[Mirrored from: http://www.ornl.gov/sgml/wg8/document/1883.htm]

WG8 N1883


Document Processing and Related Communication --

Document Description and Processing Languages

TITLE: Proposal for TC for ISO/IEC 10179
SOURCE: DSSSL Defect Editor
ACTION: For information
DATE: 11 November 1996
DISTRIBUTION: WG8 and Liaisons
REPLY TO: Dr. James David Mason
(ISO/IEC JTC1/SC18/WG8 Convenor)
Lockheed Martin Energy Systems
Information Management Services
1060 Commerce Park, M.S. 6480
Oak Ridge, TN 37831-6480 U.S.A.
Telephone: +1 423 574-6973
Facsimile: +1 423 574-0004
Network: masonjd@ornl.gov

Technical Corrigendum for ISO/IEC 10179: DSSSL

This document lists currently reported defects in ISO/IEC 10179:1996. It has been compiled by James Clark, the Defects Editor for the standard.

Defects in Clause 7: DSSSL Document Architecture

This must be aligned with AFDR in the HyTime TC.

An external-specification element can count as the first specification for the purposes of an external-specification element without a specid attribute.

Defects in Clause 8: Expression language

Delete the phrase ", or assignment of".
In the example, the expression (lambda x x) should be (lambda (#!rest x) x).
Clarify that redefinition of units works like normal redefinitions.
The description of inexact->exact is unhelpful given that any quantity with non-zero dimension is inexact.
It is not clear whether string->number operates on numbers with units in them. Production 78 defines number to allow units, but the description of string->number says it returns a number (not a quantity).
There's a superfluous space following the backslash in sentence discussing #\space.
Use of character-property-declaration and added-char-properties-declaration should require the charset feature.
Clarify treatment of redeclaration of character properties.
define-language and declare-default-language should not be part of the core expression language.
format-number and format-number-list are supposed to be in the core expression language.

Defects in Clause 9: Groves

The classnm property needs a cn="#all" attribute.
The public identifier for the property set DTD is missing a slash.
On syntax/slitasns the acnmprop attribute should have the value resname not refname.
On doctype/elemtps the acnmprop attribute should have the value "gi stem" not "gi rankstem".
entdcl/entity should have a when clause that makes it null when the entity is declaring a default entity.
On element/simplelk the acnmprop attribute should be linktype not linkset.
The subdcsds psmodule should depend on the subdcabs not the subdabs module.
In the lexdef for GI, the provide should be doctype not dtd.
The genents property of doctype should include those default entities that were referenced in the DTD (eg in CDATA attribute values).

Defects in Clause 10: Query language

A symbol should be allowed as a member of a list representing the grovepos.
Many of the definitions call siblings when they should call rsiblings.
Clarify behaviour of hierarchical-number, when the node does not have all the ancestors listed in the first argument.

Defects in Clause 12: Style language

Use of the current-node procedure in a construct-expression should not require the query feature.
The default construction rule for an ssep node should be (empty-sosofo) not (make character).
Clarify what happens if next-match is called when the current processing mode is different from the processing mode of the current match.
Use of the application-char-characteristic+property-declaration should require the charset feature.
Clarify treatment of redeclarations of properties declared with application-char-characteristic+property-declaration and character-property-declaration.
char-script-case should be optional (maybe dependent on charset feature).
Use of the procedures defined in this clause within the left/right/center-header/footer characteristics of simple-page-sequence should be handled like uses within a generate-specification.
There should be a (length-spec? obj>) procedure.
The prototype (/ x length-spec) should be removed.
In the description of the display-space procedure, the conditional?: keyword argument is incorrectly referred to as conditional:.
A length-spec should be allowed for any characteristic with a display-space type (just as a length is allowed for any characteristic with a length-spec type).
A length-spec should be allowed for any characteristic with amn inline-space type (just as a length is allowed for any characteristic with a length-spec type).
For entity-address and idref-address clarify that the entity or element addressed is in the same grove as the current node. A node identifying the grove should be allowed as an optional second argument.
The span characteristic is missing from display-group.
The specified initial value for asis-wrap-indent: is not an allowed value for the characteristic. The initial value should be 0pt.
The expand-tabs?: characteristic should be called expand-tabs: (without the question-mark). The convention is that only boolean-valued characteristics have named ending in a question mark.
Clarify interaction between expand-tabs, input-whitespace-treatment, and lines.
The initial value of last-line-justify-limit: should be 0pt rather than 0.
The line-number-side: characteristic lacks an initial value. start is the obvious choice.
The line-number-sep: characteristic lacks an initial value.
Specify that the value of the country: characteristic is the 2 letter ISO 3166 country code (as used in Internet addresses) not the 3 letter one.
Clarify meaning of space-after specified on paragraph-break.
The sideline-side: characteristic lacks an initial value. start is the obvious choice.
The sideline-sep: characteristic lacks an initial value.
Clarify that these properties cannot be redeclared.
math should not be allowed as a value for the math-font-posture character property.
inline-space-space lacks an initial value. The initial value should be the symbol #f meaning the value is whatever the font metrics say it should be (as with min-post-line-spacing)
The type of the length: characteristic should be length-spec not length.
The length: characteristic should be required to be specified for an inlined rule in the same circumstances as it is required for a leader.
12.6.14 (and others)
The display-alignment: should allow values of page-inside, page-outside, spread-inside, spread-outside as the quadding characteristic of paragraph does instead of inside and outside.
It doesn't make sense to allow #f as the value of entity-system-id.
Description of initial value of fraction-bar: characteristic is misleading. It should say that the initial value is a rule with all inherited characteristics inherited.
When the table-auto-width feature is not present, it should be permitted not to specify the width for up to one of the columns.
When there is no table-cell flow object applicable to a given geometric cell, then a table-cell flow object with no characteristics specified must be implied for that cell.
If there are conflicting borders of equal priority one from a table-border and one from a cell-border, the one from the table-border should win.
If there are conflicting borders of equal priority, one with border-present?: #t and one with border-present?: #f, the one with border-present?: #t should win.
It should be clarified that the display-size of the scroll is the width of the window and the display-size for the content of the principal port is the display-size of the scroll less the size of the margins.