> From: "Rosen, Brian" <Brian.Rosen@marconi.com> > I think this was "and html" not "or html", so I think that it is > easier to have one additional format and see how it goes, rather > than two (or three, or four) > > On of the advantages of xml is that it marks up things like references and > authors with the function, rather than the appearance. You can > much more easily generate html from xml than the other way around. > Improved formatting is good, but improved cross references/author > tracking/.... is also good. With xml, we can get both, albeit > somewhat indirectly. I think that for a small, simple step, xml > is a better choice. The same things were said the last half dozen times this issue came up. After about the second or third round in about 1989, PostScript was officially sanctioned. In some ways that went worse than opponents predicted, but in other ways better. The positive view is that PostScript for RFCs died. Since then we've had at least 2 or 3 rounds of "HTML is better" followed by at least two rounds not counting this one for XML. It's one thing to advocate PostScript, MS Word, XML, HTML, nroff, or whatever you like for the I-D submission format. This morning some people seemed to be saying that XML should only replace nroff. I don't see the point, but then I use vi and emacs to generate nroff. If the Editor will tolerate XML, or any of a zillion other markup or fancy document formats (I designed and implemented one 20 years ago that saw significant commercial exposure), no one has standing to complain. However, by this afternoon, the consensus among reformers seems to be replacing ASCII with XML for the normative documents. Also as I predicted this morning, there is no consensus on the DTD except that it will be small, simple, wonderful for generating HTML, links, etc., and different from Marshall's. The parade of wonderful, simple, easy to use, powerful, user friendly markup tools has already begun. The talk of submitting I-Ds through a validator sounds nice, but unimpressive to anyone with real world experience with HTML validators. I write very simple HTML with vi and emacs and use the X3C validator enough to worry that they might cut me off. Still, the fact that the X3C validator is no subsitute for testing your HTML against browsers demonstrates that correctness provers are as far from the perfection claimed in advertising for markup languages as for programming languages and protocols. It is just plain ***WRONG!*** to even start to consider anything but ASCII for the official documents. As hard as it is for the unscared to believe, even XML will fade away completely and be replaced by something else even more wonderful, user friendly, easier for convicted monopolies to embrace and extend, and so forth. When that happens, there will be a new calls for "reform." To put it another way, no one who is the least uncomfortable with the existing PostScript RFCs has any standing to advocate XML for official documents. That Communism or replacing ASCII RFCs didn't work before does not support the position that this time we'll get it right. Vernon Schryver vjs@rhyolite.com