[
Lists Home |
Date Index |
Thread Index
]
On Apr 4, 2004, at 3:37 PM, Elliotte Rusty Harold wrote:
> At 6:48 AM -0700 4/4/04, Rex Brooks wrote:
>
>> Lastly, who wants to cast the vote the prevents allowing the only
>> viable, public standard completely uncoupled from partisan politics,
>> that stands a chance of saving a single life in a time when another
>> incident such as 9/11 or 3/11 could happen at any time, especially if
>> that public standard, because it IS public, CAN be amended to do its
>> job better?
>
>
> The only reasonable way to judge a standard like this is to tally up
> the advantages *AND DISADVANTAGES* and see if the advantages outweigh
> the disadvantages. Sometimes it's a judgement call which different,
> rational people may come to different conclusions about.
I agree, except that it's almost ALWAYS a judgment call. XML 1.1 is a
very good example -- sensible people can disagree on the costs and
benefits, especially since the costs are now and the benefits are
sometime in the future, maybe. Furthermore, some of the costs and
benefits are intangible; I, for example, generally think XMl 1.1 is a
good idea simply because it has a cleaner alignment with Unicode and
fixes an outright bug in XML 1.0 (the infamous NEL character, which is
clearly a line terminator in Unicode). XML can't be "Unicode with angle
brackets" but then pick which bits of Unicode it finds convenient to
actually align with without losing a bit of conceptual integrity.
(This is not to restart that permathread, just to illustrate problem of
comparing the tangible costs of incompatibility vs the intangible
benefits of improved conceptual integrity.)
As far as I can tell, it's the realization that widespread consensus on
the cost-benefit tradeoff for a given proposal is more or less
impossible under the W3C rules that drove a small group of powerful
companies to invent the WS-Emperor spec development process. They got
the "important" players in a room (or email list) and hashed out the
basic architecture. I presume, but have no evidence, that lots of
horses were traded to come up with something that was mutually
acceptable. THEN, they bring in a widening circle of companies to
review the specs (I have the image of a troop of monkeys picking bugs
out of the gorillas' fur), and FINALLY send it to OASIS. This is
actually an interesting experiment -- they know that doing the whole
thing in an open way leads to 5-year-and-counting efforts such as
XQuery/XSLT 2, but they know that having IBM, BEA, and MS all go off
and invent similar but incompatible specs will kill the network effect
that keeps all the balls (bubbles?) in the air. They're trying to get
the advantages of some openness without the paralysis that all those
conflicting judgments based on very different experiences and values
tends to produce.
Watching the media events where the emperor is being draped in
imaginary standards and having the glitches be called the equivalent of
wardrobe malfunctions rather than public nudity does get a little
tiresome, I agree. I believe Don Box said it clearly the other day:
"For those of us who work on the WS-* protocols at MSFT, we look at
this exercise as defining the public "API" to our next generation of
technologies."
http://www.gotdotnet.com/team/dbox/default.aspx?key=2004-04-03T08:01:
14Z There's a word for this I just tracked down -- mokita
http://www.randomhouse.com/wotd/index.pperl?date=19991220 "the truth
that everyone knows but no one talks about".
Anyway, there seems to be a lot of mokita surrounding this topic. I
have never seen or heard of a viable, "public standard completely
uncoupled from partisan politics"; the mokita is that politics is
pervasive in the 'standards' world. Likewise, it's time to face up to
the fact that most of the WS-* specs are not 'standards'; the mokita is
that it is they are treaties between competitors that may or may not
actually work, and may or may not be a viable basis for standards in
the long run. I advocate a open mind, but healthy skepticism about all
of them; we need to consider the short run costs and benefits of
adopting any one spec, and we need to consider the necessity of
evolving the whole mess into something coherent, or allowing it to die
if the experiment turns out to be a failure.
|
