If you have noticed the MetaWrap test server has been up and down for last few weeks. Its because I have been playing with the Apache module. Finally got the framework compiling again under Linux. Win32 Apache was easy. Getting the new version in Linux to link with the MetaWrap libraries took some effort, but its dev stub is back up and running.

I am currently static linking. Going to experiment with dynamic linking.

eg:

http://test.metawrap.com/mw/@server/$html/System.Version.MIME/x=1/23?p1&p2=45

I’ll be leaving the world of JavaScript for a few weeks to re-enter the world of C/C++ so that I can get the Apache plug-in to a stage where I am serving test objects that the JavaScript can consume.

A Meaningless Diagram. Needs More Work I think.

I’m trying to figure out if I should encode the format in the end of the object or not.

so

http://test.metawrap.com/mw/@server/$html/System.Version.MIME/x=1/23?p1&p2=45

http://test.metawrap.com/mw/$xhtml/System.Version

vs

http://test.metawrap.com/mw/System.Version.xhtml

Although the latter looks cleaner, it will mean that the common extensions will become reserved words. I’m not sure if I want to place the restriction that no object can be called xml,html,xhtml,wap,flash,js etc…

I could still have defaults so that is the default for System.Version is xml, then

http://test.metawrap.com/mw/System.Version

would return the XML version.

as would

http://test.metawrap.com/mw/System.Version.xml

One possible solution would be to use the format

http://test.metawrap.com/mw/Object/Object/Object.xml

eg.

http://test.metawrap.com/mw/System/Version.xml

Which would make the Object.Mime unambiguous.