Originally posted to http://lists.w3.org/Archives/Public/public-html/2008Oct/0254.html
There have been so many things written on this subject that I have rather become neglectful in my reading of it. At first the discussion all seemed very technical and “browsery”: the sorts of things that “user-agent implementers” like to discuss. And then, lo, I saw hints that the discussions were expanding into broader issues of multi-language support, Chicago manual of style, authorial intent, … . What fun! At the upper end of these discussions we have the basic philosophy of punctuation and questions such as “ what if any difference is there between the semantics of HTML and the semantics of punctuation?” (see also )
Generally if A quotes B who quotes C, we assume that C’s speech act predates B’s, but this need not always be true.
If C quotes a previous statement by A who was quoting a putative future thought by B where the time of the thought by B is, in fact, earlier than the ostensible reading of C’s statement, then certain languages like Navajo (with its incredibly complex sense of aspect) might be expected to encapsulate it all in a most eloquent set of grammatical devices. Should our theories of punctuation not be extended to handle temporal, probabilistic and epistemic aspects of verbs?
“A said that B would have realized by now that my typing of the sentence you are now reading was, in fact, unlikely.”
Suppose an utterance is expressed in a base language, let’s call it Punctuation ML, for which the primitives of the ML are necessary and sufficient for the punctuation of human thought (hence, a language more basic than HTML). Then to what extent might we hope for our PML to be transformable into either HTML or SVG (through, say, XSLT) as appropriate?
Contextually, longer quotes might be, in HTML (as augmented by styles) ,naturally translated into indented, differently typefaced segments. Within the context of the cartoon, rendered in SVG, the very same utterance might be drawn into a thought balloon emanating from a character's mouth. A fundamental substrate language addressing universals in linguistic punctuation might be, in some sense, a logical prior step to the development of HTML. Of course, we are, currently, where we are; so, unless our theory of time involves moving backwards, it may not be a practical consideration.
There are times that it seems that HTML is striving to define itself as that more fundamental markup language in which authorial intent is conveyed through a modernized and universalized theory of punctuation; there are times it seems that HTML is something quite different, though at the moment I’m not quite able to put my finger on what that might be.