RE: experts

[Gavin Thomas Nicol <gtn@ebt.com>]

> - The result?  Specifications which do too much (and sometimes, even 
> simultaneously, specify too little.)

I think this is a valid complaint. Many specifications
do make a *lot* of assumptions about level of understanding.
Some specs even define some behaviour in terms of what 
it *doesn't* do...

> - Experts are very good at definining processing layers.
> 
> - Experts are very good at achieving optimization by blurring those 
> processing layers.  (They know where the benefits are, how to 
> achieve them, and how to stay out of trouble.)
> 
> - Those experts (perhaps that kind of expert) are willing to combine 
> multiple layers of functionality in a single process.  They know how to 
> make it work safely and efficiently, and think that even the 
> 'bozos' should  be able to figure it out.

I think, again, that *real* experts know what works, and what
doesn't, and will try to make it approachable for others. The
mark of a real expert is know how much is just enough.

A good example from recent experience is transactional integrity.
I had a long debate about the value of exposing transactions
to users... my contention was that it's probably better to
design an interface that hides them, so people don't have to
be bothered by them. The "expert" (by your definition) did
not agree.