Lists Home |
Date Index |
From: David Megginson [mailto:firstname.lastname@example.org]
Bullard, Claude L (Len) wrote (to Tim Bray):
> Whether I'm raving or not, your response to that late night hacking
> was to help create an environment in which nothing is predictable
> and anyone can stamp anything anyway they like and the
> press will pick it up and run with it. You like the credit for being the
> "co-inventor of XML" but don't accept any role in the damage done by the
> gutting of ISO and the norms of standardization that stood
> in your way.
>Len, that's a bit hard to swallow.
Try real hard, David.
>ISO is an organization designed to handle negotiations among countries.
>That's not a bad idea for things that tend to depend on heavy national
>regulation, like, say, smokestack industry or retail, but it makes little
>sense for computer technology and networking standards
That is coming soon. A regulated Internet, taxed and the whole bit, but
that isn't the issue. The issue is conformance testing such that one
can reliably cite a standard without having to be there when it is
written by those 'real smart guys' who document badly and don't understand
procurement at all. Those guys who don't believe in patents and so
now JPEGs are getting people sued left and right. Those guys who
believe 80% is good enough, so some years later, big security holes
open up and mission critical systems are compromised.
>Using ISO's national-body structure for negotiating computer
>standards is about as effective as the two of us negotiating a mideast
>plan, then expecting Yasser Arafat and Ariel Sharon to thank us and
Traceability of a process, transparency of a process, rules for
these things mean nothing to five guys hacking a design, but they mean
a lot to the people who buy them. This is what happens at scale.
>The fact is that ISO never did well with computer technology, either before
>*or* after XML.
The fact is that where ISO works with a technical consortium, they do quite
well. They can do better.
>We use the four-layer DoD networking stack, not the
>seven-layer ISO/OSI stack, and we look to the IETF, not ISO, for our
>protocols. Even modest computer tech successes like SGML have been rare
And as a result, a lot of glue is sitting on top of TCP/IP, DDNOS has
no credible defense, security is something of a shambles, and so on.
It isn't the organization; it's the people, I agree, but the people
aren't there to explain or indemnify.
Not megalomania: inexperience. A computer science degree is the
least part of that if not a trivial part. Did you really expect
the industry to walk away from patents or just give up all their
rights to intellectual property as the open sourcers worked to
drive the value of their products to zero? That is blisteringly