and now, as promised, the latest despatches from the wild, wonderful
world of corporate clulessness:
ACT III: JUST SHOOT ME NOW
(in which a $100K+ software installation is scheduled for next week,
and the vendor will be arriving with hardware known to be flaky)
since January, my division has been in negotiations with the vendor of
a searchable database package. that part doesn't really interest me,
to be honest, but i've supported the deal because the product has a
front-end parser which normalizes and categorizes a wide range of
input sources.
i'm much more interested in the parser than in the rest of the
package. one of the biggest time-demands in managing content for the
web is trying to keep it organized. the manual approach, where
someone reads everything that comes in and assigns it to a category,
is slow as hell. it's also prone to all sorts of errors and
instabilities, because people get bored and their minds wander. this
product does the same kind of work automatically, but is much faster
and much more predictable. it's not perfect, but it can reduce the
load on the human supervisors dramatically. it frees them to
concentrate on the relatively few situations where classifying a story
requires a human judgement call.
when i say all that to my higher-ups, i can see the film of
incomprehension spreading across their eyes. they know i'm excited
about something, but don't have the background to understand exactly
what. i've been tempted to switch from English to some other
language while talking, just to see if there's any change in their
expressions. what they care about is the marketing of the front-end
applications. i understand the value of that, but IMO, that's a
narrow, and somewhat brittle, view of the product. my interest is in
using the product as part of our general infrastructure, where it can
have pervasive, if less obvious, value.
i'm not too worried about that part of the debate, because our goals
aren't mutually exclusive. the differences are mostly a matter of
emphasis. they'll see tuning the search engine for a specific
feature as a strategically important business move, and i'll see it as
a training exercise that gets my staff more familiar with the system.
in the long run, i know i'll get what i want.. the geeks run the
system, and i run the geeks.
as is inevitable with any piece of software that hasn't been reduced
to shrink-wrap status, there have been teething problems with the
setup and installation of this project. the biggest problem along
the way has been a dynamic between the two companies where the
vendor's sales rep says, "sure, we can do that", and my bosses say
"okay".
now, i would've thought that anyone who reads Dilbert would understand
the comparative intelligence of taking a sales rep at their word.
all i can figure is that my bosses don't read Dilbert. i've gained a
reputation in both companies for something which hovers between
cynicism and prophetic wisdom by saying things like, "you believed a
*vendor*?" and "what if they *can't* make it work?"
the part which is most incongrouous to me is that both the guys above
me are intelligent, and have solid reputations as serious contract
negotiators. when it comes to the paper agreement, they seem to be
on top of the situation. their naivete when it comes to the
implementation side astonishes me, though. it's like drafting a
disarmament plan for Germany, but not paying attention to all those
guys with shovels.
one of the ongoing sources of trouble has had to do with the hardware
the system will run on. since this is a software product, the
suggested hardware clause was written as a throwaway item in the
contract. they offered to sell us a set of plain-vanilla servers at
about ten times the price for which we could go out and buy them
ourselves. that led to a round of finagling in which we chipped them
back to the point of agreeing to use hardware we provide.
in what's turned out to be an extremely unfortunate coincidence, the
vendor happens to have developed a version of their product for Irix,
the Silicon Graphics version of unix, and we've been running SGI Indys
as our primary webservers for some time. it was arranged that my
company would clear off two Indys, and ship them out to the vendor for
preliminary software installation and testing.
time and experience have now shown that the vendor does indeed have a
version of the product for Irix, but their technical staff is not
especially fluent with that OS. they've called us several times to
warn us of problems with the machines, most of which my systems people
have been able to resolve over the phone. to make matters worse,
though, one of the machines we sent is fairly old, and has received
enough rough handling over its lifetime that the physical hardware has
become unstable.
we didn't know this at the time we sent it, but a series of hardware
failures in our own production servers since then has shown us that
Indys get flaky with age and neglect.
the combination of the vendor's unfamiliarity with a working SGI
system, and a hardware platform which is in itself unstable, has
produced a situation where nobody knows what's going wrong, or why.
to round out the nightmare, both my bosses believe that schedule
pressure boosts productivity.
i happen to hate that particular myth with a passion, but that's a
subject for a whole different rant. in the current circumstances,
the result is a situation where nobody has time to do things right at
the outset, but everyone has an excuse for the problems being somebody
else's fault.
the installation of the system was scheduled for the end of August,
but the schedule slipped because one of the Indys would shut down if
you set a book down next to it on the table. obviously, some
component in the thing has gone marginal, but there's no way we can
learn what until the machine goes into the shop for testing.
the vendor didn't confess this problem until after the schedule
slipped, and it was too late to supply them with another, stable unit
as a replacement. we have the machine to substitute on hand, but
with the pressure to get people on site and doing stuff ASAFP,
shipping it to east coast would have caused an unacceptable delay.
the vendor has done what they think to be a good installation on the
unstable machine, and will try to swap drives when they arrive on
site.
a normal installation takes a whole business week, but we're losing a
day to the Labor Day weekend. that translates to a 20% schedule
compression, which is somewhere beyond optimism even under normal
circumstances. the hardware coming in is known to have problems, and
everyone is hoping the on-site swap will go off without a hitch and
not cost us any time loss. everyone but me, that is. i'm hoping
that the latent problems come to the surface while the vendor is here,
rather than hiding beneath the surface until we put the system into
production.
regardless of what does or doesn't happen, it's gonna be one hell of a
week.
mike stone <[EMAIL PROTECTED]> 'net geek..
been there, done that, have network, will travel.
____________________________________________________________________
--------------------------------------------------------------------
Join The Web Consultants Association : Register on our web site Now
Web Consultants Web Site : http://just4u.com/webconsultants
If you lose the instructions All subscription/unsubscribing can be done
directly from our website for all our lists.
---------------------------------------------------------------------
