- From: Michael Sokolov <sokolov@falutin.net>
- Date: Tue, 24 Jul 2012 08:04:38 -0400
- To: David Carlisle <davidc@nag.co.uk>
- Cc: public-microxml@w3.org
On 7/24/2012 7:47 AM, David Carlisle wrote:
> On 24/07/2012 12:25, Michael Sokolov wrote:
>> A design goal for me would be a 1-1 mapping between the data model
>> and the serialization form. Is that possible?
>
> Would you make attribute order and/or white space in tags significant
> then? That would be quite a departure from xml/html.
I think it's worth considering and complements many of the other goals,
mostly leading to greater simplification. The downsides are some
additional complexity in the data model ("ignorable" white space,
attribute order - I hadn't thought of that one - hmm; Regarding the
single-quote/double-quote issue - I think if we go this route, it would
be best to eliminate the use of single quotes as an attribute value
delimiter.). But the upsides are considerable, especially for
implementers. A whole class of problems becomes solvable by reference
to serialized forms. Test suites like XQTS would obviously be simpler
to define. One use case I have in mind is highlighting and annotation -
anything involving character range data that cuts across element
boundaries is currently difficult to work with.
-Mike
Received on Tuesday, 24 July 2012 12:05:38 UTC