Lawyers, Guns, and Money
Ray Dillinger
bear at sonic.net
Wed Aug 22 14:54:19 PDT 2001
On Tue, 21 Aug 2001, Mac Norton wrote:
>perhaps greater importance, in context. Faddish as it sometimes--
>well, hell, often--is, the academic side of the law is the
>*only* side of the law that even begins to reward originality.
>Those of us who actually represent people find that original
>thinking is the bane of most judges, unless you can make them
>believe the idea started with them.
>
>Any parallels in software, both as to faddishness among
>the "original" thinkers and leader-following otherwise?
At the first job I had out of college, we spent almost six
months busily "updating the look" of an application that
worked perfectly well, so that the sales guys could impress
people not only with how well it worked, but also with how
"slick" it was.
So we did color fades on title bars, and added gray to our white
scroll bars, and rearranged dialog boxes, and exchanged our
athena widgets for motif widgets that worked exactly the same,
and moved stuff from the regular menus to the right-mouse menu,
and made our "drawn oval" buttons into rectangular buttons that
were lighter on the upper and left edges and darker on the lower
and right edges (which, laughably, was called "making them 3-d"),
and converted the internal documentation from troff and TeX to
HTML, and did about a jillion other tiny things that added nearly
zero functionality but brought the old application up to the
current idea of "style".
A year after I left, I heard from someone who had stayed that
they were reimplementing the interface as a "thin client"
application, because thin clients had come into style.
Note -- it had an XWINDOWS INTERFACE to start with!!! your
"thin client" on an Xwindows interface is neither more nor
less than a standard X server running on the machine where
you want the interface to appear!
At another company, a few months ago, a lead developer came
to me announcing that they were developing a parser to
convert an application langauge into an XML schema using a
whizbang database that had somehow sprouted an XML interface,
and they were having trouble with recursive structures in the
formal schema description the database wanted. I said, "is
there a reason we're not using lex? lex handles recursive
structure just fine." The response, of course, is that nobody
wants to tell customers we're using lex to generate our XML,
because lex isn't in style.... but this whizbang database
(and why does a *database* have XML-conversion functions??) is.
I see a lot of engineering effort wasted on silly fads. Good
people spending days and sometimes weeks reinventing wheels
that represent problems that were solved decades ago, just
because the solutions developed then, despite being proven
and correct, are presently out of style. It's a waste of
resources and it pisses me off.
Bear
More information about the Testlist
mailing list