tag:blogger.com,1999:blog-6111725723134762877.post3101970692650789668..comments2022-11-16T05:32:23.171-08:00Comments on XMLSH: XML representation of xmlsh scriptAnonymoushttp://www.blogger.com/profile/06155161084044858311noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-6111725723134762877.post-35758042227497865132009-08-24T15:45:27.363-07:002009-08-24T15:45:27.363-07:00Makes perfect sense. You can actually do that tod...Makes perfect sense. You can actually do that today but not with XML tools. The xproc2xmlsh tool I've been working on is exactly that (takes xproc xml input as data, uses xmlsh to produce xmlsh output)<br /><br />But yes I could see how having an XML syntax could be useful. But would anyone use it ? Wont know until I try it :) Is anyone using XQueryX ? <br />Do the authors think it was worth the effort ? I don't know.<br /><br />Thanks for the comments !Anonymoushttps://www.blogger.com/profile/06155161084044858311noreply@blogger.comtag:blogger.com,1999:blog-6111725723134762877.post-88005530810362845332009-08-24T11:57:49.975-07:002009-08-24T11:57:49.975-07:00Dave Pawson beat me to it. Authoring in the RNC no...Dave Pawson beat me to it. Authoring in the RNC non-XML format is much better for a human, but having the ability to convert to XML opens up the whole XML ecosystems of tools.<br /><br />Ant uses an XML syntax, but it's not great to author, and a JavaScript-like or Python-like syntax would be better to author the script in, and as long as they translate with full fidelity to XML (with proper error lines conversions).<br /><br />Just like in XML, what really matters is the *info-set* and the actual format used is relevant to the parser/tool only, not human being.<br /><br />The advantage of having an XML representation of the "info-set" is to:<br />1) contains lots of meta-data, similar to how the AST of a C++ source file tells you a lot more about the code that what just looking at the source tell you (calls to implicit conversions, type coercion, etc...)<br />2) ability to transform / validate using XML tools.<br /><br />Another analogy is with Lisp, with the code and the data using the same syntax, allowing to write code (macros) that "write" code.<br /><br />If XMLSH had an XML syntax, you could have XMLSH "code" that generates XMLSH "code" based on XML "data" that furthers processes more XML "data".<br /><br />Hopefully that makes sense ;-)--DDhttps://www.blogger.com/profile/03097117701435249366noreply@blogger.comtag:blogger.com,1999:blog-6111725723134762877.post-72318164099829623162009-08-23T12:24:36.808-07:002009-08-23T12:24:36.808-07:00Relax NG is a good example. Has both an XML format...Relax NG is a good example. Has both an XML format<br />and a 'simpler' format?<br /><br />The XML format may be generated by such as XSLT,<br />the 'other' format is easier for humans.<br /><br />Both are useful. JC even provided a tool to transform one into another.<br /><br />So yes, I guess the XML format is useful!<br /><br />HTH DavePDave Pawsonhttp://www.dpawson.co.uknoreply@blogger.com