Joshua Thorp wrote:
On Sep 14, 2006, at 3:17 AM, Marcus G. Daniels wrote:
While possible, the idea that
university or hobby software can be better
than software developed by a
multi-billion dollar corporations doesn't
jump out as a likely scenario.
Interoperability is God, and failing to
provide it is a fine reason for
a software project to fail!
If you are a multi-billion dollar company why interoperate?
Just declare the rest of the market for suckers and dilettantes.
Unfortunately for billion dollar companies its turtles all the way down
and they struggle mightily just to interoperate with their own
products--and largely fail.
Shrinkwrapped software is a tiny minority these days of the total
software pool.
Back end business systems are much larger, and one of their prime tasks
is B2B and B2C operations
to automate sales and distribution, supply chains, information access,
customer tailoring, etc.
Interoperability is a very important component of that task. RPC,
CORBA, XML, Web Services/SOAP, Indigo
are all approaches that are supposed to make this interoperability
easier, but we seem
to always be stuck in unhappy land.
Here's an interesting new attempt: http://www.zeroc.com
Seems to me the very concept of a multi-billion dollar company
as software producer put up against small groups of hackers is absurd
anyway. It really hinges on the make up of the small teams of people
inside that large capital structure that are doing the real work
anyway. Plenty of fine coders exist inside and out of such large
companies and depending on management and marketing or acquisitions
they may have more or less time to deliver a finished product. But
more often than not, though a billion dollar company may be good at
well crafted design process, I would bet they find their best ideas
from those who do something for the sake of art as an amateur, or to
push forward the frontier of ideas as a scholar.
Not to get into a flame pit, but isn't that a bit elitist?
Some people are very creative when working on a directed,
corporate-goal style problem.
Many of those goals are very pragmatic, customer- and market-driven,
and so have a pretty
high payoff in satisfaction if solved elegantly - the solution will be
used extensively.
I would also point out one of my long-term examples of how DEC or HP
came up with
a good robust extended file permissions system on Unix, while the open
source world took
years more effort and AFAIK never came up with anything usable for
serious applications.
Yes, it partly comes down to the teams of programmers, plus their
support, their motivation
and incentive, the maintainability of the project, including the
oft-maligned marketing aspect
(even in the open source world, if a good project isn't marketed well,
it won't attract new
programmers, bug reports, etc.)
'Can't we all just get along?'
What we are talking around here is just as personal as race and
politics--where do you fall on the artist<--scientist-->engineer
spectrum. Engineers are most comfortable in slow moving vehicles with
plenty of restraints and air bags. Artists are most happy in new
concept cars that are untried and untested--they might die but at least
it will be a statement of some sort. And of course testing cars is for
scientists. (Some may quibble with this, but I would have to say
check out the difference between math departments in an Engineering
school versus an arts and sciences school -- Engineers are most
comfortable working with equations from a table and processes from a
lab manual, scientists get a big kick out of deriving equations that
are already in that engineering text. And I think artists are largely
there for the drugs and the women..)
I think George Michaels and Gary Glitter are there for the boys ;-)
I couldn't imagine the same languages appealing to all three
crowds. And why should they?
We can certainly tell a lot by the tools a person uses (and the
company they keep). And if you don't like engineers I would say better
to avoid C++, project managers, and multi-billion dollar companies.
Another angle to this whole mess is that it is possible to write
very unstable and largely un-useful code in C++, it just takes a long
time to get there. If you want, you can get there faster in python.
--joshua
============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at http://www.friam.org
|
