- From: Ian Hickson <ian@hixie.ch>
- Date: Tue, 28 Dec 2004 13:53:03 +0000 (UTC)
On Tue, 28 Dec 2004, Henri Sivonen wrote: > On Dec 9, 2004, at 03:19, Ian Hickson wrote: > > On Mon, 6 Dec 2004, Henri Sivonen wrote: > > > > This is incorrect, U+FEFF is a valid character in its own right > > > > (albeit deprecated in favour of U+2060) and is only the BOM if > > > > found at the start of a string. > > > > > > Isn't it the point of deprecation that implementors may opt not to > > > support deprecated stuff? > > > > It depends (the word "deprecate" doesn't imply this on its inverse; > > you have to check with each instance of a deprecation to find what the > > case is). However, in many cases, as, I believe, with U+FEFF, > > deprecation is intended to discourage potentially harmful or confusing > > use, without breaking backwards compatibility with existing content > > (and thus still requiring it of implementations). > > According to http://www.unicode.org/faq/utf_bom.html#38 a data format or > protocol may choose to ignore the BOM in the middle of a string. HTML doesn't choose that, though, so that isn't relevant to us. (Thanks for the pointer, though.) > Anyway, I'm still uncomfortable with using a deprecated character that > has a very special other meaning as a magic marker in WF 2.0. I'm not overjoyed with it myself, but I haven't got any better ideas. The current system works quite well, and certainly works better than the "[]" prefix that I first suggested. Do you have a better idea? > > > (Still, limiting the choice of syntactic sugar in the submission > > > format has the feel of Appendix C to it.) > > > > There aren't really any restrictions beyond not breaking XML 1.0 rules > > and being compatible with both non-namespace-aware and namespace-aware > > parsers. Which, of the remaining restrictions, did you want to remove? > > In principle all restrictions on the choice of syntactic sugar, but I > guess it is now good enough for practical purposes. As far as I can tell the only restrictions now are about not using a prefix, which we need to have to handle non-namespaced servers easily. > > > > The spec doesn't say what the recipient is supposed to do with > > > > _recognised_ elements or attributes, either. What servers do is > > > > pretty much up to the servers, not much we can do about it. > > > > > > Attempts to influence the ad hoc works of the ViewSourceClan might > > > indeed be futile. However, what you write in the spec could > > > influence what the developers of server-side frameworks (eg. Struts) > > > do. > > > > It could, I guess. Did you have any particular ideas in mind? I don't > > really know what to add. > > I had a general "must ignore unknowns" rule in mind. For unknown > elements, treat as if the start and end tags were not there. After this, > ignore text children of the submission element. Ignore unknown PIs. I've added a note that gives a suggestion on how to handle unexpected content. Let me know if you think that's ok. I don't think making it normative makes any sense; it's not like we can do anything to ensure compliance anyway, so making it law would only serve to create a generation of law-breakers (as it were). -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 28 December 2004 05:53:03 UTC