[ag-automation] neuer Mitleser

Robert Schwebel r.schwebel at pengutronix.de
Mon Apr 10 11:14:17 CEST 2006


On Mon, Apr 10, 2006 at 10:35:27AM +0200, Wolfgang Denk wrote:
> > 2.4 is obsolete.
> 
> Repeating a wrong statement again and again does not make it true.

It is. The fact that you still support it commercially doesn't change
the fact that the community is focussed on 2.6 these days and doesn't
support 2.4 any more, and that for several years now.

Don't understand me wrong - it is surely a fact that there are people
out there who still work with 2.4, and due to the open source nature of
Linux they can, and they still find commercial support (for example from
your company). But people have to note that "the" Linux development,
stable and unstable, is light years ahead of the old 2.4 days.

Statements like yours result in people asking for support for modern
interfaces like for example wifi, usb or spi ontop of three year old
technology. Things which have long been solved in 2.6. Telling them that
you just have to backport three years of USB development just doesn't
solve this. And if I see the code chaos in the current ppc/powerpc part
of the 2.6 tree I'm really wondering if people shouldn't better focus
their creative energy into cleaning that up and make the mainline better
instead of wasting time with prehistoric stuff. 

No offense intended, this is just a different view, and OSS is about the
possibility to have different views - in the end a customer decides
which variant does better fit to his needs. Our trees and BSPs do all
sit ontop of the latest Linus GIT tree and all development is done
there. 

> Yes, 2.6 is the current development kernel, and new  work  should  be
> done  in  this  environmwent  if  possible.  But 2.6 has some serious
> problems: it's big and  slow  (especially  on  low-end  systems)  and
> anything  but  stable. It's a moving target. 2.4 is the stable kernel
> tree at the moment.

Well, it might be due to the case that we don't have anything slower
than 75 MHz (H720x). I personally don't think that long term development
strategies should be focussed on corner cases; we once ported Linux to
the DigiConnect ME, which is a 40 MHz network plug with Linux inside,
but that's definitely not the platform we build our strategic processes
around (and it worked just fine with 2.6).

Last time the statements from your website have been discussed on LKML,
http://www.uwsg.iu.edu/hypermail/linux/kernel/0603.1/index.html#1722
there wasn't much activity from your side. Woudldn't it be a good idea
to discuss performance problems there?

Side note to occasional readers: Please don't take this discussions too
serious. Open Source people usually have strong opinions. That's what
makes Linux better. It doesn't mean that we all don't love each other ;)

Robert
-- 
 Dipl.-Ing. Robert Schwebel | http://www.pengutronix.de
 Pengutronix - Linux Solutions for Science and Industry
   Handelsregister:  Amtsgericht Hildesheim, HRA 2686
     Hannoversche Str. 2, 31134 Hildesheim, Germany
   Phone: +49-5121-206917-0 |  Fax: +49-5121-206917-9



More information about the ag-automation mailing list