DB design document

[Jean-David Beyer]

David Merrill wrote (in part):

> Don't you know that is one of the big problems with enterprise
> software design? It's all designed by committee!
> 
I sure do. I worked at a large, formerly well regarded, research and
development organization, and we had to deal with that all the time
(with little success IMAO). No doubt the people at Microsoft have the
same problems, since it is my understanding that the individuals there
are extremely intelligent, yet the results ... .

I had the good fortune to work for a small system engineering
organization for about four years. Management was enlightened. No design
by committee. There was a principal architect for system design. While
he would listen to reason, he was the dictator and his decisions were
administratively superior to those of the CEO as far as the system
architecture was concerned. (Fortunately he was extremely competent in
this role. I suspect that even if he were a little less competent it
would still be the way to go. Napoleon remarked that one bad general is
better than two good generals.) This ensured a consistent point of view
for the entire system. Once the modules passed unit test, the
integration testing of the entire system took only a few minutes,
because the only problems would be ambiguities in the interface
specifications, which were extremely few, and they were quickly
resolved.

I wish all projects were as easily done as that one was.

I imagine in an open-source context, management details are even more
difficult to achieve, since there is no enlightened CEO who can decree
that the decisions of the principal architect are law. It seems to have
worked with the Linux kernel, where L.T. had the principal architect
role, perhaps by acclamation, rather than executive fiat.

-- 
 .~.  Jean-David Beyer           Registered Linux User 85642.
 /V\                             Registered Machine    73926.
/( )\ Shrewsbury, New Jersey
^^-^^ 12:20pm up 11 days, 21:08, 2 users, load average: 2.14, 2.12, 2.10