Re: RE: James Clark: XML versus the Web

    And perhaps some further movement forward.

    To this relatively inexperienced viewer, this seems like going back to 2002 is an interesting way to get started, if the goal is a short specification.   

    However, it seems that fundamentally XML is a specification language, for a lot more than markup, and that many of the basic extensions over the last decade have made it, extremely powerful in dealing with data for many purposes.  The issue is that these extensions are complex and inconsistent.

    It seems that the basis for some of the problems are with the current language itself,  but many are with the additional standards and especially how they  interact.  Many of these problems with the additional standards, though, stem, not from themselves, but from deficiencies in the basic language capabilities.

    It would appear then, as noted in other posts on this topic, that the starting point should be on a set of fundamental language extensions, such as those that deal with data types, expressions and probably most importantly - modularization to simplify and standardize interactions among related specifications.  Development would proceed based with a clear focus on what is fundamental to the language, what is needed for application models, such as presentation, and what is needed for compatibility.  

  Syntax simplification would provide a sugar coating to encourage implementation and adaptation.

  As James Clark put it in his original post on this thread “the challenge is how to  . . . create technologies . . . that bring to the broader Web developer community some of the good aspects of the modern XML development experience”.  It would seem that, for the long term, this must require a considerable rebuilding based on architectural fundamentals, rather than just updating.  And this indeed seems much larger, and certainly more comprehensive, than the transition from SGML to XML.