Mukul Gandhi a écrit :
> Please see: http://www.research.ibm.com/xj/
>
Thank you for this link.
Xj is indeed a very ambitious project but I'm afraid it's not anymore
active and it doesn't seem to have Sun approval...
I think the preprocessor approach is good when it can be completely
separated from the effective compilation.It can be done for static
internal XML document loading while xj goals need a much more integrated
approach...
Human brain can easily read sophisticated syntaxes but XML is human
readable only when there is a perfect indentation and it is a problem
with the string approach. JSON is more compact than XML and it's an
advantage when reading this notation. Function calls approach (DOM,
SAX-like, XLINQ,...) are, from my point of view, less human readable
than well-indented XML and could also be automatically generated by an
XSL transformation so why not use the XML notation instead !
Comparing Java versus Visual Basic, I would say Visual Basic is easier
to read because indentation is almost mandatory (only one instruction
per line) and no closing brakets ('}') but explicit end keywords (End
If, End Sub,...). For many reasons, I think source files should be
stored in full XML (even expressions such as "a+b") while intelligent
editors would display and allow to modify them using a syntax such as
the ones we know (the very old C syntax is still alive !) : it would
then be easy to have static internal XML documents embedded.
Alain COUTHURES
<agenceXML>
http://www.agencexml.com
