Standards are made in such a way that when a standards compliant browser recieves a page that contains things it doesn’t understand (such as a new CSS property, or a new style sheet language entirely, or a new scripting language), it will ignore the parts that it doesn’t understand.—Slime
Within a particular Recommendation’s evolution, I can agree with this statement. The rules for evolving HTML were well understood. The rules for evolving XML is well understood. The rules for evolving CSS is well understood. Yep.
But an old standards-compliant browser that doesn’t understand XSLT will be unable to do anything with an XML file which was intended to be transformed. Indeed, the “content” rendered by untransformed XML will likely be unusable, even if the browser chooses to render it, because logical filtering, calculations, and tree restructuring will not have occured.
Likewise, that browser which doesn’t understand XML namespacing will not render an XHTML document correctly, should it also contain namespaced SVG, for example.
In such a case, different representations must be presented to the old browser on the basis of client-side or server-side content negotiation. Backwards compatibility of HTML 4.01 with HTML 2.0 is irrelavent in the situation I am describing.
The standards themselves are what are backwards compatible. For instance, XHTML was created in such a way that a browser that doesn’t understand XML will still recognize it as HTML 4.01.—Slime
Actually, this is incorrect for a couple of reasons.
First, XHTML 1.0 -can be- made to also conform to HTML 4.01, if the XHTML complies with restrictions made in the XHTML spec [XHTML].
Such a document can be served as either text/html or application/xhtml+xml. 
Any XHTML document which does not adhere to the restrictions made in the previous reference (for whatever reason) may cause problems with HTML browsers.
In this case, it is not valid to describe the content as text/html, and should be served as application/xhtml+xml [XHTMLMediaTypes]. In this case, a browser which does not understand the application/xhtml+xml MIME type will generally popup a Save As dialog, or some such catch-all behavior.
Slime, thanks for taking the time to respond, but your rebuttal only illustrates my point—people don’t seem to understand the purpose of MIME types, or the concept and necessity of content negotiation.