[vwnc] Set equality testing...
Locked
 
|
Hi all..
Just a quick question.. I ran across a difference between Squeak and VW last night that had me wondering, and while I could extend the Set class to cover it, I'm not sure it wouldn't have ramifications with other areas within VW that use the Set class.. Anyway, the issue revolves around the use of #= on a pair of Sets.. In Squeak, the pair of sets are considered equal if that have the same contents but not necessarily if they're the same underlying objects. In VW the Set class defers the comparison to the Object base class which maps #= to #== so in this case the sets are only equal IF they both are the exact same object.. Anyway, not sure if there's some underlying reasoning behind this but thought I'd ask regardless.. TIA! -- Rick _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: [vwnc] Set equality testing...
|
|
Rick,
There's #elementsEqual: in Com-Extensions that'll do the trick for you, Collection>>elementsEqual: otherCollection "Answer <true> if the elements of the receiver and the <otherCollection> are equal." self size = otherCollection size ifFalse: [^false]. self do: [:each | (otherCollection includes: each) ifFalse: [^false]]. ^true Cheers, -Boris -- +1.604.689.0322 DeepCove Labs Ltd. 4th floor 595 Howe Street Vancouver, Canada V6C 2T5 http://tinyurl.com/r7uw4 [hidden email] CONFIDENTIALITY NOTICE This email is intended only for the persons named in the message header. Unless otherwise indicated, it contains information that is private and confidential. If you have received it in error, please notify the sender and delete the entire message including any attachments. Thank you. -----Original Message----- From: [hidden email] [mailto:[hidden email]] On Behalf Of Rick Flower Sent: September 23, 2008 9:43 AM To: [hidden email] Subject: [vwnc] Set equality testing... Hi all.. Just a quick question.. I ran across a difference between Squeak and VW last night that had me wondering, and while I could extend the Set class to cover it, I'm not sure it wouldn't have ramifications with other areas within VW that use the Set class.. Anyway, the issue revolves around the use of #= on a pair of Sets.. In Squeak, the pair of sets are considered equal if that have the same contents but not necessarily if they're the same underlying objects. In VW the Set class defers the comparison to the Object base class which maps #= to #== so in this case the sets are only equal IF they both are the exact same object.. Anyway, not sure if there's some underlying reasoning behind this but thought I'd ask regardless.. TIA! -- Rick _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: [vwnc] Set equality testing...
|
|
In reply to this post by Rick Flower
On Tue, Sep 23, 2008 at 9:42 AM, Rick Flower <[hidden email]> wrote: Hi all.. VW let's the perfect be the enemy of the good. Because a Set could contain itself (or the two sets being compared could include each other) there's potential for the comparison to never terminate, or to be slow when complicated by adding a visited set to prevent recursion.
Squeak follows the 95% rule and assumes that most of the time when you compare sets they're not recursive and so the simple implementation will suffice. I think Squeak's approach is far superior. There's nothing to prevent adding a more sophisticated comparison function that handles recursion, and the naive #= implementation is useful and works most of the time. The 95% rule was used when Smalltalk was being implemented in the first place. IMO, it is still a good maxim.
Eliot
_______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: [vwnc] Set equality testing...
|
|
In reply to this post by Rick Flower
Hello,
> In VW the Set class > defers the comparison to the Object base class which maps #= to #== so in > this case the sets are only equal IF they both are the exact same object.. > Not anymore in 7.6... now Sets are equal if they are of the same size and if they have the same (recursively checked) contents. Both #= and #hash tend to assume the receivers are not recursive. Andres. _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: [vwnc] Set equality testing...
|
|
On Tue, September 23, 2008 11:13 am, Andres Valloud wrote:
> Hello, >> In VW the Set class >> defers the comparison to the Object base class which maps #= to #== so >> in >> this case the sets are only equal IF they both are the exact same >> object.. >> > > Not anymore in 7.6... now Sets are equal if they are of the same size > and if they have the same (recursively checked) contents. Both #= and > #hash tend to assume the receivers are not recursive. Hmm.. I am using 7.6 and stepped into the code for #= and it was down in the Object class which in turn invoked #== (also in the Object class). Perhaps I've loaded something which removed the stock functionality? I'll take a look when I get home.. Thanks everyone for the suggestions/info! _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: [vwnc] Set equality testing...
|
|
Affirmative,
Set new = Set new -> false This is in 7.6, Andres? -Boris -- +1.604.689.0322 DeepCove Labs Ltd. 4th floor 595 Howe Street Vancouver, Canada V6C 2T5 http://tinyurl.com/r7uw4 [hidden email] CONFIDENTIALITY NOTICE This email is intended only for the persons named in the message header. Unless otherwise indicated, it contains information that is private and confidential. If you have received it in error, please notify the sender and delete the entire message including any attachments. Thank you. -----Original Message----- From: [hidden email] [mailto:[hidden email]] On Behalf Of Rick Flower Sent: September 23, 2008 11:53 AM To: Andres Valloud Cc: [hidden email] Subject: Re: [vwnc] Set equality testing... On Tue, September 23, 2008 11:13 am, Andres Valloud wrote: > Hello, >> In VW the Set class >> defers the comparison to the Object base class which maps #= to #== so >> in >> this case the sets are only equal IF they both are the exact same >> object.. >> > > Not anymore in 7.6... now Sets are equal if they are of the same size > and if they have the same (recursively checked) contents. Both #= and > #hash tend to assume the receivers are not recursive. Hmm.. I am using 7.6 and stepped into the code for #= and it was down in the Object class which in turn invoked #== (also in the Object class). Perhaps I've loaded something which removed the stock functionality? I'll take a look when I get home.. Thanks everyone for the suggestions/info! _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: [vwnc] Set equality testing...
|
|
In reply to this post by Andres Valloud-3
On Tue, Sep 23, 2008 at 11:13 AM, Andres Valloud <[hidden email]> wrote: Hello, Good news!
_______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: [vwnc] Set equality testing...
|
|
In reply to this post by Boris Popov, DeepCove Labs (SNN)
Set lookUpSelector: #= ifAbsent: [nil] -> CompiledMethod Object>>=
-Boris -- +1.604.689.0322 DeepCove Labs Ltd. 4th floor 595 Howe Street Vancouver, Canada V6C 2T5 http://tinyurl.com/r7uw4 [hidden email] CONFIDENTIALITY NOTICE This email is intended only for the persons named in the message header. Unless otherwise indicated, it contains information that is private and confidential. If you have received it in error, please notify the sender and delete the entire message including any attachments. Thank you. -----Original Message----- From: [hidden email] [mailto:[hidden email]] On Behalf Of Boris Popov Sent: September 23, 2008 11:59 AM To: [hidden email]; Andres Valloud Cc: [hidden email] Subject: Re: [vwnc] Set equality testing... Affirmative, Set new = Set new -> false This is in 7.6, Andres? -Boris -- +1.604.689.0322 DeepCove Labs Ltd. 4th floor 595 Howe Street Vancouver, Canada V6C 2T5 http://tinyurl.com/r7uw4 [hidden email] CONFIDENTIALITY NOTICE This email is intended only for the persons named in the message header. Unless otherwise indicated, it contains information that is private and confidential. If you have received it in error, please notify the sender and delete the entire message including any attachments. Thank you. -----Original Message----- From: [hidden email] [mailto:[hidden email]] On Behalf Of Rick Flower Sent: September 23, 2008 11:53 AM To: Andres Valloud Cc: [hidden email] Subject: Re: [vwnc] Set equality testing... On Tue, September 23, 2008 11:13 am, Andres Valloud wrote: > Hello, >> In VW the Set class >> defers the comparison to the Object base class which maps #= to #== so >> in >> this case the sets are only equal IF they both are the exact same >> object.. >> > > Not anymore in 7.6... now Sets are equal if they are of the same size > and if they have the same (recursively checked) contents. Both #= and > #hash tend to assume the receivers are not recursive. Hmm.. I am using 7.6 and stepped into the code for #= and it was down in the Object class which in turn invoked #== (also in the Object class). Perhaps I've loaded something which removed the stock functionality? I'll take a look when I get home.. Thanks everyone for the suggestions/info! _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
Re: [vwnc] Set equality testing...
|
|
In reply to this post by Boris Popov, DeepCove Labs (SNN)
Oh, doh... it's in for 7.7, but I thought it was out there already in
7.6. Sorry! Andres. Boris Popov wrote: > Affirmative, > > Set new = Set new -> false > > This is in 7.6, > > Andres? > > -Boris > > _______________________________________________ vwnc mailing list [hidden email] http://lists.cs.uiuc.edu/mailman/listinfo/vwnc |
«
Return to VisualWorks
|
1 view|%1 views
| Free forum by Nabble | Edit this page |