0002183: Only simple identifiers allowed in solve-before constraint

(0005960)
Neil Korpusik (administrator)
2008-02-09 18:21

The proposal was unanimously approved by the sv-ec
in the conference call of February 4, 2008.

(0006167)
Neil Korpusik (administrator)
2008-03-08 19:15

The proposal was sent back to the SV-EC by the
Champions in the February 25th, 2008 conference call.

    Brad
       I don't know what "simply identify" means? Sounds like a
       "simple_identifier" to me, which is roughly what was already there.
       What's an example of an "expression" that can "simply identify"
       something without being a "simple_identifier"? Also, just an
       observation, not a strong objection, this is hardly the only place in
       the BNF that would have the content "expression { , expression }". See
       case_item, rs_case_item, and assignment pattern. So why the need for a
       special production for it here?
    Stu
       The new sentence "Each expression in an 'expression_list' must simply
       identify a random variable." is not well worded. What is meant by "must
       simply"? The word "must" should be "can", "may" or "shall" (I am
       guessing that the intent is "shall, but I am not sure). The word
       "simply" should be deleted or rephrased.

    Neil - both Stu and Brad have concerns about the same sentence.

(0006344)
Ray_Ryan (developer)
2008-03-24 14:45

In considering the champions feedback and the SV-EC discussion, I
thought about better ways to state restriction on expressions in a solve-before constraints (that still allowed)
- names of random 'variables' should be allowed
- rand fields of random variables should be allowed.
- bit selects of random varialbes shoulde be allowed.
- references to elements of a random array should be allowed.

Finally my conclusion is that we don't need special restrictions on the expression in a solve before constraint. This raises the question how
about the case:
    rand int x,y,z;
    ...
    constraint C {
        solve (x+y) before z;
    }
This should simply have the same semantics as
    rand int x,y,z;
    rand int t1;
    constraint C {
        t1 == x+y;
        solve t1 before x;
    }

Since this can be coded with the current LRM, allowing expressions in a solve-before constraint doesn't add any new solver capabilities - rather it is a simplification.

I have added a modified proposal (errata_2183d.htm) that allows expression in solve-before constaints and removes the prior resrictions.

(0006616)
Neil Korpusik (administrator)
2008-04-25 18:08

The proposal was unanimously approved by the Champions in the
April 24th, 2008 conference call.

(0006657)
Neil Korpusik (administrator)
2008-05-02 17:21

The proposal was unanimously approved by the Working Group
in the conference call of May 01, 2008.

(0007244)
Ray_Ryan (developer)
2008-07-21 11:31

Reviewed in Draft6

Issue History
Date Modified	Username	Field	Change
2007-11-12 08:16	Ray_Ryan	New Issue
2007-11-12 08:16	Ray_Ryan	Type	=> Errata
2007-11-12 19:12	shalom	Issue Monitored: shalom
2007-11-30 15:15	Ray_Ryan	Status	new => assigned
2007-11-30 15:15	Ray_Ryan	Assigned To	=> Ray_Ryan
2007-11-30 15:17	Ray_Ryan	File Added: errata_2183.htm
2007-12-17 09:54	Ray_Ryan	File Added: errata_2183b.htm
2007-12-17 09:55	Ray_Ryan	File Deleted: errata_2183b.htm
2007-12-17 09:55	Ray_Ryan	File Added: errata_2183b.htm
2008-02-04 11:34	Ray_Ryan	File Added: errata_2183c.htm
2008-02-04 11:38	mehdi_mohtashemi	File Added: 2183_mantis_c.pdf
2008-02-09 18:21	Neil Korpusik	Note Added: 0005960
2008-02-09 18:21	Neil Korpusik	Status	assigned => resolved
2008-02-09 18:21	Neil Korpusik	Resolution	open => fixed
2008-03-08 19:15	Neil Korpusik	Status	resolved => feedback
2008-03-08 19:15	Neil Korpusik	Resolution	fixed => reopened
2008-03-08 19:15	Neil Korpusik	Note Added: 0006167
2008-03-24 14:28	Ray_Ryan	File Added: errata_2183d.htm
2008-03-24 14:45	Ray_Ryan	Note Added: 0006344
2008-03-25 19:24	Ray_Ryan	File Added: errata_2183e.htm
2008-03-25 19:27	Ray_Ryan	File Deleted: errata_2183e.htm
2008-03-25 19:30	Ray_Ryan	File Added: errata_2183e.htm
2008-03-25 19:31	Ray_Ryan	File Deleted: errata_2183e.htm
2008-03-25 19:34	Ray_Ryan	File Added: errata_2183e.htm
2008-03-28 07:25	Ray_Ryan	File Added: errata_2183f.htm
2008-04-14 12:13	mehdi_mohtashemi	Status	feedback => resolved
2008-04-14 12:13	mehdi_mohtashemi	Resolution	reopened => fixed
2008-04-14 12:13	mehdi_mohtashemi	Additional Information Updated
2008-04-25 18:08	Neil Korpusik	Note Added: 0006616
2008-05-02 17:16	Neil Korpusik	Status	resolved => approved
2008-05-02 17:21	Neil Korpusik	Note Added: 0006657
2008-06-05 08:27	Stuart Sutherland	Status	approved => completed
2008-06-05 08:27	Stuart Sutherland	Fixed in Version	=> P1800-2008/D6
2008-07-21 11:31	Ray_Ryan	Status	completed => closed
2008-07-21 11:31	Ray_Ryan	Note Added: 0007244

Notes
(0005960) Neil Korpusik (administrator) 2008-02-09 18:21	The proposal was unanimously approved by the sv-ec in the conference call of February 4, 2008.

(0006167) Neil Korpusik (administrator) 2008-03-08 19:15	The proposal was sent back to the SV-EC by the Champions in the February 25th, 2008 conference call. Brad I don't know what "simply identify" means? Sounds like a "simple_identifier" to me, which is roughly what was already there. What's an example of an "expression" that can "simply identify" something without being a "simple_identifier"? Also, just an observation, not a strong objection, this is hardly the only place in the BNF that would have the content "expression { , expression }". See case_item, rs_case_item, and assignment pattern. So why the need for a special production for it here? Stu The new sentence "Each expression in an 'expression_list' must simply identify a random variable." is not well worded. What is meant by "must simply"? The word "must" should be "can", "may" or "shall" (I am guessing that the intent is "shall, but I am not sure). The word "simply" should be deleted or rephrased. Neil - both Stu and Brad have concerns about the same sentence.

(0006344) Ray_Ryan (developer) 2008-03-24 14:45	In considering the champions feedback and the SV-EC discussion, I thought about better ways to state restriction on expressions in a solve-before constraints (that still allowed) - names of random 'variables' should be allowed - rand fields of random variables should be allowed. - bit selects of random varialbes shoulde be allowed. - references to elements of a random array should be allowed. Finally my conclusion is that we don't need special restrictions on the expression in a solve before constraint. This raises the question how about the case: rand int x,y,z; ... constraint C { solve (x+y) before z; } This should simply have the same semantics as rand int x,y,z; rand int t1; constraint C { t1 == x+y; solve t1 before x; } Since this can be coded with the current LRM, allowing expressions in a solve-before constraint doesn't add any new solver capabilities - rather it is a simplification. I have added a modified proposal (errata_2183d.htm) that allows expression in solve-before constaints and removes the prior resrictions.

(0006616) Neil Korpusik (administrator) 2008-04-25 18:08	The proposal was unanimously approved by the Champions in the April 24th, 2008 conference call.

(0006657) Neil Korpusik (administrator) 2008-05-02 17:21	The proposal was unanimously approved by the Working Group in the conference call of May 01, 2008.

(0007244) Ray_Ryan (developer) 2008-07-21 11:31	Reviewed in Draft6

Relationships