Re: [MiNT] New MiNT distribution is n.. ..pro

[maxilys@normandnet.fr (Remi Villatel)]

Hi Martin-Éric_Racine, on 20 Feb 00 you said:

 > (En passant, Rémi, tes dates sont en année 00, plutot que 2000)

(Elles sont correctes dans l'entête du message, c'est tout ce qui importe. Le 
reste vient des limitations de LED...)

 >> We can't stay stuck to the 680x0 forever.

 > Please explain what you find so limiting on m68k processors.

Jo-Even already summarize all my thoughts: Speed! I don't think that the m68k 
are limited but they need a little more speed. I'm the proud owner of an Hades 
060, I wouldn't exchange it for a hundred of any other computer... 
Nevertheless, I found it too slow sometimes, especially when I do what MiNT is 
supposed to allow us to do: multitasking, i.e. use several programs at the same 
time. When I use GEM programs that do nothing but wait for me to use the mouse 
or the keyboard, the speed isn't an issue but if any of the programs starts to 
manipulate files, the speed drops and I can't even type at my full speed in 
Qed. Maybe this is only due to the lack of real DMA on Hades. Anyway, I'd like 
more speed in Calamus, in CAB, in PoV, in Photoline, etc. And 20 more MHz ain't 
just enough...

Beside this, Motorola doesn't support the m68k range any more. They make chips 
for those who want some but they don't develop any more. We'll never see even a 
100 MHz 68060. They concentrate on the ColdFire and the PPC.

 > If you really expect people to change architecture (even if a 68k
 > emulator is included), you must give them some real-life reasons
 > to do so, not a circular argument about Coldfire versus G4 like
 > Rodolphe did on fr.comp.sys.atari.

What's "circular" in what I said? I proposed an alternative with two choises: 
The ColdFire (which I find "crippled" because not so compatible with the m68k, 
neither very fast at the moment) and the PPC (which is fast and well supported 
by Motorola). I don't care about Rodolphe, the PPC was the best choice at the 
moment I wrote. If the ColdFire becomes faster and more compatible, it will 
then become the best choice. That's _even_ what I'd prefer!

    [---SNIP---]
 > A technological argument means absolutely zippo, when the whole
 > point is to make something for real-life users who want to surf,
    [---SNIP---]

Let's talk about surfing... Did you never dream that CAB calculates the pages 
and decompress the JPEGs twice faster? Optimisation reaches its limits very 
fast when dealing with complex things. And if we want to have Javascript too, 
optimisation won't do miracles, the program will have to become bigger and to 
do some more things on a machine which doesn't accelerate. The final result 
will be definitively slow without a real increase in speed.

And the 20 MHz of the Milan don't impress me. The benches they proudly display 
on the web site are only a few % higher than the benches of my Hades at 60 MHz, 
and even not all results. So...

See you later,

=============================
Remi Villatel
E-mail: maxilys@normandnet.fr
=============================