RE: [xml-dev] More predictions to mull over

["Len Bullard" <cbullard@hiwaay.net>]

Life is dead or dieing, Elliotte.  And more is born to do that everyday.
Is HTML actually evolving or is it on life support because of the deployment
(ie, in a certain part of the lifecycle)?

OTOH, certain fundamentals remain the same and we keep tweaking them to
emphasize some part of a design to achieve some term of optimization for a
given period of constraints.  Note Mike and I both have used the combustion
engine examples.  I've cited the influence of the Grand Prix over that and
so on.

What you might want to dwell on is the fundamental aspect of REST over the
application zeitgeist of WS.  REST is fine and most of us understand how to
use it.  The next question is, is it enough if you need reliable predictable
contractable testable replicable server side services and do you want to
write very dense difficult noisy contracts to procure that or do you want to
cite a document that does that for you.  I suspect that you haven't had that
job yet.  When you finally enter the "Wow! He still writes HTML?  How
antedeluvian is that??" phase of your career and find yourself being
promoted up and out into the proposals department, you may change your mind
about contract-based services.

Or not.  Some people are making very good money because they are the only
ones in their company who know how to code in Cobol and Fortran.   Think of
them as 'coroner coders'.  It comes to all living things to carry their dead
around boxed or shrinkwrapped or left out on the back porch to let the
chickens nibble on.

It's not the velocity.  It's the mass.  You can handle a two-ton satellite
in orbit at 17,500 MPH with four-gloved hands, but if you make a single
mistake, it still crushes your fingers.  It may not be moving fast but given
the relative platform density, there is still a lot of WS-platform out
there.  Is that pretty or edgy?  Well, no.  Should it be?

VRML is dead except we are still using it because we have the tools and the
books. Raph Koster says 'no one serious uses it' but he writes games and
hasn't paid attention to the contracts coming out of the Beltway. X3D is
alive and the books are in the pipeline.  The tools are immature.   As I
said, convenience trumps elegance. VRML and X3D are citable.  Koster is
merely quotable.  

If the majority of desktops and profitable applications support WS because
of the deployment, ugly lives for as long as it is convenient.  Enterprise
apps may not be heroic but they pay better.   Adorable is great for
Hollywood and edgy is great for New York.  London gets by on ugly but
well-scripted.  

len


From: Elliotte Harold [mailto:elharo@metalab.unc.edu] 
 
Len Bullard wrote:
> We've had the same '. is dead' predictions for lots of languages and 
> technologies.   We've been having the same debates on the VRML/X3D lists 
> since those statements from the gamers that 'no one serious does 
> anything with.." some months ago.  Fortran is still out there, Cobol is 
> still out there and any time someone says 'yes, but who cares', check 
> out which language is running a lot of missile control systems and which 
> one is still running a lot of banks.  Densities change but not the fact 
> of for a sizable x there is some evidence of n.
> 

Dead or dying. CORBA is dead. DOS, Cobol and Fortran are dying. I teach 
at an engineering school, and I'm one of the few people there (including 
among the faculty) who knows anything about Fortran. I don't know anyone 
who still uses it, though probably there are still a few physics 
professors who are happily coding away in Fortran, but they'll all 
retire in the next few few years.

Languages like COBOL and Fortran are petrified fossils. So are operating 
systems like DOS, OS/2, and Mac OS 9. There is a huge amount of useful 
legacy software that keeps chugging along on these platforms, but nobody 
is dong anything new with them. At most they're carefully maintaining 
the old stuff. They're like decommissioned satellites left in a slowly 
decaying orbit for years or decades. Eventually however they will burn 
up in the atmosphere.

On the other hand tech like CORBA and OSI that never achieved orbit or 
even left the launch pad in the first place are thoroughly dead. They 
don't even have much of a legacy to support. Maybe VRML was written off 
prematurely and will still take off. I don't know. But there are a lot 
of dead technologies out there.

Java, C++, Windows, Mac OS X, Linux, XML are all alive and are the 
hearts of vibrant ecosystems of new code. It's not just the size of x 
that matters. It's the velocity.