Profiling: best practices to exclude nodes rather than includingthem

[Georges Schmitz <georges.schmitz@arcor.de>]

Hi,

we are writing a modular documentation (manual, help, tutorial) for a 
modular software. There are gui components in the software, that appear 
in nearly all software modules, but sometimes with small changes. I'm 
using "conformance" as attribute to express that it belongs to a certain 
module and "condition" for manual, help and tutorial.

(A)
Now, a tutorial should be less crowded with extensive explanations of 
every gui label, than should eg. the help system. Thus I want to say: 
these 2 paras should go into a tutorial and the rest of the section 
(maybe 8 other paras) not.

Do I have to tag 2 paras with condition="manual help tutorial" and the 
other 8 ones with condition="manual help"? In no time, you don't see 
anymore, which part of your document appears in which output.

(B)
Similar problem with my conformance attribute: there is a central gui 
component in nearly all software modules, but eg. in module "mod7" there 
are missing 2 buttons, the other 7 modules use these buttons. As far as 
I can see, I have to write conformance="mod1 mod2 mod3 ... mod8", 
leaving out mod7 in the list.

It would help a lot, if I could use a profiling attribute by "negation", 
saying in pseudo-style something like conformance="not(mod7)".

My question is: are there some methods to get profiling complexity down 
in my environment?

Thanks for any suggestion,
Georges