Issue IGNORE-USE-TERMINOLOGY Writeup

Forum:		Public Review

Issue:		IGNORE-USE-TERMINOLOGY

References:	Barrett's public review comment #30

		IGNORE, IGNORABLE

Category:	CLARIFICATION

Edit history:	21 Dec 1992, Version 1 by Loosemore

Status:		Proposal VALUE-ONLY passed (7+1)-2 on letter ballot 93-302



Problem description:


  In the description of the IGNORE declaration, there are several

  occurances of "use", "used", "referred to", and "referenced", which

  are not well specified.  The question is, what constitutes a use of a

  binding.  Are only references for value considered to be "uses~, or

  should references as a place to modify also be considered "uses"?

  Implementations differ on this question, making it very difficult to

  use the IGNORE declaration in portable code.


  There are two proposals, VALUE-ONLY and VALUE-AND-ASSIGNMENT.



Proposal (IGNORE-USE-TERMINOLOGY:VALUE-ONLY):


  Change the description of the IGNORE and IGNORABLE declarations to

  use the glossary term "reference" throughout.


  Clarify that, for the purposes of IGNORE and IGNORABLE, a "reference"

  includes only references for value.  



Proposal (IGNORE-USE-TERMINOLOGY:VALUE-AND-ASSIGNMENT):

 

  Change the description of the IGNORE and IGNORABLE declarations to

  use the glossary term "reference" throughout.


  Clarify that, for the purposes of IGNORE and IGNORABLE, a "reference"

  includes both references for value and variable assignments.



Rationale:


  Either one of these proposals would address the problem.



Current practice:


  Unknown.



Cost to implementors:


  Probably small.



Cost to users:


  Probably small.



Aesthetics:


  Specifying this behavior is more aesthetic than leaving it unspecified.



Editorial impact:


  The changes are confined to the dictionary entry for IGNORE and

  IGNORABLE.



Discussion:


  Barrett has indicated he prefers proposal VALUE-ONLY, but Loosemore

  thinks that having variables that can be assigned to but not

  referenced are of questionable utility.


  While researching this issue, Loosemore noted that the provision

  about the behavior of IGNORE (and presumably IGNORABLE) from issue

  MACRO-DECLARATIONS had not been incorporated into draft 12.24.  Can

  we give the editor authority to remedy this?