[rules-users] Re: [drools-user] More Fun with JBoss Rules IDE and JBoss Rules in general

[Geoffrey Wiseman]

On 5/24/07, Scott Reed <sreed at spamcop.net> wrote:
>
> Unless I never looked at the name except in the source code, I don't see
> the much point in having
> two names that can be distinguished (due to hashcode inclusion) but have
> been semantically
> disconnected from the source code rules. If an error occurs in one of
> those rules, how will you know
> which one the message refers to? How will you make sense out of the log?
> It makes more sense to me
> to replace non-alphanumeric characters with alphanumerics as Ron's team
> member proposed.


Yeah, it's true that including a hashcode allows you to go from the "real"
name to a unique generated one, but doesn't really help you go back the
other way, unless you ALSO store the original name as a string in the rule,
where you can use it to refer to the rule by name in exceptions and so
forth.

An alpha replacement could be difficult given the number of potential
non-alpha characters.

The other option, probably more work, would simply be to improve the
validation and error reporting such that it was clear which two
special-characters-included rules conflicted, and why they conflict.

e.g. "RNR Qty == RNR Adjustment Qty : 100% Tolerance" has the same
translated rules name as "RNR Qty >= RNR Adjustment Qty : 100% Tolerance";
please rename one of these rules.

That'd make the problem a little more clear, but I don't like it as much.

Anyway -- I'm not writing the code, so I'll just let Edson et. al figure it
out.  :)

  - Geoffrey
-- 
Geoffrey Wiseman
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/rules-users/attachments/20070524/b417ace1/attachment.html