Re: sampling algoritm based on indexed constrains

Albretch wrote:


> I have been thinking about an implementation of a sampling algorithm
> that would conditionally constrain selections preferably in Java (ANSI
> C/C++ are fine, too).
>
> Say, you have a number of items you may conditionally select. Cases
> would be:
>
> 1._ you may select any and any number of them,
> 2._ you may select any, but only a given number of them, say 1, 2 or
3
> 3._ you either select the first or second and any other one,
> 4._ you may select any as long as the total of certain weight remains
> in a range
> 5._ you -must- select all
> . .


> I did search and could not find anything similar to what I was
> looking for. Maybe you know of something like this or can avail me
> of some leads in the right direction.




Yes, you *could* write a simple scanner and parser to determine the
validity of sampling criteria. What you do with valid criteria after
that is up to you!


I would type "Scanners", "parsers" and "simple examples" into Google
and try and find a tutorial on how to write them. You will also need
to define the grammar.


Flex would be a good tool to use to see how parsers and scanners work,
although for your problem it might be a little overcomplex :)


A Java binary is available here and there are also some useful papers
and tutorials:
http://www.flex-compiler.lcs.mit.edu/Harpoon/


John

Return to the comp.compilers page.
Search the comp.compilers archives again.