Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: (Fwd) Re: [IEEE P-1788] Motion P1788/M0013.01:ComparisonOperat



Dear all:

To each "<= " a "<" can be defined by "u < v := u <= v and u 'notequal' v" and
to each "<" a "<=" can be defined by "u <= v := u < v or u = v".
See, for instance, [3] in the proposal.

The <= and the < in the proposal are not related that way. However, this is not unusual in mathematics. For vectors and matrices we usually define (aij) <= (bij) and (aij) < (bij) if the relation holds for all aij and bij. Here the < is not the < defined by (u <= v and u notequal v). Let me for short call the latter 'lessthan'. If 0 lessthan (bij) is true, then (bij) may well have components which are zero. While for 0 < (bij) no component of (bij) may be zero. So for vectors and matrices <= and < are not related as in the first paragraph and nobody ever complained about it.

Intervals can be interpreted as special two dimensional vectors.

Best regards
Ulrich



Svetoslav Markov schrieb:
Dear George,

I said erroneously:

"The first and fourst order relations are of different nature."

Sorry, I meant:

The second and fourth order relations are of different nature.

Svetoslav


------- Forwarded message follows -------
From:           	"Svetoslav Markov" <smarkov@xxxxxxxxxx>
To:             	"Corliss, George" <george.corliss@xxxxxxxxxxxxx>,
       	P1788 <stds-1788@xxxxxxxxxxxxxxxxx>
Date sent:      	Wed, 07 Apr 2010 10:40:14 +0300
Subject: Re: [IEEE P-1788] Motion P1788/M0013.01:ComparisonOperations up for discussion
Priority:       	normal

[ Double-click this line for list subscription options ] Dear George,
before your question is correctly answered,
I would like to make the following notational
suggestions.

The first and fourst order relations are of different nature.

To avoid notational confusions I would
suggest that the first order relation is denoted
differently than \le, for instance by
\preceq , as  already  done in several works.

I also suggest that for uniformity the last
notation is changed to  \le (less_than_or_equal_to)

In algebra it is accepted that the strict analogues
of these two order relations are defined by the respective relations and the equality relation.

For example, see:
Garrett Birkhoff. Lattice theory, Providence, Rhode island, 1967.
There the order  relation \le, that is:

\less_than_or_equal_to is first defined by the properties reflexivity, antisymmetry and transitivity. Then it says:
"If x \le  y and  x \neq  y, then one writes  x < y."

Of couse, it could be done reverse, see e. g.

B.L.Van der Waerden. Algebra, Springer, 1971.

    There   first  the (strickt) order relation  x < y is defined,
then it says:

"x \le  y  means that either  x < y  or  x =  y.

At any case one of the two order relations x \le y and x < y
is always a consequence of the other one plus relation "=",
same with the order relation  \preceq

I strongly support the inclusion of \preceq, which has
been often used in our group since 1970.  G. Birkhoff
also advocates this relation in his paper :

Birkhoff, G.: The Role of Order in Computing, in: Moore, R. (ed.), Reliability in Computing, Academic Press, 1988, pp. 357–378. Regards,

Svetoslav

On 7 Apr 2010 at 2:41, Corliss, George wrote:

From:           	"Corliss, George" <george.corliss@xxxxxxxxxxxxx>
To:             	P1788 <stds-1788@xxxxxxxxxxxxxxxxx>
Copies to:      	"Corliss, George" <george.corliss@xxxxxxxxxxxxx>
Subject: Re: [IEEE P-1788] Motion P1788/M0013.01:ComparisonOperations up for discussion
Date sent:      	Wed, 7 Apr 2010 02:41:34 +0000

P1788,

On Apr 6, 2010, at 6:58 AM, Ralph B Kearfott wrote:
Since Motion 13 (comparison operations) has been proposed (by Bo Einarsson) and seconded (by Dan Zuras),
it is now up for discussion as a position paper.  Discussion will proceed
until the end of April 27, after which the voting period will begin.
I try to argue consistently for simplicity.  My interpretation of Motion 13 is to propose a minimal set of comparisons.  That is, given the set described in the motion, which I think of as

   \a == \b iff a1 = b1 && a2 = b2;
         //  \a:   |--------|
         //  \b:   |--------|
\a <= \b iff a1 <= b1 && a2 <= b2;
         //  \a:   |--------|
         //  \b:        |--------|
\a \contained_in \b iff b1 <= a1 && a2 <= b2;
         //  \a:      |--------|
         //  \b:   |--------------|
\a < \b iff a2 < b1;
         //  \a:   |-----|
         //  \b:              |-----|

Can we express any (most?) comparisons I might want to make between intervals in terms of these four comparisons without further recourse to extracting and testing endpoints?

We might want
\a certainly_less_than \b  (\a < \b)
\a certainly_less_than_or_equal_to \b
\a certainly_equals \b  (\a == \b)
\a certainly_greater_than_or_equal_to \b
\a certainly_greater_than \b  (\b < \a)

\a possibly_less_than \b
\a possibly_less_than_or_equal_to \b  (not(\b < \a))
\a possibly_equals \b
\a possibly_greater_than_or_equal_to \b  (not(\a < \b))
\a possibly_greater_than \b

Since the proposed minimal set includes containment, we might add
\a certainly_contained_in \b  (\a \contained_in \b)
\a possibly_contained_in \b

Perhaps ..._contained_in_interior?

and the negation of all the above.  Perhaps others?


Can anyone code each of these in terms of the four proposed comparisons without further recourse to extracting and testing endpoints?  Presumably also allowing not(), negation(), and perhaps some other operations?

John had posed that as a challenge to me.  After an evening of thought, there are several I could not get.  Can you?

If we can do that without TOO ugly expressions, I'm a strong supporter of Motion 13 for its simplicity.

Alternatively, is there a well-developed theory of a minimal set of comparisons?

Dr. George F. Corliss
Electrical and Computer Engineering
Marquette University
P.O. Box 1881
1515 W. Wisconsin Ave
Milwaukee WI 53201-1881 USA
414-288-6599; GasDay: 288-4400; Fax 288-5579
George.Corliss@xxxxxxxxxxxxx
www.eng.mu.edu/corlissg



 Prof. Svetoslav Markov, DSci, PhD

 Head, Dept.  "Biomathematics",               phone: +359-2-979-3704
 Inst. of Mathematics and Informatics,       fax: +359-2-971-3649
 Bulgarian Academy of Sciences,              e-mail: smarkov@xxxxxxxxxx
"Acad. G. Bonchev" st., block 8, BG-1113 Sofia, BULGARIA mobile (gsm): 0885871584

 URL: http://www.math.bas.bg/~bio/
------- End of forwarded message -------


--
Karlsruher Institut fu"r Technologie (KIT)
Institut fu"r Angewandte und Numerische Mathematik (IANM2)
D-76128 Karlsruhe, Germany
Prof. Ulrich Kulisch

Telefon: +49 721 608-2680
Fax: +49 721 608-6679
E-Mail: ulrich.kulisch@xxxxxxx
www.kit.edu
www.math.kit.edu/ianm2/~kulisch/

KIT - Universita"t des Landes Baden-Wu"rttemberg und nationales Gro?forschungszentrum in der Helmholtz-Gemeinschaft