Re: [xml-dev] Whither XML ?

[BillClare3@aol.com]

   This thread started out to be about evolving XML to be a language for application development. 

   Not particularly surprising though, much of what has been discussed here deals with problems with browser support and why that is largely the problem with standards adoption.   Also there is discussion of serious issues related to efficient data formats for processing, communication and exchange, particularly with mobile devices. 

   What I am advocating is fundamental changes to the language foundations for XML, which would incidentally ameliorate both of these issues.

   With regard to data formats, the fundamentals of a specification language need to deal directly with issues of effective data types and formats - in addition to UTF variants in a markup language.

   While not directed at browser issues, this approach would have the following consequences:

·         By simplifying specifications for XML standards, browser support for them becomes considerably easier.

·         Support for various browser extensions can avoid the need for new base function in the browser.

For instance, a new standard involves new elements and attributes, and often complex constraints on their use.

Schema allow new elements to be defined.  Properties and behavior for these elements can also be defined so that infosets essentially become objects.  Implementation as objects then requires only the ability to add appropriately fenced execution libraries that can be accessed by the browser through standard interfaces – create, get, put, compare, run, etc. 

End user output can be routed to a rendering engine, possibly implemented as a plug-in.

New standards also imply both syntax and semantic constraints. These can be addressed, in many cases with  fundamental flexibility in the base language, and in other cases with “rules’ statements.

Thus a new standard can be born with no change to the browser.

·         What is suggested here is primarily directed to application development.

The intent is to have a fundamental influence on how computer applications are developed.  This is driected more to developers, rather than to end users.   Hence the adoption path is different –implementation by a single browser is sufficient for application support.  Then, for other browsers - “if you build it they will come” – maybe ?

Platforms do compete based on their “app” libraries.

   This discussion highlights but two of the issues that an extended XML specification language can address. 

   But then, the questions remain:

·          Is a simpler and fundamentally more powerful language for models based application development a reasonable goal for the future of XML standards development ?

·         Is there energy in the community to address such a fundamental change?

 

In a message dated 11/6/2010 4:59:16 P.M. Eastern Daylight Time, BillClare3@aol.com writes:

  This note is probably on a somewhat different topic .  It deals broadly with issues for future XML standardization efforts.

   During the last decade there has been an amazing development of XML based standards and of the capabilities that they can support.  However many of these are in limbo, many are not supported, others are poorly coordinated, inconsistent, and overly complex.  For new capabilities the development and, more tellingly, the adoption cycle is long and complex.

    At one time, there was a considerable simplification made in going from SGML to HTML.  This simplification also allowed for a development of greatly expanded capabilities.  It seems that now is the time for an analogous effort in transforming XML with the objective of also leading to considerable simplification and expanded capabilities. 

   Clearly there are issues of syntax simplification which can be addressed.  More fundamental though are semantic issues.

   The first insight here is that XML standards generally support various models, including models for other models.  Fundamental generic models to be supported with broad and consistent capabilities are application models for data( physical and logical views, metadata, etc.), for presentation(HTML, Open office, SVG, CSS, etc.), for communication and for control (data and work flow, state machines, etc.).   As with the Model/View/Controller paradigm, these together can provide a complete, unified and consistent framework for application development.

    The second insight here that these models are based on a specification language, rather than a procedural language, and that, as a specification language there are considerable improvements that can be made to XML fundamentals to ease development and interactions among these and other models.  For instance it should be possible to deal with schema, style sheets, metadata, data content, etc. in a common and consistent way. 

    Basically then this is a recognition that what started as a “markup” language can serve well as a “models” language.

    The objective here then is to create a complete and consistent specification language that forms the basis for developers to create executable application frameworks based on generic models for data, presentation, communication and processing.  A secondary objective is to create a base whereby it is simpler to develop and adopt new standards.

    In that this is meant to be a somewhat revolutionary approach, a primary concern is compatibility.  This needs to be addressed; for instance, with alternative language processing and through extended and alternate “infosets” which can be supported by a common agent. 

   Much of this is outlined in more detailed documentation that is under development and can be made available.

   This note is to inquire if there is interest in the community in pursuing such a fundamental look at approaches for the further development of XML standards.