*BSD News Article 75122


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!munnari.OZ.AU!news.ecn.uoknor.edu!news.eng.convex.com!newshost.convex.com!newsgate.duke.edu!news-feed-1.peachnet.edu!paperboy.wellfleet.com!news3.near.net!news-server.moat.platsol.com!news
From: Harold Lockhart <hal@platsol.com>
Newsgroups: comp.unix.bsd.bsdi.misc,comp.soft-sys.dce
Subject: Re: DCE secure core and DFS for BSDi?
Date: Wed, 31 Jul 1996 11:52:25 -0700
Organization: Platinum Solutions Inc.
Lines: 56
Message-ID: <31FFAB69.738E@platsol.com>
References: <4rumio$qhc@news.more.net> <glvITFG00iWZI8aH5x@andrew.cmu.edu> <DuqpA8.29n@wiz.com> <w5w687dysyo.fsf@mantis.austin.ibm.com> <4tctvg$fg1@paperboy.osf.org>
Reply-To: hal@platsol.com
NNTP-Posting-Host: hal_pc.bos.platsol.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: Mozilla 3.0b4 (Win16; I)
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.bsdi.misc:4433 comp.soft-sys.dce:3584

Rich Salz wrote:
> 
> In <w5w687dysyo.fsf@mantis.austin.ibm.com> Chris Cowan <cc@mantis.austin.ibm.com> writes:
> >There is nothing that would prevent someone from writing a
> >"work-alike" based on the OSF RFCs, is there?
> 
> Primarily the AES volumes, not the RFC's.  Nope.  The only thing getting
> in the way is that it is almost impossible to write the specs correctly
> enough, and with enough detail, so that it is practical to do this. But
> it's always been OSf's stated policy that doing this is a goal.
> 
> >I have always harbored the hope that someone would do this.
> 
> Me too.
>         /r$

Just for the record, it is my understanding that Tandem re-engineered CDS (or perhaps 
just NSI) using their own database engine.

But considering re-engineering the whole thing, unless someone did it and then made the 
code available for free (as was done initially with UNIX and TCP/IP) what would be the 
advantage?  Everyone complains about how long it takes for DCE to come out.  A 
separately implemented version would trail by a year or more.  (This is being 
optimistic, since the AES has not been updated for 1.1 and 1.2 has now been shipped to 
vendors.)

I understand the politics of the "single spec, multiple implementations" ideal, but does 
it really make sense for infrastructure?  Building robust distributed systems is so 
hard, that I for one would sacrifice almost anything for being able to count on rigorous 
interoperability.  When I am looking for bugs, I don't want to have to consider the idl 
compiler, etc. (Yes, I know the idl compiler has bugs, this follows from Lockhart's 
assertion, but they are the same everywhere.)

In fact, my main complaint with the "one stop shopping" approach is that paradoxically, 
a single dominant vendor (IBM in the old days, Microsoft now), for a variety of reasons, 
does not work hard enough at interoperability even among their own products. I assert 
that there is better interoperablity today among DCE versions from five different 
vendors than between, say five Microsoft applications or five 1980 vintage SNA products.

Less anyone misunderstand, I fully endorse the current policy of allowing any product 
that passes the conformance tests to be labeled "DCE" regardless of whether a license is 
purchased from OSF. I joined many others in urging the previous policy be changed back 
in '92.  However, I see see bundled runtimes everywhere as a far more important priority 
than multiple implementations. 

Now if the Open Group could figure out an economically justifiable way to put the whole 
thing in the public domain...

Hal

=================================================================
Harold W. Lockhart Jr.            Platinum Solutions Inc.
Chief Technical Architect         8 New England Executive Park
Email: hal@platsol.com            Burlington, MA 01803 USA
Voice: (617)229-4980 X1202        Fax: (617)229-2969
=================================================================