Re: Where is Oracle’s Grid ?

Daniel Morgan <damorgan_at_x.washington.edu> wrote in message news:<1072914858.897670_at_yasure>...
> Joel Garry wrote:
> > Daniel Morgan <damorgan_at_x.washington.edu> wrote in message news:<1072810464.715634_at_yasure>...
> >
> > Saying "just buy good software" is ridiculous, there simply isn't any
> > such thing. Saying "write it yourself, look at Amazon" is a little
> > more subtly ridiculous, as it ignores the timing of the dotcom madness
> > in relation to their capital investment needs. Software houses with a
> > small staff can come up with some excellent stuff, but the scale is
> > all wrong (or even if it can scale, there is no efficient mechanism to
> > sell it to decision makers). The only answer is to explicitly
> > recognize that the length of software life cycles requires long term
> > stable software to base applications upon. This is in conflict with
> > the reality of quarterly-biased financial results in the capital
> > markets.
> >
> > Put another way, Larry is better off hacking up Peoplesoft than taking
> > responsibility for bad Forms, Reports and Apps.
> >
> > jg
> > --
> > @home.com is bogus.
> > http://www.wired.com/wired/archive/12.01/internet.html?tw=wn_tophead_2
>
> If you are talking about recreating an Enterprise sized accounting or
> distribution system I agree. The investment is huge. But there are a lot
> of places other than Amazon that have done well too. Heard any
> complaints from Visa? Mastercard? American Express? New York Stock
> Exchange? Their apps are not COTS.

I think we agree here. But I must point out that we aren't likely to hear about failed projects at those places, and certainly aren't going to hear complaints if there are any. Did we ever find out exactly what happened at Orbitz? We sometimes do hear about failed projects at government agencies because of the eventual accountability and the fourth estate. Funny how Dilbert can apply to just about any commercial site but nobody ever admits they are a PHB.

>
> But the truth is that the vast majority of apps are not huge projects.
> Most could be written from scratch by a team of a half-dozen to a dozen
> really competent developers in 6-12 months.
>
> The problem is that managment has yet to figure out how to make 10 guys
> at $200,000/yr look more appealing than 100 guys at $20,000/yr. So they
> keep ending up with trash.

A range of skills should be best, don't you think? And don't most 100-guy projects have that? Otherwise you have too-many-chiefs syndrome. Saying 10 $200K guys is good is making the same skillset interchangeability error that green MBA's make with the 100 guys.

jg

--
@home.com is bogus.
hypergraphia! http://www.signonsandiego.com/news/uniontrib/wed/currents/news_1c31writer.html