Cart
[Home] [By Thread] [By Date] [Recent Entries]
Title: Message > Can anyone address the Why and include the
perspective of a parser requirements writer / standards committee
member?
In my view the problem arises
because the architectural layering in XML isn't strong enough. There are
different layers of protocol superimposed on each other and they aren't clearly
separated. The XML spec makes an attempt to define a physical level and a
logical level, but there's a single grammar that intertwines both, and there's
no clear separation that says applications should work at one level or the other
but not both. In consequence, there's a continuum from things the application
doesn't care about (like character encoding), to things that it just might care
about (character references), to things that it's quite likely to care about
(entity references with semantic significance). Every processing model that's
layered on top of XML has to make its own decisions about what details to expose
to the application. If you expose too much, you make life awkward for people who
don't care. If you expose too little, you make life awkward for those who
do.
Michael
Kay
http://www.saxonica.com/
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] |
