Sorry for jumping in ...
Marcus Bingenheimer さんは書きました:
> Dear all,
> I want to confirm Wendell's perceptive remarks
> I think this is interesting and highly arguable. I've actually found
> that while RNG is easy and fun for people to learn who have
> already mastered DTDs, to those new to the topics of document
> analysis and content modeling, especially those unpracticed in
> "going meta" on their documents, it's not really easier. On the
> contrary, for some, its elegance and smooth surface make it simply
> baffling, whereas the fairly concise but "scratchy" syntax of DTDs
> gives a bit more of a handle. RNG really comes into its own under
> maintenance and when things get complex, not at beginner level.
> This is the first semester where I teach a Relax NG module without
> having introduced DTDs before.
> It does indeed seem more difficult for the students, even if the
> beginnings are fairly straightforward. I am thinking of doing DTDs
> again next year simply for didactic reasons (and why not, if LISP is
> still taught at MIT with a similar rationale?)
> In any case, content modeling is at the core.
> Again, I agree wholeheartedly. Does anybody have a curriculum with
> graded exercises to introduce content modeling?
Not 100% an anwer to your question, but as a base for developing such a
curriculum still a useful book IMO and since a few months freely available:
Of course lot's of this is SGML specific, but many parts are pretty general.
> It really seems desirable to develop a course like this.
> all the best
> Dr. Marcus Bingenheimer 馬德偉
> Director, Library and Information Center 圖書資訊館館長
> Dharma Drum Buddhist College 法鼓佛教研修學院 (DDBC)
> No. 2-6 Xishihu, Jinshan 20842, Taipei County, Taiwan, R.O.C.
> 台灣，20842台北縣金山鄉西勢湖2-6號 Tel: +886-2-2498-7171 # 2381