Exploring the alternative Syntax of XML-reading and thinking

Http://www-900.ibm.com/developerWorks/cn/xml/x-syntax/index.shtml

Great article! It involves issues that I have been thinking about, how to simplify XML editing (it clearly points out one of the root causes of XML editing difficulties I don't actually realize:The most commercialized XML editor also needs to be positioned and clicked to a certain extent, making it a way to write content quickly and freely.).

Like wiki, more concise hypertext language is also the starting point of my early naxt project attempt. In fact, in the middle of the term, I realized that even if I strictly split the module/level, to maintain a consistent set of features and syntax, I needed to limit the scope of the application more clearly. So now I tend to define naxt as an equivalent compact Syntax of xhtml2 (or its subset. This avoids excessive semantic entanglement.

In my attempt at the naxt project, I gradually switched to compiling a common txt2xml grammar parser and converter. BTW, while downgrading XML to a simple text format in a specific field, another line I am concerned with is the general expression of a specific XML editor (instead of writing a Special Editor for each dialect in a heavyweight language ), for example, I have great expectations for using XForms as the Interface Definition Language for general XML editing.

Continue, the above mentioned many designs worth reference (such as very successful Relax NG and its compact syntax), especially commit TXT. In fact, from the first time I came into contact with resin's stylescript, I also considered XSLT's compact syntax for a long time. In this way, I can learn from the following: XSLT itself, XSLT-lite (stylescript), xsltxt, XQuery, jxpath, and gpath. Of course, I still have major design issues. I have not considered whether it is a direct XSLT correspondence, or whether it can be closer to an independent template language (for example, it is allowed to destroy well-formed, even beyond declarative side effects )? Despite the emergence of general and convenient agile languages such as groovy, I think it is still necessary to consider the XSLT scheme because it is standard, platform neutral, widely accepted, has the maximum degree of interoperability (so you need to be cautious with such extensions as side effects), and has a unique and powerful "automatic template application ". The general template technology does not have this abstract mechanism.

The conclusion of the article raises some important questions, which are the troubles and principles I have realized or have not yet realized.

...... The disadvantage of the compression syntax is the lossInteroperabilityAnd can it surviveTime Test. Most of the results come from independent third parties. Not completely clearWhat kind of support is provided for the character encoding of different kinds and the rarely used XML components (such as processing commands )?. In addition, in most casesThere is only one founder, so his idea may gradually disappear.

...... Do not ignore XML 1.0InteroperabilityAnd its impact on the network.The main value of XMLIt is also the main reason for us to take this step.

...... In addition to common XML substitution syntax,The non-XML syntax specific to the application seems to have greater value., Especially in places where a large amount of content needs to be created ....... However, there isReasonable trade-offs. Relax NG Compact is a good example to illustrate that non-XML syntax can indeed reveal the secrets of the underlying concepts and data structures of a language ......

Ease of content creation remains the most intense debate I have ever seen about using alternative XML Representation. Editor development seems to be one of the most difficult problems in computing, and is rarely helpful.

After reading and thinking, I should first rewrite a general txt2xml tool. Before that, however, we need to define a language that represents the text grammar. I have already considered many similar tools, especially rules defined by XML syntax, such as XI. But now I guess RNG can be directly used to some extent?

The road is still long ......