[dsdl-discuss] Montreal WG1 meeting notes

Dear all,

Here are Martin's notes from yesterdays WG1 meeting. I'm posting them as
Martin doesn't have access to the email account he uses with
dsdl-discuss ...

---- snip

The following issues need to be checked on Part 8 (also required for
Part 5):

1) Are all IRIs valid as namespace URIs? If not, what restrictions
should we specify for the use of IRIs as namespaces?

Comments on Part 7:

1) How will CVDL be invoked? In the introduction it says "Descriptions
in this language may be referenced from schemas". What schema languages
provide means of referencing character sets? How do they do this? The
introduction also states "Furthermore, they may also be referenced from
forms and stylesheets." How do you reference a character set from a form
(e.g. an XForm) or a CSS stylesheet?

2) Should clause 6 refer to the whole of the W3C Schema regular
expression language, or just to subclause F.1, which defines how
character classes are defined? What is the relationship between the
"Unicode regular expressions" referred to in clause 6 and XML Schema
character classes?

3) What the relationship between graphemes and grapheme clusters? For
example, how can you say that a character set may not include an e acute
character or an e with a combining acute, but may have other letters
combined with the combining acute code?

4) The definition of hull in Clause 5 is somewhat circular: it reads "A
hull gives an outer boundary so that characters or grapheme clusters
which are not in the hull are guaranteed not to be in the collection;"
We suggest that "not in the hull" be changed to "not defined within the
specified range". This would also require a change to the definition of
hull in 3.3. We think that it may be necessary to state explicitly that
a hull is a single range of characters, rather than a "set of characters
and grapheme characters". If a hull can be defined using any XML Schema
regular expression it could be defined as a set of single characters
plus a set of ranges of characters. Is this really what you had in mind
for a "hull"?

5) Need to clarify both notes at the end of clause 6: in the first one
there is the word See on its own at the end of the note, without a
qualifying reference; in the second it is unclear how a single LF could
ever match a Unicode expression. Please simplify these notes.

6) The reference to "RELAX NG schema in the compact syntax" at the start
of clause 7 should be replaced by a formal reference to the amendment
for Part 2.

7) The grammar of the first para of 8.7 needs adjusting. It probably
should read: "IANA charsets and other implementation-dependent encodings
are named collections. The semantics are implementation dependent." How
are "implementation-dependent encodings" assigned names?

8) What is meant by the term "decomposition" in the second paragraph of
clause 9. Is decomposition defined within the Unicode specification? If
so this should be explicitly stated.

9) In clause 6 the phrase "strings u and v matches e" should be adjusted
to "strings u and v match e".

10) In clause 8.2, Case 3, change both occurrences of "has" to "have"

11) Ensure consistency of use of periods after entry in lists. Sometimes
you have them on all items, sometimes on only the last entry in the
list. The use of periods should be consistent.

12) In the last para of clause 9 please change "but rather examine the
string directly" to "but rather examines the string directly."

13) An example instance would make it much easier to understand the
specification.

--
DSDL members discussion list
To unsubscribe, please send a message with the
command  "unsubscribe" to dsdl-discuss-request@dsdl.org
(mailto:dsdl-discuss-request@dsdl.org?Subject=unsubscribe)