Re: Processor requirements

[Michal Suchanek]

On 11/21/05, Ludovic Courtès <address@hidden> wrote:
> Hi,
>
> Bas Wijnen <address@hidden> writes:
>
> > Yes, just like 32 bit CPUs. ;-)
> >
> > We all know what trouble the ext2 translator has because of them.  I think 
> > we
> > should learn from this that we should not depend on something becoming 
> > really
> > obsolete.
>
> [...]
>
> > I fully agree that we don't need support for the old hardware initially.  
> > But
> > we shouldn't make the mistake to design critical parts of the system in ways
> > that cannot be ported without a complete rewrite of large parts of code.
>
> These two statements seem contradictory.  According to your second
> paragraph, mapping a whole filesystem into the ext2 server's address
> space can be seen as a mistake because it "cannot be ported [to 32-bit
> hardware] without a complete rewrite of large parts of code".

This is not contradictory. It was probably assumed it will be good
enough on 64bit cpus but majority of people is still using 32bit cpus
today.

>
> Besides, I think that Free Software in general *must* support "old"
> hardware.  At least, it shouldn't participate in the everlasting rush to
> new hardware.  That's an ethical problem.

Well, i686 is not really that new.
If it turns out that some features not present on earlier CPUs are
critical for the design it is not that bad. But the possibility to
support older hardware would be certainly nice.
After all, Linux never supported (well, there are some later ports as
separete projects) 286, and it was not that old when it was released.
It was just unsuitable for running it.

Thanks

Michal

--
             Support the freedom of music!
Maybe it's a weird genre  ..  but weird is *not* illegal.
Maybe next time they will send a special forces commando
to your picnic .. because they think you are weird.
 www.music-versus-guns.org  http://en.policejnistat.cz