Public and Commercial Software Disasters

Daz darrenp at bruce.cs.monash.edu.au
Wed Jun 8 22:19:09 EST 1994


David Huen (s.d.huen at bham.ac.uk) wrote:
: In article <2svifr$smu at mserv1.dl.ac.uk>, Duncan Rouch
: <drouch at molbiol.ox.ac.uk> wrote:

: <stuff deleted>

: > The prime cause is usually management ignorance combined with bad commercial
: > consultants.  So, it might be useful if some of the computer-science know-how 
: > at universities was used in consultancies to help stop wasting all that
: > public money on modern dinosaurs, not to mention to help the police,
: > firebrigade and ambulance services do better jobs. 
: > 
: I doubt if many academics have ever implemented large systems, involving
: large development teams, with tight performance and downtime specs and
: deadlines that had to be used by relatively untrained users. If the average
: university computer centre represents the level of service an academic
: consultancy can yield you in a life-critical program, I'd invest in a
: decent ambulance-chasing legal practice.. :-).

: The skills and experience you'd want are at the top end of the commercial
: sector, in businesses and organisations which have better things to do than
: make large donations to a certain political party.

: > 
: > Duncan Rouch
: > New Chemistry Lab, University of Oxford, UK.

: Hi Dunx, So that's where you are!!!!

I'd come in somewhere in between - as a PhD student at university and having
a part time interest in consulting and commercial computing, I'd say that both
sies of the fence are lacking certain skills - certainly universities have a
fairly disasterous management record for biggish projects, as they tend to
encourage unrestrained investigation rather than disciplined adherence to
a plan. I wouldn't equate a university CC with university programming either,
since they have typically a different kind of staff to a computing faculty for
example. 

On the other side of the fence, I am continually sickened at the
generally low amounts of thought that seem to go into most commerical pieces
of software. Certain large software vendors who shall remain as anonymous
as their outstanding incompetence permits seem to have an ongoing commitment
to vending solutions with a very low technological component. Most of the
human race doesn't reaslise for example that it's unreasonable for their
entire machine to crash when one application misbehaves, or that the whole
machine doesn't have to grind to a halt whilst formatting a floppy or doing
a high speed modem transfer. Watching an X server crawling when you open
a few thousand windows (quite reasonably) b/c the programmer probably used
a link list to hold the window list rather than a hash table.

The programmers who tend to remain at uni will tend to be smarter on average,
think there's scope for some more intelligent programming, and for universities
to learn better project management skills.

My 2 cents (australian) worth,

Daz.




More information about the Bio-soft mailing list