>This sounds like a positively cryptic way to structure a VI.

Not necessarily.

>What is good about a BD or FP that is pretty or clean or sterile, 
>if it communicates nothing about its function?

Well what it does communicate is a structured and designed approach, and my argument 
would be that a simple diagram will always communicate it's function. (See my comments 
below). I would argue that having a non-sterile BD doesn't necessarily include the 
case that it communicates the VI's function! You could also argue that you can always 
work out what's going on, but taking the time to work out what's going on is exactly 
the reason to make it sterile and easy to read - to reduce this time.

>And further, can such a VI be composed that doesn't require a large
>recomposition to clean it up to this state?  Mostly, I never get 
>to do anything over.  So it is all experimental.  Then when it works I 
>don't have the luxury of rewriting it for aesthetic reasons. Because LV 
>is self-documenting to a great degree if you habituate yourself to a 
>few things, like naming states and sub-vis cogently, why would I want 
>to make that stuff any harder to get to?
>
>I am actually offended by the ideas of a sterile top level.

Hmm. I guess this probably comes back to software development origins or development 
constraints. In programming for whatever the reason (time/money/research 
based/experience), the development process is predominantly incremental and when you 
arrive at the end point you're done. A structured approach  is only possible with some 
sort of planning and some sort of requirements before the development begins. Often 
this is not possible (as well I know;-), but even the most rudimentary "unspecified" 
project can benefit from planning. And incremental doesn't mean that it has to end up 
non-sterile. The primary benefits here are maintainability and extendability. Neither 
of these are requirements in a project that only has one goal; eg a research based or 
sole functional target, where it is never going to be maintained or enhanced. However 
if the complexity grows and grows, one day you will find yourself restarting from NULL 
because you cannot possible add that one last feature a client has requested. <This 
usually occurs at 2am with a deadline fast approaching>

Having developed in the non-sterile way and the sterile way I would prefer sterility 
;-) As my wife once said having looked over my shoulder, "How boring, all your code 
looks the same." I was pleased but I think she meant to it as a jibe!

In our company, it is no longer acceptable to program in such an ad hoc fashion. Not 
because this doesn't achive results, but because the additional benefits of the a 
structured design are that (1) all members of the development team know where 
everything is located in any project (2) there is decreased learning time on software 
when the structure follows one of a pre-selected set of design patterns. These lead 
directly to a decreased development cost.

cheers, Alex.


Reply via email to