Hey guys,
Several good points have been brought up regarding the cost of WO
vs. other application servers, and the relative cost of WO as a component
of an overall project. It's true that when you compare WO to say, Oracle
App Server or NetDynamics in the context of a large scale project replete
with big UNIX iron and a top-shelf RDBM$, the cost is not unreasonable.
However, that describes about 5% of the projects I actually see
in my environment. Mostly we do a lot of onesy-twosy stuff where we're
using either a cheap database (i.e., MS-SQL) or an existing DB server.
Typically, the app may have a dozen components, and maybe one subproject,
and is generally intended to "webbify" some old Visual Basic application
or provide web-based access to an existing database. These projects get
done on budgets that don't exceed $10K, hardware and all -- and often
they're "unofficial" projects that don't have a budget at all. Maybe
an application server is overkill for something like this, but WO seems
pretty lightweight in terms of overhead, and as a developer I appreciate
the benefits of working in that environment. As a systems engineer, I
happen to think that the application server model is far superior to
other architectural options.
But all that is hard to sell, and the pointy-haired managers
here don't understand and don't care to. Their attitude is "hey, if
ASP works for Dell with all the traffic they get, why won't it work
for me? Why should I spend all this money?"
And all I'm saying is, they're right. Sure, I crack on ASP
as much as the next guy, but I defy any one of you to describe an
application that *couldn't* be written with it. I can hear all the
"yeah, but"s, and I have several myself, but the fact is if you throw
enough hardware at it, you can make that pig perform -- although you'll
find yourself building a fault-tolerant network of NT boxes if you want
the app to stay up. But for the $25K WO costs, I can buy a big-ass
4-6 CPU machine with 2-4 Gigs of RAM that will run any app I'm likely
to write. And these apps typically will never scale very large.
If I was a contractor or a software house that built stuff for
deep-pocketed clients, it wouldn't be a real issue for me. But I'm
just one of those overworked systems guys shoehorned into doing
frequent in-house development of what amount to low-end intranet apps.
I guess what I'd really like to see is some low-end version
similar to the stuff NeXT gave away for free in WO 1.0 and 2.0 --
maybe it just does webscript, no Java or Obj-C (well, Obj-C would
be *really nice*, but one has to make some concessions). It *has* to
have database connectivity. Maybe give customers the option of adding
various features "a la carte". And whether you stick with tpm or
per-user licensing, offer affordable "bump packs" in small increments
of say, 10 users or 10 tpm.
My boss is fond of saying "Only Microsoft understands the
corporate purchasing model". Unfortunately, at least in the case
of my employer, he's usually right.
Off the soapbox is
Troy
--
Troy D. Casey -- [EMAIL PROTECTED]
"I can retain neither respect nor affection for a Government which has
been moving from wrong to wrong in order to defend its immorality."
-- Mohandas Gandhi (on Britain)
