mission critical apps; was Re: cross platform ide

[Alex Rice]

On Feb 8, 2004, at 12:11 AM, Ken Norris wrote:
> We can plan exactly how and where we want to trap messages and provide
> callbacks that doublecheck for a proper response, i.e., one that falls
> within predetermined limits, can't we?

Formalize it and be able to explain it to managers and other 
programmers. Make a case study for runrev to post on their website!

I've been using runrev for a couple of years and I'm not writing code 
as well-planned as you describe. I'm still trying to get errorDialog 
working in my standalones, and figuring out why exception handling 
seems quirky sometimes.

I know someone who claims to write all his code so it's "mathematically 
tested to be correct" through some kind of elaborate unit testing (a C 
programmer actually, not runrev). That's one extreme. I don't know 
anyone else that has that kind of discipline or even if this person was 
really on the level.

With xtalks, and smalltalk, one just cannot know if a message is going 
to be accepted by an object until *runtime*. That is, run the app and 
see what happens. To an manager in charge of software for big $$ 
transactions, that translates into a craps shoot.

How do you ensure correctness in your app? Some languages have features 
like static typing, assertions, and robust exception handling, features 
that facilitate getting the correct, required behavior at runtime. How 
do you compensate for the absence of those features in your xtalk 
programming?

You've got xtalks at one end of the spectrum and and java, ada, eiffel, 
etc. at the other end of the spectrum. Just something to be aware of 
when you take runrev into corporate IT environments.

--
Alex Rice | Mindlube Software | http://mindlube.com