Catching Error Messages

Gregory Lypny gregory.lypny at videotron.ca
Thu Aug 10 08:36:53 EDT 2006


	Hello Mark,

	Sorry for my delay in acknowledging your reply.  I'm not a  
statistician, at least not a competent one.  I've created a function  
that spits out p-values for Fisher's Exact Test, and this involves  
computing large factorials.  Revolution can handle at most 175! or  
200! (I forget the exact number), and I've been able to surpass this  
by simplifying fractions in my calculations (i.e., simplifying x!/ 
y!).  However, if a user enters data that would cause Revolution's  
limits to be exceeded, I want to be able to suggest alternative,  
equivalent tests.

	Thank you for recommending the Try control structure.  That's  
exactly what I was looking for.  Unfortunately, there is no example  
provided in the documentation, but I'll mess around with it.

	Regards,

		Gregory Lypny


Mark Schonewille responded to my message:

> Hi Gregory,
>
> Are you a statistician? Sample size should not play any role in the
> size of numbers that Revolution can handle. What exactly are you
> having a problem with? You an catch error messages by using the try-
> catch control structure. Just put this around the syntax that does
> the calculations and show an error dialog if you "catch" an error.
> See the docs for more info.
>
> Some of the information is difficult to retrieve if a stack is
> password protected. I made a shareware library that allows you to
> catch errors in password protected stacks. The library gives you
> several ways to parse and display errors. If this sounds useful, have
> a look at the developers section of the Economy-x-Talk website.
>
> Best regards,
>
> Mark



More information about the use-livecode mailing list