With reference to yesterday's conference call [Re: notes on teleconference, 9 sep], and in particular about the discussion about macros, style sheets, etc: I had the sense at the end that we all thought it would be useful to come up with a definite spec for a simple macro facility roughly equivalent in power to lambda-expression-expansion (e.g. not necessarily allowing sensitivity to parameter-types in the pattern matching, and perhaps having only one "namespace" for macros without attempting to make some of them non-overridable), which together with a simple "annotation" facility for semantic info or rendering options, would be sufficient to allow people to establish conventions (sets of macro names and their intended meanings) for which authors could write their documents, so that they would render properly when used with various style sheets consisting of bindings (expansion rules) for the macros established by the convention. This simple facility would avoid many (though probably not all) of the complex questions and subtleties of macro semantics, while allowing what we all see as their most important use. The fact that sophisticated renderers could not be written directly in the macro language, and that style sheets using only the macro language could not be sophisticated enough to use heuristics to try to interpret the meaning of documents not authored for a recognized convention, might be an inconvenience, but would not prevent such utilities from being written some other way, since the information they need to work could (and usually would) still be present in the HTML-Math. So I suggest that we try to come up with a specific simple proposal for annotation and macro syntax to provide the above-described level of solution to these problems, and then spend more time on these issues only when these relate to proposed extensions to whatever simple, good-enough system we've agreed on (or if we have second thoughts about the adequacy of what we've agreed on). (I'll try to find time to propose one possible form of it in a few days.) - BruceReceived on Tuesday, 10 September 1996 16:10:22 UTC
This archive was generated by hypermail 2.4.0 : Saturday, 15 April 2023 17:19:57 UTC