Kurt
> > > > the next dentistry software being written by dentists:
> > > This is incorrent because it presupposes that a dentist is the best
> > > person describe a dentistry practice
> > But surely he is?
>
> Probably....
The point was that as the impedance to developing information systems drops, so
it becomes more accessible to any given domain expert. Therefore if there is
someone controlling the Galactic Dentistry Federation, that person can
computerise his tasks just as easily as the lone dentist at Pluto Dental
Services.
> And where does Open Source then figure in this? Is it entirely
> irrelevant,
I dont see how it changes the point of making systems easier to create, unless
you factor in the wider audience of experts, which would contribute positively
to the argument of making computer systems easier to craft.
>
> > In my experience you ask most people about their businesss and they
> > will give you long convoluted rerason why it can't be systemised
> Is this the manager of that business or the owner?
> And why don't they consider their accounts, and their internal
> "systems" - certain ways of doing say petty cash, staff salaries etc
> as a particular implementation model?
this is exactly the point of most small business teachings (see 'the E-Myth'
for example). The argument goes that small businesses can reduce the chance of
failure by focusing on the Business rather than the Trade aspects of their
venture (a painter who suffers cash flow problems is better served painting
LESS and correcting the cashflow failure cause). Clever business owners will
grasps their business systems, as you say. The point I was making was that the
neophyte painter would tell you that you cannot systemise painting, yet that is
precisely what the experienced painter/businessman should do.
> > The secret to your system modeling skills is a cyclic process of
> > RAD-deploy-trial, RAD-deploy-trial, in a spiral fashion. System
> > modelling in the big bang model fails because the user behaviour is
> > changed by the system that is applied. There is no process that
> > allows system modelling to anticipate how the user will adapt his
> > work to the new model.
>
> True, but with Incis as an example, wouldn't it be fair to say that
> sometimes the user (ie the local bobby) has no say in the new model,
> and will be (attempted to be) forced to change to the new model
that was the flaw in the process IBM used, in fact the local bobby should have
been given scope to craft or at least define the system.
> > Given this, there is no
> > reason why the dentist cannot mold the product. And the end result
> > will be exactly right for the dentist, but will be fundamentally
> > different to the model of the dentistry requirements one would
> > create at the beginning.
>
> As above, there is a case - the most often one (?) - where the
> reason that the dentist cannot mould the product is that his manager
> wants it to go in one direction, and his manager wants it to go
> in another, and the accountant wants it cheaper etc. And the people
> who can get closest to the users level are system modellers -
> analysts, developers or software engineers of some kind.
You assume that best IT result is a large integrated system. This is the old
way and is flawed in the future model as I see it. The manager would have a
system relevant to his domain, and so the accountant, etc. Sure, there exists
data sharing between them now, and by computerising the individual entities
there would be shared data in the new system. Perhaps this would happen to a
greater degree, and perhaps more efficiently. But the salient point is that
each individual is served by systems (s)he creates to suit the moment, and
which will then be adapted as the work adjusts to the new process. The fallout
might be a reduced need for, say, accountants, at the low level, and they would
then contribute their skills at a more executive level in the process. Same as
the programmers
{end of rave}
Leo
---------------------------------------------------------------------------
New Zealand Delphi Users group - Delphi List - [EMAIL PROTECTED]
Website: http://www.delphi.org.nz
To UnSub, send email to: [EMAIL PROTECTED]
with body of "unsubscribe delphi"
