John McKown wrote: <begin snippet> Neither is good "pseudo-xml", if that is what was intended <joke type="technical"> ... </joke> is at least syntactically correct XML. </end snippet> I write XML and the like on occasion, but I was not writing either XML or pseudo-XML on this occasion. Backus used broken brackets, '<', and '>', as delimiters in what came to be called BNF, as in <decimal digit> ::= 0|1|2|3|4|5|6|7|8|9 ^, in the 1960s, and I and others find it convenient in many definitional contexts: it permits a brief descriptive phrase to be used as the manipulable and largely self-explanatory name of an entity. None of this should be interpreted as an effort to delegitimate the use of broken brackets in XML. I can see no objection to that, but in a period of radical historical amnesia it is perhaps worth noting that notation thought by some to have been introduced with XML in fact has a long prior history.
John Gilmore
