OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

relax-ng message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Subject: [relax-ng] Minutes RELAX NG TC 2002-04-11

Minutes for a RELAX NG TC telecon meeting held 11 April 2002 at 9:00 am EDT
(UTC -04:00)



Not Attending

Kohsuke (called in briefly but asked to be excused)


1. Local namespace declaration for the compact syntax

We agreed that we ought to do this, but we will decide on a syntax later.

2. Character escapes \x{n}

James: Should we allow \{n} anywhere?
John: I think we should allow it everywhere.
James: What would be the disadvantages of not allowing it everywhere?
John: People care what they get...it simply is a representation of the
James: XML has been criticized for not allowing character references

The general consensus was to allow escaped characters everywhere, but this
is not the final vote, as attendance was light at this meeting.

3. Long string literals

James: Sometimes regular expressions can be very long things...
John: We need to allow them to be commented.
James: Should we have a concatenation character [such as +] or an adjacent
double quote delimiter?
Mike: I like adjacent or "mere" concatenation. I think it's simpler.
John: I like adjacent strings...it's mere lexical machinery.
James: C++ and Python use it.
Josh: Is adjacency harder to implement?
James: There is one ambiguity that we would need a solution for...we can't
allow it everywhere.
John: Do we need to delimit ns declarations?
James: I think we do all or nothing.
John: We can allow views of URIs?
James: I think URIs should be only one string.
John: This may be ugly...
James: We can express it in the grammar, use some expression in the grammar.
John: C tokenizer tokens and strings concatenate themselves.
James: Should we have comments between strings?
John: And make URIs single string only?
James: You can use an annotation for this.
John: Always []? And use multiple strings only in params?
James: It's not that simple. Double and single quotes don't appear in a URI.
John: You can use percent escapes in URIs.

General consensus was to resolve this issue after more discussion (via

4. Allow literal newlines in annotations?

John: I'm in favor of this.
James: Should we use a special syntax or a newline token?
Mike: It seemed odd to me [to allow literal newlines], but I capitulate.
James: Yes well Java and JavaScript allow literal newlines.
John: I'm unsettled; I'm still conflicted on this.
James: Let's take some time to consider this.

Literal newlines in strings has not been decided.

5. Allow an encoding declaration?

James: We should allow UTF-8 and UTF-16 differentiated by a BOM [Byte Order
Mark]. [When BOM appears as first character in the file, it signals to the
processor that the file is UTF-16, else you get UTF-8.]
John: I don't like the arbitrary encoding implications.
Mike: I think if we do !rnc encoding="UTF-8" [which would produce <?xml
encoding="UTF-8">] we may as well do !xml encoding="UTF-8" and give them
control over the XML declaration, rather than add a new construct.
John: I don't think this is necessary.
James: I think this would be confusing.
John: It's a problem to implement...are we looking at character strings or
XML blobs?
James: I don't think that is our problem.
John: I am reneging my proposal that we have an encoding declaration.
James: Let's not decide now.
John: Let's give it two weeks.

6. How shall we handle divs and annotations?

John: # is a lexical construct, ## is an annotation...let's not make things
excessively complex.
James: I hear you. Right now ## does not work well with groups of divs [as
in ## annotate [] {...}]...what about multiple stars?
John: I think this would be painful.
James: ## [alone] works well for <a:documentation>
John: The trick is to make it work well.
James: We could put the { on the same line as ##.
John: We could concatenate multiple lines of ##.
James: That seems OK. But with divs its ugly.
John: Should we allow the div to persist until the next div?
James: That may work better.
John: Give the div a name?
James: What about div1, div2, div3, etc.
John: This doesn't seem so bad, especially if there is no whitespace and the
nest limit is 9.
James: How about a div key other than braces?

This will be discussed further through e-mail.

Note: John officially asked for regular TC membership.
[I believe that according to the rules of TC membership
(http://www.oasis-open.org/committees/process.shtml#sec1e) that because John
asked to participate in the TC on 27 Feb, he could have been granted
membership after the third meeting he attended on 28 March.]

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]

Powered by eList eXpress LLC