For those who didn't know this: one of the great things about XML is that it allows for the mixing of namespaces in a single document.
This means that different XML technologies (like XHTML, SVG, MathML, and yes, XForms) can be used in a single file.
Now if the display device (ie. browser) has support for all the used technologies, some funky things become possible.
For example, a web page could use XHTML for structuring a document, SVG for graphics, MathML for formula display and XForms for data input.
Throw in some scripting, and you could for example do a function plotter as a web application, in a single document.
At this point it's not just about structured data anymore, but also about mixing-and-matching technologies to create applications. And XForms is one of the building blocks that will hopefully make this possible.
As far as i can see, this would really be a nice application for a neural network. The only problem I see is that only 1000 turns are played, which is probably not enough to train a neural network into a winner.
Basically it's all just pattern recognition, and neural networks are good at that.
Slashdot Mirror
For those who didn't know this: one of the great things about XML is that it allows for the mixing of namespaces in a single document.
This means that different XML technologies (like XHTML, SVG, MathML, and yes, XForms) can be used in a single file.
Now if the display device (ie. browser) has support for all the used technologies, some funky things become possible.
For example, a web page could use XHTML for structuring a document, SVG for graphics, MathML for formula display and XForms for data input.
Throw in some scripting, and you could for example do a function plotter as a web application, in a single document.
At this point it's not just about structured data anymore, but also about mixing-and-matching technologies to create applications. And XForms is one of the building blocks that will hopefully make this possible.
As far as i can see, this would really be a nice application for a neural network. The only problem I see is that only 1000 turns are played, which is probably not enough to train a neural network into a winner.
Basically it's all just pattern recognition, and neural networks are good at that.
Anyone wanna give it a try?