doesn't work that way. For example, Ubuntu and Debian are not
> binary compatible.
>
I don't see the connection; my point was that if (say) Ubuntu added a
whizzy new feature to Python then Debian and everyone else would be able
to pick up the source to that new feature and include it in their
Pythons, thereby keeping the implementations compatible. That doesn't
require binary compatibility, just source compatibility.


Ewan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <https://lists.ubuntu.com/archives/ubuntu-users/attachments/20060326/340b86d0/attachment.pgp>


More information about the ubuntu-users mailing list