[linux-elitists] PengPod engineers for a new project

Greg KH greg at kroah.com
Thu Aug 8 15:46:17 PDT 2013


On Thu, Aug 08, 2013 at 05:30:40PM -0500, Tilghman Lesher wrote:
> On Thu, Aug 8, 2013 at 4:33 PM, Greg KH <greg at kroah.com> wrote:
> > On Thu, Aug 08, 2013 at 09:36:00AM -0500, Tilghman Lesher wrote:
> >> On Thu, Aug 8, 2013 at 4:27 AM, David Edmondson <dme at dme.org> wrote:
> >> > Linux will win completely and be everywhere, but we will have none of
> >> > the freedoms that it was intended to allow.
> >>
> >> That Linux will "win completely" is actually a bad thing, historically
> >> speaking.  In any market where a single product has "won", stagnation
> >> was the result.  It's the fight for dominance which keeps us moving
> >> forward, and thus, I hope for a future where there are always at least
> >> 3 options, continually fighting for dominance, sometimes one pulling
> >> forward, but with no one product achieving the dominance for which it
> >> fights.  It's a process, not a destination.
> >
> > You used the word "historically" correctly here.
> >
> > Remember, Linux has done something that has never been done before, so
> > why would history be any indication that this is going to cause
> > "stagnation"?
> 
> Remember the dot-com boom and bust, where everybody insisted that they
> had a new way of doing business, and the old rules didn't apply?
> You're telling me the old rules don't apply to Linux.

I'm not talking about "business rules", I'm talking about software
engineering rules, and the development of ecosystems and huge systems.
Linux has broken all of those "rules" in the past, so much so that no
one teaches them anymore (well, good schools don't, plenty of old ones
still do...)

> > Remember, there isn't just "one" Linux, unlike there has been in the
> > past with single-systems.  There are thousands of Linux versions out
> > there, for hundreds of thousands of different types of systems, being
> > used in a myriad of different ways.
> 
> There are thousands of Windows versions out there, too

No there isn't.  There are only a handful, all running on one processor.

Don't count the windows embedded stuff, that's a totally different
operating system.

If you have to write new userspace code for the operating system
version, that's a new operating system.

> , just as there are thousands of MacOS (probably hundreds, if you only
> count Mac OS X) versions,

There are not thousands of OSX versions either.

> but we generally count each of those as
> being singleton
> competitors.  Yes, the versatility of Linux means that if you live on
> the bleeding edge, there are thousands of different kernels in use at
> any one time, but most of us who aren't kernel developers tend not to
> run anything other than what is packaged.

The Linux system you use in your phone is very different from the
version you interact with when you use Google and it's different from
the version you use on your desktop and it's different from the version
you use when riding in an airplane (both from the seat-back
entertainment system and the air-traffic control system that guided you
in safely) and it's different from the version used by the airline to
bill your credit card, and so on...

Yes, they all share a common kernel at one point in history, but all are
different at the same time.  They all require different things, making
the overall ecosystem quite varied.

> We have more important things to do than rice up our kernels like some
> sort of Gentoo user.

I don't think you really understand who uses Gentoo these days, do
you?  NASDAQ and ChromeOS are just two such usages, both succeeding
quite well, don't you think?

> > It is being created by 3,000 different developers, from over 450
> > different companies every year, pulling it in ways that are anything but
> > common.  There is a constant churn and change happening, never stopping.
> >
> > And this rate of change keeps increasing, stagnation is no where to be
> > seen at all.
> 
> Correct, because Linux hasn't won, yet.  It has not achieved an
> operating system monopoly, yet.  Competition thrives, so innovation
> thrives.

So you are saying that the only reason Linux is being developed so much
is because it is competing with something else?  What exactly is it
competing with that is causing this change? [1]

> > To quote an IBM executive, "Are you done with Linux yet?"  And my
> > response, finally, after 10 years of hearing this, "We will be when you
> > stop making new hardware."
> 
> Similarly, I hope this never happens, because hardware innovation
> would be dead, and we'd all be poorer for it.

I agree, and I don't think it ever will happen.  Hence, Linux needs to
keep changing to keep up with new hardware.

> > As long as the world keeps changing, Linux should be changing to keep up
> > with it.  If it doesn't, then it will die, but that's a different story.
> 
> I don't think that you're even advocating a position where Linux has
> to win the operating system market, beating out every competitor and
> obtaining what is essentially a monopoly.  If we were ever to get
> there, however, I'm not certain if you see that as wise or if you're
> simply indifferent to the possibility.  Care to take a more definitive
> position?

I think it is wise in that Linux really is better than anything else
right now.  But also sad, I like those other operating systems, and it's
sad to see them die, but I fail to see anything else succeeding these
days, in anything other than "niche" markets.

greg k-h

[1] Linux is really competing with Linux these days, not other operating
    system kernels, we are so far ahead of everyone else its scary, and
    troubling in that we are now making it up as we go along, and that's
    way harder than trying to catch something else.


More information about the linux-elitists mailing list