[e2e] the evolution of deployability
cannara at attglobal.net
Tue Dec 17 18:53:50 PST 2002
I must agree with Noel on many points. Those of us who've been on even one
stds committee can surely recall the process -- one that often diverges from
optimality, even engineering sensibility, as the task moves along.
Small groups of individuals often do so much better. Go back to modem days
and recall what Hayes did. Proprietary stds fostered competition, as well as
engineering selection for excellence. The ultimate supremacy of Hayes stds
helped everyone and provided a firm base for good, non-proprietary stds.
Ethernet was an even simpler example, requiring less sophisticated engineering
than modems. Too bad the IETF, over the decades, seems incapable of this,
especially when confronted with the harsh realities of M & C marketing &
Yes, NAT is evil, but it's less evil than IPv6, because of how v6 was
corrupted from the ideas around needing a v4 successor. The history and
documents that describe the v6 process are as intriguing as the result is
disappointing. IPSEC is another example. Looking at other stds bodies, one
can find more as well, need we mention ATM?
We've unfortunately entered the millenium with 3 mediocrities as dominant: M
as the leading software supplier; TCP/IP (and associated stds) as dominant
protocols; and I as dominant in chip making & processor architecture.
It's always interesting to check into the trade rags, trivial research
projects and technical discussions once every few years, just to remind us of
how slowly we progress, and how scattered that progress is in relation to
computer networking's promise that we actually lived 30 years ago. Such
Roop Mukherjee wrote:
> Indeed it seems that standards allow mass production of compatible things.
> But, the question under debate here is why does evolution of such things
> seem to slow down.
> I am merely offering a point of view that standards themselves or their
> governing bodies may be responsible at least for a part of it.
> If indeed the standards bodies allowed technologies to play themselves out
> and restricted their role to ensuring interoperability of the important ones
> that emerge then perhaps we may have systems that evolve better and
> faster. In other words, my objection is toward standardization of design
> (how to build a fax machine) as opposed to interfaces (how to display
> faxed characters).
> -- Roop
More information about the end2end-interest