🚀 go-pugleaf

Hey there mjr...  couldn't resist:          :)


mjr@hussar.dco.dec.com writes:
>
>	I don't think it's *POSSIBLE* to define "Open Systems" meaningfully.
>I've heard some folks define "Open Systems" in terms of applications
>portability - but - do you see the hook? It's not the vendor that
>provides applications portability!! (As Geoff should know, if anyone does!)
>

You're probably right --- at this point there's probably no way to define
the term meaningfully.  Just like "AI", "OOP", and "MIP", "Open Systems"
is a relative term that's going to be used in the most market-advantageous
way possible by the person using it.

Let me throw in my $0.02, tho;  a reasonable definition of an "open system"
might be a "system" (whatever that is) that is based on interfaces 
(hardware, system call, whatever) that are well known, well understood, and
available in a variety of implementations from a variety of vendors.

By this definition, it's easy to see that "UNIX" is an open system, defined
by the familiar UNIX API and the more-or-less standardized environment that
comes with it.  VMS loses on the basis of its single-vendorness.  NT will
also lose on the basis of its single-vendorness, plus its new and likely
quirky system-level API.

PC *hardware* is open.  DOS and Windows are not.  SQL per se is open,
whereas older versions of, say, Progress were not.  Even DBase is open,
more or less.  SCSI is open.  ADB isn't.   Etc, etc.

What about NeXTStep?  Well, certainly, the UNIXy side of it is open.  The
rest of the stuff that makes up NeXTStep is closed as all get out --- for
now.  NeXT appears committed to "opening up", tho:  witness the GNU ObjC
project and NeXT's support for it.


>	Other people have told me that "Open Systems" means that the
>system conforms to all relevant industry and de facto standards. That's
>also nice, but that still means that you get to ensure the portability
>yourself, and that you're still at the mercy of the quality of the
>implementation of the standard. Is it possible that "Open Systems" is
>much like saying, "Here's a bunch of industry standard sheep fur. If
>you want a coat, go for it!"?

De facto standards don't make open systems.  At one point, IBM's
various incarnations of its mainframe OS's were *the* defacto
standard --- but they sure as hell weren't "open".

>	Is it possible that when someone says they want "Open Systems"
>they're saying they want "cheap, fast, and good?" - most likely. One
>thing for sure - in the open systems arena, there ain't no such thing
>as a free lunch. Or is that an "Open Lunch"?

When I say I want "open systems", I mean I don't want to have to learn
another idiosyncratic environment, a new environment-specific lingo,
and a new set of tools to do whatever it is I want to do.  Every time
somebody re-invents the wheel and everybody has to ramp up on the new
environment, overall progress slips.  On the other hand, small differences
between "standard" environments are going to happen, and that's no big
deal (esp. if, as in the case of the NeXTStep UI, they add value).

I want to be free to choose between hardware and software platforms with
basically similar features, but differentiated on the basis of value.
I want to be free to move ahead as technology progresses and not be bound
environmentally to a single-vendor.  I want to have a foundation of concepts
and tools that is common to all the platforms I have to use.

Probably, more than anything, "open systems" is the opposite of "single
vendor".  Power to the people. ;)


jb



--
-- Jeff Bone ------------------------------------------------------------
        NT...  Just say N0T.   (Viva UN*X!)   Standard disclaimer. 
------------------------------------------------------- jbone@dell.com --
                                                                       --