Re: [xml-dev] Re: ***SPAM*** [xml-dev] Re: The Goals of XML at 25, and t

[Arjun Ray <arayq2@gmail.com>]

 On Thu, 22 Jul 2021 13:04:19 +0200, Marcus Reichardt
<u123724@gmail.com> wrote:

| (technically, SGML has MSOCHAR, MSICHAR, MSSCHAR but these are not
| assigned in the default concrete syntax). Verbatim inclusion of 
| markup delimiters is performed through entity references (eg &lt;). 

Unfortunately, ISO 8879 screwed the pooch here, ensuring that this
escaping mechanism wouldn't be widely used.  Some fine print in the
standard winds up requiring the escaping characters to be echoed in
the parsed output!  So, e.g., if MSSCHAR were defined as '\', then the
sequence '\<' in the source data would have to be presented as '\<',
and not as '<', to the application. 

Could this be the subject of a useful override via SEEALSO? (I don't
know, as I have yet to grok SEEALSO.)  

| So isn't SGML/XML already doing what you're postulating ie avoiding 
| the need for arbitrary lookback

I think the case for parallelized processing (where, apparently, you
could start parsing anywhere) is over-stated.  It seems to implicitly
assume that the entirety of the source document is already available,
which is certainly not true in streaming contexts (even as simple as
data coming in over the network).  And it militates against modal
parsing, which imposes severe syntactic constraints.