Go backward to About this document
Go up to Top
Changes from version 0.99
- 0:
- New version number.
- 0:
- Was "The structure of this document is as follows:".
- 0:
- Version description.
- 0:
- Citation of Note M-4 added.
- 1:
- Clarification.
- 1.3:
- Clarification.
- 1.3:
- Strong equations are derivable from existential ones without using disjunction or negation.
- 1.4:
- Clarification.
- 2:
- Clarification.
- 2:
- Well-formedness of terms is not defined.
- 2:
- Clarification.
- 2:
- Clarification.
- 2.1.1.1:
- S for sorts in concrete syntax examples replaced throughout by s.
- 2.1.1.1:
- Unhelpful explanation removed.
- 2.1.2.1:
- ON for operation names in concrete syntax examples replaced throughout by f.
- 2.1.2.1:
- Clarification.
- 2.1.2.1:
- Clarification.
- 2.1.2.1:
- Clarification.
- 2.1.2.1:
- Clarification.
- 2.1.2.2:
- Clarification.
- 2.1.3.1:
- PN for predicate names in concrete syntax examples replaced throughout by p.
- 2.1.3.1:
- Typo.
- 2.1.3.2:
- Clarification.
- 2.1.4.1:
- A warning about selectors usually being partial has been added.
- 2.1.4.1:
- Further restrictions are imposed on datatype declarations, to avoid pathological examples that would undermine the relationship between free datatype declarations in basic specifications and the corresponding structured specifications.
- 2.1.4.2:
- Constructors in free datatype declarations are injective.
- 2.1.4.2:
- Selectors in free datatypes are as undefined as possible.
- 2.3.1:
- Clarification.
- 2.3.4.3:
- Typo.
- 2.3.4.3:
- The mixfix identifier `__ __' is regarded as an infix.
- 2.3.4.3:
- Clarification.
- 2.4:
- Clarification.
- 2.4:
- Clarification of the distinction between mixfix and ordinary identifiers.
- 2.4:
- Removal of an inconsistency with Appendix C.
- 2.4:
- Clarification.
- 4.1.2:
- Introductory comment added.
- 4.1.2.1:
- Clarification that unions of sorts in free datatype declarations are always disjoint unions.
- 4.1.2.1:
- Free datatypes must not affect the overloading relation of the local environment.
- 4.2.1:
- Well-formedness of terms is not defined.
- 5:
- Clarification.
- 5:
- Implicit views have been eliminated.
- 5:
- Morphisms determined by symbol maps between signatures are as much as possible identity.
- 6.1.5:
- Clarification.
- 6.2:
- Introductory remarks added, and sectioning adjusted.
- 6.2.1:
- Sentence moved to below.
- 6.2.1:
- Sentence moved from above.
- 6.2.1:
- Explanation of the rôle of imports.
- 6.2.2:
- Clarification.
- 6.2.2:
- The imports are added to the local environment for arguments.
- 6.2.2:
- Clarification.
- 6.2.2:
- Correction of " typo.
- 6.2.2:
- Clarification.
- 6.2.2:
- Push-out explained.
- 6.3:
- Introductory remarks changed, and sectioning adjusted.
- 6.3.1:
- A view definition is well-formed only if the signature morphism is uniquely determined.
- 6.3.1:
- Clarification of the scope of parameters in generic view definitions.
- 6.3.2:
- Clarification that views have to provide the required models.
- 6.3.2:
- Clarification that instantiated generic views have to provide the required models.
- 6.3.2:
- Implicit fitting views have been removed.
- 6.4.1:
- QUAL-ID, TYPE introduced, since a:s might be either a constant function or a unary predicate.
- 6.4.1:
- Clarification.
- 6.4.1:
- Symbol lists determine sets of qualified symbols.
- 6.4.2:
- Clarification.
- 6.4.2:
- Symbol maps determine maps between qualified symbols.
- 6.5:
- Subsort embeddings do not extend to compound identifiers.
- 6.5:
- Ensuring semantics gives push-outs when defined.
- 8:
- The concrete syntax of architectural specifications is no longer tentative.
- 8.1.1:
- Clarification of the types of implicit generic functions.
- 8.1.1:
- Explanation of unit specifications extending signatures of imported units.
- 8.2:
- A couple of changes to the abstract syntax of unit specifications, not affecting the way they are written: UNIT-SPEC-NAME is replaced by SPEC-NAME (since they cannot be distinguished context-freely); and an explicit `closed' construct is introduced (cf. closed structured specifications).
- 8.2:
- Implicit closing of defined unit specifications, and no longer any distinction between names of unit specifications and of ordinary specifications.
- 8.2.3:
- An explicit `closed' construct for unit specifications is introduced (cf. closed structured specifications).
- 8.3:
- Clarification.
- 8.3.1:
- Sharing is related to the overloading relation.
- 8.3.1.1:
- Renaming should not introduce new sharing.
- 8.3.1.5:
- Explanation of unit applications made more uniform with that of instantiations.
- 8.3.1.5:
- Each argument should fall into the domain of the generic unit.
- 9:
- Downloaded specifications are not necessarily from the latest version of the remote library.
- 10:
- The concrete syntax of libraries is no longer tentative.
- 10.2:
- A library identifier may be a path of file names.
- 10.2:
- Library identifiers determine library location.
- A:
- Changes in grammars are indicated by `!' at the left of each changed line.
- B:
- Changes in grammars are indicated by `!' at the left of each changed line.
- C:
- The reference to the original concrete syntax document has been removed. Also, the adjective `proposed' is no longer applied to the concrete syntax.
- C:
- Obsolete citation.
- C:
- Clarification.
- C.2:
- Changes in grammars are indicated by `!' at the left of each changed line.
- C.3:
- Corrected keywords.
- C.3:
- Disambiguation rules for architectural specifications added.
- C.4:
- Minor addition to the list of rejected ISO Latin-1 characters.
- C.4:
- Lexical syntax of URL is distinct from that of PATH.
- C.5.2.1:
- Label annotations may be allowed more generally in lists.
- D.3:
- Names are displayed with mixed upper and lower case when `small-caps' are not available.
- E:
- The introduction and the examples themselves have been revised.
- F:
- Updated with the progress that has been made since version 0.99 appeared.
CoFI
Document: CASL/Summary-v1.0 -- Version: 1.0 -- 22 October 1998.
Comments to cofi-language@brics.dk