OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: [xml-dev] Don't Let Architecture Astronauts Scare You

[ Lists Home | Date Index | Thread Index ]

Blaming the US for global economic trends is like blaming a 
bartender for drunks.

Let's get back to real cases:

Quality:  MAC plug and play generally works out 
of the box.  How do they do that?  Closed architecture.
One doesn't design the MAC on the web.  It's a  
consensus-phobic system.  Apple has made that into 
a marketing feature.

Reliability:  don't use a feature that isn't proven. 
Can the software market produce innovations that way?   
So far, no and if it is left to the web to simplify, 
it becomes a long wait these days because standardization 
and specification are conflated under consensus.  The web 
has become design by a very very very large committee.  It 
is dissent-phobic.  Seen any innovative W3C apps lately?  

The web browser was not designed in Redmond.  
Who designed it?  Grad students feeding at the DARPA 
trough.  Was it a modern innovation for hypertext? 
No. It was a giant leap backwards.  It was cheap too.

Why do it then?  Scalability across the dimension of 
human competence.  If the web was to be something 
any idiot could design for and use, it had to be simple. 
But that means we can only go as fast as the 
slowest idiot.  It's innovation-phobic.  Anything 
really new is also unrecognizable in and of itself, 
so unless it produces remarkable results quickly and 
for a lot of observers, it lays there.  See VRML (not 
that new, but remarkably powerful; just not that useful).

The modern GUI originates not at PARC, but in the labs 
of Douglas Englebart and Ivan Sutherland.  Was it an 
innovation?  Yes. Was it profitable to build?  Not when 
it was invented.  Profit required piracy and cheap cheap 
cheap processors and memory.  One can't honor patents 
and use expensive platforms.  It's low-resources-phobic. 
Remember, cheap scales.   High quality and low cost 
don't usually come in the same package unless one 
can sell quantity or give away product and not starve. 
See the deLorean.

Linkbases are an old idea.  Cool or not, MS didn't 
do us out of them on the web.  Tim Berners-Lee did.     
One can't shift the blame to Redmond.  That is 
history-phobic.  The beasties are an old idea. They 
died under the banner of YAGNI at CERN.

Would linkbases be 'cool'?

We are revisiting old ideas with new names.  Linkbases were 
replaced by relational database systems. The question 
that has to be answered is are we getting complexity 
with no gain in power, flexibility, etc.?  IOW, before 
XHTML gets harder, ask why?  What will it be used for? Are 
the easy cases compelling enough to live with the hard 
corner cases? Are there easy ways to do the same things 
without them?  Are there harder ways that don't cost 
as much in terms of deployment and learning curve?  

Architecture astronauts don't scare me.  I live in 
the town that put engines under jet jocks and sent 
them to space.  I know it can be done; it isn't 
done by consensus.   Extraordinary innovations are 
usually the results of small teams that work together 
on many projects for many years developing their 
techniques, understandings, and team communication 
strategies.  Then a leader with an agenda gives them 
a fat budget with a firm deadline.  It doesn't 
scare me because it is thrilling to witness.  But 
don't think it can be done on the cheap, and don't 
count on sustaining it because one purpose systems 
have a way of being optimized past any reuse once 
the original agenda is met.  They produce results 
but they are hard to keep extending for new uses. 


From: Matt Gushee [mailto:mgushee@havenrock.com]

On Tue, Sep 17, 2002 at 11:11:41AM -0500, Bullard, Claude L (Len) wrote:
> They make Mercedes SUVs in Alabama (no joke).  
> Would a modern day software vendor stay in 
> business if it had to use the Mercedes 
> design approach?  Remember, the Mercedes 
> SUV is made in America, so the "Americans 
> build junk" phil won't wash.

It's a gross generalization with an element of truth, depending on what
country you are comparing the US with. But I didn't and don't say that
"Americans make junk." Every country has some junk, and Americans make
some high-quality products*.

But I will make, and defend, the statement that the American economic
system and the emerging, US-dominated global system, make it difficult
to invest in good design and quality manufacturing. I think the reasons
are well-known: primarily the pressure from financial markets for
short-term profits--which has no doubt always been there but has been
exacerbated by deregulation over the last two decades.

And the popularity of that moronic mantra "better, faster, cheaper"
hasn't helped. "Better" usually loses because it is hard to measure.

> It is a matter 
> of what will a customer pay for something 
> considered disposable.

Which partly depends on cultural attitudes towards things being

> Software is eminently 
> disposable.  Data isn't.  That is exactly why 
> markup came to be.

Interesting point.

 * However, I believe it is fair to say that some countries (Japan, 
   Germany, and Scandinavia come to mind) have a higher proportion of
   quality products and services than the US. The fact that it's been
   said before, and the fact that the quality of some American products
   (notably cars) has improved in recent years, don't make it untrue.
   One concrete example: what happens when you stick a bill in a US
   vending machine or change machine? It works, what, maybe 60% of the
   time? Whereas Japanese machines very rarely reject bills--in seven
   years there, it happened to me about three or four times.

   I wish I had time to continue the dissertation. But let me close
   with the thought that junk and bad practices exist everywhere, but
   they are unevenly distributed, and it is instructive to examine why 
   that is so.


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS