In a message dated: Tue, 01 Feb 2000 21:08:11 +0100
Ferenc Tamas Gyurcsan said:
>Another civil engineer will add that it is definitely true:-). In every aspect
>of engineering. You can't even tell what's the peak value. You can just tell
>the peak value that happened in the last 100-200 years (in water management),
>and then you can evaluate with the cost/benefit ratio. But this is kind of far
>from CS:-).
Not really, since everyone always seems to want to decide what to do in the
future based upon what has happened in the past. Management alwasy attempts
to budget based on what they did last year. Occasionally they'll try to
figure in some growth but that's rarely accurate.
For instance, budgeting time comes around and you know you need to buy
another NFS server. Management will ask several questions like:
How much disk space are we using now?
How much were we using this time last year?
What's that growth rate?
How much do we use per person?
Based upon these numbers, and the expected growth, how
much should we buy?
All very good questions. And they might even get you in the ballpark. But
past performance is *never* a guarantee of future performance (as wall
street will tell you :)
So, invariably, something changes making all these planning numbers completely
irrelevent :) And you know it's always going to be on the negative side,
since Murphy is always looking over your shoulder. The growth of the company
may be negative, but for some reason people are using more disk space (r�sum�s
most likely :)
So, in general, the same kind of capacity planning that applies to other forms
of engineering also apply to the IT world, and in the same ways:
- Management will never let you buy what you need to get the job done
correctly
- They will never make any approvals on time
- By the time they do, you need more than you asked for
- By the time you get what they gave you, you need to get more
- No matter how well you do implementing the project with what you're
given, it's never good, cheap, or fast enough
- You'll always have to revisit this issue again within 6-12 months
Note, these guidelines only apply to *HUGE* corporations or gov't entities :)
In the small company world, things work just fine (I know, I've been on all
sides!:)
--
Seeya,
Paul
----
Doing something stupid always costs less (up front) than doing
something intelligent.
Bean counters are *always* wrong!
A conclusion is simply the place where you got tired of thinking.
If you're not having fun, you're not doing it right!
**********************************************************
To unsubscribe from this list, send mail to
[EMAIL PROTECTED] with the following text in the
*body* (*not* the subject line) of the letter:
unsubscribe gnhlug
**********************************************************
