[CoFI] CASL Summary v0.98 available

[pdmosses@brics.dk]

Here is a copy of some pages from version 0.98 of the CASL Summary,
now available in full on WWW and FTP at URLs:

  http://www.brics.dk/Projects/CoFI/Documents/CASL/Summary/
  ftp://ftp.brics.dk/Projects/CoFI/Documents/CASL/Summary/

The versions of App. C and D on WWW provide links for easy browsing of
the listed items.  (Kindly let me know if there are any problems with
links or servers.)

PLEASE REACT, especially to the ISSUES FOR DISCUSSION, and well in
advance of the DEADLINE SPECIFIED BELOW! - so that the Amsterdam
meeting of the Language Design Task Group will have a proper basis for
deciding on what adjustments to make to the design, before preparing
the final version.

Sorry that I was unable to produce this document any earlier...

Peter

----   --------------------------------------------
\  /  | Peter D Mosses         <pdmosses@brics.dk> |
CoFI  | Common Framework Initiative  - Coordinator |
/  \  | WWW URL: http://www.brics.dk/Projects/CoFI |
----   --------------------------------------------

                                    CASL
                 The Common Algebraic Specification Language
                                   Summary

Abstract

     This Summary is the basis for the Design Proposal [LD97b] for
     CASL, the Common Algebraic Specification Language, prepared by the
     Language Design Task Group of CoFI, the Common Framework
     Initiative. It gives the abstract syntax, and informally describes
     its intended semantics. It is accompanied by the Rationale for
     CoFI [CoF97], the Rationale for the Proposed Design of CASL
     [LD97c], a draft of the Formal Semantics of CASL [Sem97c], and
     proposals for Concrete Syntax, with Examples of how CASL
     specifications might look [KB97][VBC97][Mos97a].

     Version 0.97 of the CASL Design Proposal was submitted for
     approval to the sponsoring IFIP Working Group on Foundations of
     System Specification, WG 1.3. It received tentative approval,
     together with a referees' report recommending the reconsideration
     of some elements of the design [IFI97]; a response has already
     been made [LD97a].

     The present version of this Summary shows just which bits of CASL
     are currently subject to reconsideration or revision, in view of
     the referees' comments and the recommendations made by the CoFI
     Semantics Task Group [Sem97c]. Changes that were made since the
     previous version are highlighted in the same way as this sentence.
     Points where further discussion is required are highlighted like
     this sentence.

     All comments on this Summary should be sent to the CoFI Language
     Design mailing list(cofi-language@brics.dk).

     Deadline for comments to be considered at the Amsterdam Meeting of
     the Language Design Task Group:

          12 noon on Tuesday, 23 September 1997.
_________________________________________________________________________

Appendix C: Changes from the Proposed Design v0.97

Only the major changes are listed.

()
     CASL now stands for Common Algebraic Specification Language.
()
     The abstract and About This Document have been updated.
(footnotes 6, 7, 8, 9, 10)
     Some footnotes with supplementary details have been added.
()
     LOCAL-BASIC-SPEC has been removed.
(2.1)
     An ID should not be used simultaneously for constants and variables of
     the same sort.
(2.1.2)
     The syntax of FUN-DEFN has been adjusted.
(2.1.3)
     PRED-DEFN has been added.
(2.2)
     The same variable must not be declared with different sorts by basic
     items in the same BASIC-SPEC.
(3.1)
     Overloading relations are defined on qualified symbols. The definitions
     have been adjusted to make the relation symmetric.
(4.2)
     Implicit embeddings can be applied to well-sorted terms.
(6.1.2)
     The semantics of the conservative extension is undefined instead of
     inconsistent.
(6.2)
     GEN-SPEC-INST and FITTING-ARG have been adjusted to use SYMB-MAP*
     instead of SIG-MORPH?.
(6.2)
     When instantiating a generic specification, each model of an argument
     is required to be a model of the corresponding parameter, otherwise the
     instantiation is undefined.
(8.1)
     An ARCH-SPEC can be used as a UNIT-SPEC.
(10.2)
     The placing of the optional item name in an ITEM-NAME-MAP has been
     changed.
(10.2)
     The explanation of DOWNLOAD has been made less operational.
(10.2)
     An extra, more concise grammar has been added.

_________________________________________________________________________

Appendix D: Points for Discussion

(1.2, 2.2)
     Are empty carrier sets to be allowed or not? Are there any interesting
     examples where it makes a difference?
(2.1.4)
     The abstract syntax of a COMPONENT has been generalized to allow
     several selector functions with the same result sort to be specified
     together.
(4.1)
     The abstract syntax and semantics of ISO-DECL have been adjusted to be
     more uniform to those of SUBSORT-DECL. Are there any objections?
(footnote 11)
     Can the treatment of compound identifiers be adjusted to permit a fully
     institution-independent treatment of structuring constructs?
(footnote 12)
     What abstract syntax is to be used for making explicit which parts of
     the parameter signatures may share with the body of the generic
     specification, so as to obtain pushout semantics?
(5)
     Should naming of morphisms be allowed? See Note L-6 [Ber97], which also
     proposes changing the treatment of instantiation and sharing.
(8.2)
     What changes are needed to the abstract syntax for UNIT-TERM?

----------------------------------------------------------------------------
CoFI Document: CASL/Summary --Version 0.98-- 9 September 1997.