On Fri, Feb 12, 2010 at 12:55 AM, Shawn Walker <swal...@opensolaris.org> wrote:
> On 02/ 1/10 06:33 PM, Danek Duvall wrote:
>>
>> I'm trying to run a query with the following structure:
>>
>>     Z AND (X OR Y)
>>
>> where X is a compound (ANDed) query but I can't figure out how to do it.
>> There is no provision, AFAICT, for putting parens around expressions, and
>> no way to decompose the entire expression via deMorgan's laws into
>> something that can be expressed.  This would require either the ability to
>> negate a group of charts:
>>
>>     NOT (NOT X AND NOT Y)
>>
>> where "NOT X" and "NOT Y" would be separate charts or the ability to OR
>> charts together, as straightforwardly implied by the original statement.
>>
>> Is there something I'm missing, either in the bugzilla interface or in the
>> boolean algebra toolchest?

I think I gave Danek an outline which should actually work for this problem.

The trick is to generate one chart which is the union of all possible
bits you might need.
And then you add additional charts which delete portions of the set.

Basically, consider the world using a Venn Diagram.
The universal set is all bugs.
Each condition you want or don't want should be a single ring in your
Venn diagram.
In chart number one, you describe the union of all of those rings
(using OR), which makes a set which excludes bugs which aren't matched
by any condition described by your query.
In each subsequent chart, use "negate this chart", and then use AND to
describe a set which you need to exclude because it is some set of
conditions which overlaps with the set you want but has some criteria
which doesn't belong.

You can end up with many additional charts of this type, however
usually you won't need to. I believe for the specific case Danek
wanted it was sufficient to use two or three charts.

Note that if the basic range for your query can be described in the
non boolean charts section, then you can just describe it there and
use the negation part for all of the corrections.

I'm very sorry I didn't write this up earlier, I did discuss it with
danek, but asked him to try the query and get back to me / reply to
the list.
_______________________________________________
tools-discuss mailing list
tools-discuss@opensolaris.org

Reply via email to