From: Lee Sau Dan on
>>>>> "blmblm" == blmblm <blmblm(a)myrealbox.com> writes:

blmblm> So, your proposed GUI also needs to be able to store its
blmblm> results for later processing? Okay.

blmblm> Or maybe you're just saying "do two separate searches".
blmblm> Which is probably what most people would do anyway, and in
blmblm> fact the only benefit I can think of in doing it as I
blmblm> originally described is that the combined search would
blmblm> only have to traverse the directory structure once.

Doing it in 1 search has another advantage: no duplicates in the
result list.


It's actually related to optimization. Doing 2 separate searches by
invoking 'find' 2 times makes that the 2 queries cannot be related to
one another by 'find'. 'find' will be unable to do any optimization
(if the implementation does it). Doing it in 1 single 'find' command
opens up the possibility of optimizing the whole query to 'find'.




--
Lee Sau Dan §õ¦u´° ~{@nJX6X~}

E-mail: danlee(a)informatik.uni-freiburg.de
Home page: http://www.informatik.uni-freiburg.de/~danlee
From: Lee Sau Dan on
>>>>> "Peter" == Peter T Breuer <ptb(a)oboe.it.uc3m.es> writes:

Peter> In comp.os.linux.misc Lee Sau Dan
Peter> <danlee(a)informatik.uni-freiburg.de> wrote:
>>>>>>> "Peter" == Peter T Breuer <ptb(a)oboe.it.uc3m.es> writes:

>> >> How do I enter something that gives the functionality of "(A
>> >> and B) or C"?

Peter> Chose A, C checkboxes. Click "again". Chose B, C
Peter> checkboxes. Click "finish" ("search", whatever).

>> So, you've done the conversion to CNF in your brain, haven't
>> you?

Peter> I don't need to "figure" for something so simple. It's
Peter> obvious.


>> And what should you do when C is not an atomic predicate,

Peter> You will have to express yourself to me in atomics in order
Peter> to explain yourself. I'm afraid I don't speak the language
Peter> you do.

So, original query:

(A and B)

fined query:

(A and B) or (P and (Q or R))

where A, B, P, Q, R are all atomic.


How do you do that *conveniently* in your GUI?


--
Lee Sau Dan §õ¦u´° ~{@nJX6X~}

E-mail: danlee(a)informatik.uni-freiburg.de
Home page: http://www.informatik.uni-freiburg.de/~danlee
From: Lee Sau Dan on
>>>>> "Tobias" == Tobias Brox <tobias(a)stud.cs.uit.no> writes:

Tobias> [Lee Sau Dan]
Peter> Chose A, C checkboxes. Click "again". Chose B, C
Peter> checkboxes. Click "finish" ("search", whatever).

>> So, you've done the conversion to CNF in your brain, haven't
>> you?

Tobias> A user accustomized to such a tool will think in CNF, no
Tobias> conversion needed.

And this tool is "easy" to learn?


--
Lee Sau Dan §õ¦u´° ~{@nJX6X~}

E-mail: danlee(a)informatik.uni-freiburg.de
Home page: http://www.informatik.uni-freiburg.de/~danlee
From: Lee Sau Dan on
>>>>> "Peter" == Peter T Breuer <ptb(a)oboe.it.uc3m.es> writes:

>> So, that means you can't optimize the query to make the search
>> more efficient.

Peter> You can optimize the query all you like subject to the
Peter> condition that any ORs in the query will have to be handled
Peter> by you, though, not the google DB engine.

So, you have to do that optimization in your UI implementation,
instead of simply leaving it to 'find'?


Peter> So what?

So, you're reinventing the wheel.


Peter> Using two google searches underneath is mere
Peter> implementation! Leave that sort of thing up to the
Peter> implementation.

>> Unoptimized, inefficient implementaion.

Peter> Leave it up to the implementation.

Why reinvent something that 'find' already does?


Peter> Google has plenty of horsepower.

But not to be monopolised by you a single user.


Peter> Do you have more? I think your task is just to do an cat
Peter> and sort! The lower-level queries can go to google in
Peter> parallel.

If Google supported disjunction in the engine, more optimizations are
possible by pushing the OR operation down to the engine level.



--
Lee Sau Dan §õ¦u´° ~{@nJX6X~}

E-mail: danlee(a)informatik.uni-freiburg.de
Home page: http://www.informatik.uni-freiburg.de/~danlee
From: Peter T. Breuer on
In comp.os.linux.misc Lee Sau Dan <danlee(a)informatik.uni-freiburg.de> wrote:
> It's actually related to optimization.

Look - if google wanted to optimize they have my FULL permission to
take mine and everybody elses (who is searching at the same time)
searches and optimise over them simultaneously.

It's not our business.

Peter