David Carlisle wrote -- > I'll get ticked off by Chris for being seriously off topic but... Tick, tick!! My feeling is that yes, the details of *ML sybtax are not relevant, but the topic of "authoring structured documents" is pertinent. I think, judging by his later remarks, that perhaps Sebastian misled me when he wrote: > millions and millions of > people write HTML happily every day, I though he meant "by hand" but I susepct he means they use authoting tools that produce HTML. A very diferent meaning of "write" as well as of "is"?? > a tiny proportion write > LaTeX. what does that suggest? 1. That a lot more applications van be used as authoring tools for HTML. 2. That nearly all documents at present prepared for wweb publication can be adequately desribed in HTML. Of ocurse, there is a dinosaur-and-egg set-up here: people may really want to produce large quantities of easily browsable, well-designed, complex documents for we publishing but cannot do so whilst they are constarined by many factors to using HTML. This is pertinent since many of the detailed contributions seem to me to imply that people both do and will and should produce LaTeX by hand-coding. I do not think that they should; I am sure that they no longer need to do so. Thus I feel that whatever document-level syntaxes(sp??) a future version of LaTeX will read, they should be designed to clearly represent the full complexity of the documant and its structure, not to be easily hand-codable. Of course, backward-compatibility needs to be sensible dealt with but we do not need to be compatible at the authoring-tools level since we can assume that old authoring methods will never need to be used again. Maybe a controverial paragraph, but something that needs discussion and resolution. I hope that Frank made clear the imprtance of your "homework": (my trouble is that, as a teacher, the "importance of homewrok" has never been questionable:-) and that you can see that it is relevant, at least somewhat, to the above. chris