`open systems': no one can agree...

H. L. Rogers rogers at ofc.Columbia.NCR.COM
Thu May 26 23:22:43 AEST 1988


In article <11671 at mimsy.UUCP> chris at mimsy.UUCP (Chris Torek) writes:
>I get the feeling that every single person who has posted an opinion
>about OSF has a different definition of an `open system'.  So:  Just
>what *is* an `open system'?  What makes a system `open'?
>

Wow!  A chance to define *open*!  Let's start with my American Heritage
Dictionary of the English Language:

	open. 1.a. Affording unobstructed entrance and exit; 
              not shut or closed. b. Affording unobstructed 
	      passage or view; spacious and unenclosed.  
	      2. Having no protecting or concealing cover; 
	      exposed.  3. Not sealed, tied, or folded: an 
	      open package.  4. Having interspersed gaps, 
	      spaces, or intervals: open columns.  
	      5.a. Accessible to all; unrestricted: open meeting.  
	      b. Unhampered by restrictions.  6.a. Susceptible; 
	      inviting. b. Unprotected; vulnerable: open to 
	      attack.  7.a. Available; obtainable: The job is 
	      still open. ......etc, etc, etc.

Unobstructed entrance?  Accessible to all?  Unhampered by restrictions?
Available?  Obtainable?

Seems to me that the way to meet all this criteria in the context of
an *open* operating system is to fund an independent organization
which will, without bias, develop and provide the software to any
and all who are interested.  This organization would receive its
input from *end users* relative to the features/functions which are
desired by the buying community, and from ISV's (stress the 
*Independent* in ISV) for the features/functions needed by 
applications.  It would not be swayed by the prejudices of hardware 
vendors.  

So, you might ask, where is the incentive to fund this *super-open*
organization if you cannot influence what they develop?  One answer
is provision of favorable royalty fees to the funding companies, or
the right to purchase a paid-up-forever-fee.  

On what hardware would this super-open OS execute?  Well, as 
delivered from the developing organization, probably none, if you
really want it to remain *open*.  Each and every hardware machine
dependency could be documented, but not supported by real code.
Then, each and every hardware vendor could develop his own hardware-
dependent code.  And if he's smart, he will not forget about the
ABI for his chip(s) as he develops his hardware-dependent code.

Using the American Heritage definition and my interpretation,
there is *no* movement to provide an open system.

Sigh.....I guess I was just being a romantic today.



More information about the Comp.unix.wizards mailing list