[cgl_discussion] Re: [Ksummit-2004-discuss] CGL

Mika Kukkonen mika at osdl.org
Mon Jan 26 13:09:28 PST 2004


(dropping ksummit mailing list from CC to avoid offending them
with offtopic discussion; feel free to put them back if you want)

On Mon, 2004-01-26 at 12:40, Christoph Hellwig wrote:
(...)
> I think most people who commented on CGL on here have their fair deal
> of having to deal with it.  Maybe it be by working for a distributor
> that was talked into implementing the spec or having to deal with the
> completely brainded CGL patches as a subsystem or driver maintainer.

You could replace string "CGL" in above paragraph with "POSIX" with no
change. I agree "design by committee" has it's problems (the old joke
about "camel is horse designed by committee" comes to mind). So nothing
new or CGL specific there.

> Sorry for disrupting your nice marketing speech, but their intent was of
> course not to use linux in the telco industries.  Their intent was to
> get the same thing they always did, just cheaper.  And Linux + commodity
> hardware seemed to be the best way to get that (and probably it _is_)

While price of an OS is a factor, it really is not the major one. The
real pain comes from two factors:
	a) change from circuit-switched technology to packet switched
	    (i.e. from SS7 to IP-based protocols), which radically
	    challenges all current telecom product platforms
	b) The pain of dealing with multiple different OS'es (some   
	    in-house, some commercial), with all having too little
	    weight to get enough ISV support
Linux is attractive because it has leading edge IP-stack and starts
to have significant ISV support.

> Because a bunch of id^H^Hphbs coming up with a spec is never helpful.

I have always thought that information is better than no information.
Would you have preferred that we would have made the CGL spec only
available to OSDL member companies instead of making it publicly
available?

> > existence. Of course it can always be argued that projects like OpenHPI
> > (http://openhpi.sourceforge.net/) would exist even without CGL, but my
> 
> And it wouldn't be a big deal if it didn't existed.  One project more or

To you, probably not. To several companies, it is.

> less implementing a stupid spec more or less doesn't change the world.

Lot of people could have said that about an OS project started by a
student in Helsinki University 15 years ago ...

--MiKu





More information about the cgl_discussion mailing list