Options:
- # Session Start: Sat May 26 00:00:00 2007
- # Session Ident: #html-wg
- # [00:00] <hsivonen> nn
- # [00:00] <sbuluf> web apps
- # [00:01] <Philip`> [...]just a collection of technologies that lie behind web applications
- # [00:03] <mjs> it's not the collection though, the collection is supposed to be the things that will use this hypothetical future backplane (I think):
- # [00:03] <mjs> "The common lifecycle allows for application fragments (potentially still built with different technologies) to be composed more readily by plugging into a shared backplane of behavior."
- # [00:03] <mjs> of course, given that this document does not even appear to be aware of basic HTML (see the figure labelled "HTML submit") I don't know how seriously it needs to be taken
- # [00:05] <Philip`> Hmm, they appear to have mixed up HTML and VoiceXML
- # [00:05] * Philip` never knew that VoiceXML had its own forms language
- # [00:07] <mjs> their AJAX example is also hilariously broken
- # [00:08] <Philip`> It's almost as if they wanted to show it in a bad light
- # [00:10] <Philip`> They should have written it as "var req = new XMLHttpRequest(url); req.onload = function() { ... }" because that's almost as well implemented as xf:submission in most UAs
- # [00:11] <Dashiva> Meaning it's not? :)
- # [00:12] <Philip`> Indeed :-)
- # [00:12] <Philip`> though I assume it could be implemented trivially, which is a more convenient strategy than implementing all of XForms
- # [00:15] <Zeros> That's a really confusing spec
- # [00:16] <mjs> Zeros: it's not a spec I don't think, although I am not sure what it is exactly
- # [00:16] <Dashiva> Something like this seems a lot easier than xforms: function EasyXMLHttpRequest(url,handler) { var x = new XMLHttpRequest(); x.open('GET',url,true); x.onreadystatechange = function() { if (x.readyState == 4) handler(this); }; x.send(); }
- # [00:17] <mjs> I think the XForms / XHTML2 people missed the boat on web apps and are trying to figure out how they can associate their stuff on currently trendy buzzwords
- # [00:17] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [00:17] <Zeros> "We hope you take this paper in the spirit with which it was intended..."
- # [00:17] <mjs> the problem is that modern web apps are becoming less server-oriented, so the idea of having something like a "submission object" itself sounds anachronistic
- # [00:17] <Zeros> Clearly I haven't a clue what spirit that document was intended to have.
- # [00:19] <mjs> whisky, I would guess
- # [00:20] <Philip`> It'd be nice if you could submit a form directly over a P2P network to another client
- # [00:21] <Dashiva> Web 3.0 will be p2p?
- # [00:22] <Zeros> Philip`, that opens all kinds of security issues though :/
- # [00:23] * Joins: gavin_ (gavin@74.103.208.221)
- # [00:24] <Philip`> Just make sure it opens enough security issues that potential attackers will be overwhelmed by the choices, so they'll give up trying to decide and will go away and do something else
- # [00:25] <Philip`> But if there was something that would let me add multiplayer to Canvex without relying on a server that will collapse when a few dozen people start playing at once, I'd quite appreciate that :-)
- # [00:26] <Dashiva> I propose fish school security be applied
- # [00:26] * Quits: Sander (svl@71.57.109.108) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
- # [00:26] <Dashiva> With enough targets, the chance of it being you is low
- # [00:28] <Hixie> hsivonen: i don't see what benefit this gives us. (what i meant by the BOM not being a BOM is that if the encoding is different than the BOM's, then you've got some other character that would never be treated as U+FEFF)
- # [00:32] * Quits: kingryan (rking3@208.66.64.47) (Quit: kingryan)
- # [01:02] * Quits: heycam (cam@203.214.56.87) (Ping timeout)
- # [01:18] * Joins: heycam (cam@203.214.6.6)
- # [01:20] * Joins: Sander (svl@71.57.109.108)
- # [01:20] * tH gets déjà vu from Schalk Neethling's last post
- # [01:22] * Quits: billmason (billmason@69.30.57.156) (Ping timeout)
- # [01:25] * Quits: Sander (svl@71.57.109.108) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
- # [01:41] * Parts: hasather (hasather@81.235.209.174)
- # [01:54] * Quits: zcorpan_ (zcorpan@84.216.41.249) (Ping timeout)
- # [02:02] * Parts: hyatt (hyatt@24.6.91.161)
- # [02:23] * Quits: tH (Rob@87.102.84.23) (Quit: ChatZilla 0.9.78.1-rdmsoft [XULRunner 1.8.0.9/2006120508])
- # [02:25] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [02:30] * Joins: gavin_ (gavin@74.103.208.221)
- # [02:32] * Quits: h3h (bfults@66.162.32.234) (Quit: |)
- # [02:36] * Joins: asbjornu (asbjorn@84.48.116.134)
- # [02:48] * Joins: zcorpan_ (zcorpan@84.216.41.7)
- # [03:14] * Quits: dbaron (dbaron@71.198.189.81) (Quit: 8403864 bytes have been tenured, next gc will be global.)
- # [03:38] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
- # [03:43] * Quits: Zeros (Zeros-Elip@67.154.87.254) (Ping timeout)
- # [03:51] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [04:05] * Quits: zcorpan_ (zcorpan@84.216.41.7) (Ping timeout)
- # [04:05] * Joins: zcorpan_ (zcorpan@84.216.41.7)
- # [04:08] * Joins: hyatt (hyatt@24.6.91.161)
- # [04:11] * Quits: hyatt (hyatt@24.6.91.161) (Quit: hyatt)
- # [04:32] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [04:37] * Joins: gavin_ (gavin@74.103.208.221)
- # [04:44] * Joins: Zeros (Zeros-Elip@69.140.48.129)
- # [04:45] * Joins: Zeros_ (Zeros-Elip@67.154.87.254)
- # [04:47] * Joins: hyatt (hyatt@24.6.91.161)
- # [04:47] * Quits: Zeros (Zeros-Elip@69.140.48.129) (Ping timeout)
- # [05:00] * Quits: zcorpan_ (zcorpan@84.216.41.7) (Ping timeout)
- # [05:16] * Quits: hyatt (hyatt@24.6.91.161) (Quit: hyatt)
- # [05:26] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Get thee behind me, satan.)
- # [06:06] * Joins: hyatt (hyatt@24.6.91.161)
- # [06:40] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [06:45] * Joins: gavin_ (gavin@74.103.208.221)
- # [07:02] * Parts: hyatt (hyatt@24.6.91.161)
- # [08:27] <hsivonen> Hixie: does the real world require BOM dropping when HTTP says UTF-16LE or UTF-16BE?
- # [08:27] <hsivonen> Hixie: if not, I think the spec should follow Unicode correctness
- # [08:28] <hsivonen> Hixie: I do think that declaring UTF-16BE or UTF-16LE on the HTTP level is a bad idea, though, and authors should always use the BOM and UTF-16 if they insist on using 16-bit code units
- # [08:36] <Hixie> i think having a BOM at the start of UTF-16LE makes as much sense as having it at the start of UTF-8
- # [08:37] <Hixie> and i think it should be allowed for the same reason -- it's harmless and arises naturally in the creation of text files
- # [08:37] <Hixie> furthermore, as the character itself is invisible, it will be significantly complicated for authors to comprehend how to fix the error if we report it anyway
- # [08:47] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [08:52] * Joins: gavin_ (gavin@74.103.208.221)
- # [09:08] * Quits: Zeros_ (Zeros-Elip@67.154.87.254) (Ping timeout)
- # [09:14] * Joins: Zeros_ (Zeros-Elip@69.140.48.129)
- # [09:18] <hsivonen> Hixie: RFC 2781 disagrees with you. what do current browsers do?
- # [09:19] <hsivonen> that is, is it really harmless for existing doctype sniffing and all that?
- # [09:21] <hsivonen> since the BOM arises naturally, it is really really dumb for an author to specify the endianness on the HTTP level. no gain and a possibility to get it wrong
- # [09:29] <Hixie> not sure what browsers do. I imagine they ignore the BOM though.
- # [09:30] <hsivonen> I guess I should write test cases before discussing this further, but it would be nice not to violate RFCs gratuitously if real-world compat doesn't require doing so
- # [09:31] <Hixie> looks like rfc2781 make it moot what the html5 spec says, as it is then a violation at the encoding layer, not the parser layer
- # [09:31] <Hixie> the rfc doesn't require anything of the UAs as far as i can tell
- # [09:32] <hsivonen> it requires the BOM not to have effect, i.e. the decoded stream to include it
- # [09:32] <hsivonen> if the HTML layer wants to chop it off after that, it is being weird
- # [09:32] <hsivonen> of course, this whole thing is about a labeling case that authors should avoid anyway exactly for issues like this
- # [09:34] <hsivonen> as for encoding layer errors vs. html layer errors, I think the spec should do some hand holding for implementors here, because we cannot assume every parser implementor to be a spec lawyer as well
- # [09:34] <Hixie> i think that in any case a leading BOM should be stripped, whether it is compliant or not
- # [09:35] <Hixie> it would be the height of uselessness to have the BOM, which authors can't see, imply a <body>
- # [09:35] <Hixie> i would be a "gotcha" whose only clear purpose was to confuse authors, in an area where authors can barely understand what's going on anyway
- # [09:36] <Hixie> regarding whether it's an error or not, that's only a conformance checker matter really
- # [09:37] <hsivonen> what if the HTTP-level label is UTF-16 and the byte stream has two leading BOMs? the decoded character stream then has one BOM. should the HTML layer drop it?
- # [09:38] <Hixie> the UTF-16 RFC doesn't as far as i can tell require the leading BOM to be stripped before sending it to the next layer
- # [09:38] <Hixie> I would say that only the HTML (application) layer strips the BOM
- # [09:51] <hsivonen> hmm. I guess my view here is biased by decoder implementations that do drop the BOM in the decoder when the label says UTF-16.
- # [09:53] <hsivonen> anyway, I think this needs test cases to figure out what the current crop of browsers does and then a bit of spec elaboration plus test cases to help implementors
- # [09:53] * hsivonen mumbles something about UTF-8 being the one true encoding and UTF-16 and UTF-32 being time sinks for diligent implementors
- # [09:57] * Joins: ROBOd (robod@86.34.246.154)
- # [10:01] * Quits: gavin (gavin@63.245.208.169) (Client exited)
- # [10:03] * Joins: gavin (gavin@63.245.208.169)
- # [10:11] <Hixie> hsivonen: fair enough
- # [10:11] <Hixie> hsivonen: (note that UTF-8 is not really great for CJK)
- # [10:12] <hsivonen> Hixie: gzip and a very large chunk of characters in CJK HTML is markup and, therefore, in the Basic Latin block
- # [10:42] * Quits: Zeros_ (Zeros-Elip@69.140.48.129) (Quit: Leaving)
- # [10:46] * Quits: jgraham (jgraham@85.210.7.238) (Quit: Ex-Chat)
- # [10:54] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [10:59] * Joins: gavin_ (gavin@74.103.208.221)
- # [11:00] * Joins: hyatt (hyatt@24.6.91.161)
- # [11:07] * Quits: hyatt (hyatt@24.6.91.161) (Quit: hyatt)
- # [11:10] * Joins: frippz (fredrikfro@193.11.209.47)
- # [11:13] <Hixie> hsivonen: yeah, maybe for html utf-8 balances it out
- # [12:11] * Joins: hasather (hasather@81.235.209.174)
- # [12:16] * Joins: tH (Rob@87.102.84.23)
- # [12:17] * Quits: Lachy (Lachlan@203.206.243.149) (Quit: Leaving)
- # [12:42] * Quits: sbuluf (esdtutb@200.49.140.142) (Ping timeout)
- # [12:46] * Joins: Lachy (Lachlan@203.206.243.149)
- # [12:53] * Quits: hasather (hasather@81.235.209.174) (Connection reset by peer)
- # [12:54] * Joins: hasather (hasather@81.235.209.174)
- # [13:02] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [13:07] * Joins: gavin_ (gavin@74.103.208.221)
- # [13:19] * Quits: inimino (inimino@75.71.88.233) (Ping timeout)
- # [13:35] * Joins: zcorpan_ (zcorpan@84.216.40.112)
- # [13:45] * Parts: hasather (hasather@81.235.209.174)
- # [13:50] * Joins: hasather (hasather@81.235.209.174)
- # [15:10] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [15:15] * Joins: gavin_ (gavin@74.103.208.221)
- # [15:29] * Joins: anne (annevk@81.68.67.12)
- # [16:04] * Quits: anne (annevk@81.68.67.12) (Ping timeout)
- # [16:32] * Joins: loic (loic@90.29.27.6)
- # [16:55] * Quits: hasather (hasather@81.235.209.174) (Ping timeout)
- # [16:57] * Quits: beowulf (carisenda@91.84.50.132) (Ping timeout)
- # [16:58] * Joins: beowulf (carisenda@91.84.50.132)
- # [17:00] * Joins: hasather (hasather@81.235.209.174)
- # [17:17] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [17:22] * Joins: gavin_ (gavin@74.103.208.221)
- # [17:47] * Quits: loic (loic@90.29.27.6) (Ping timeout)
- # [18:17] * Quits: zcorpan_ (zcorpan@84.216.40.112) (Ping timeout)
- # [18:20] * Joins: loic (loic@90.29.32.76)
- # [18:33] * Joins: anne (annevk@86.90.70.28)
- # [19:25] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [19:25] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
- # [19:30] * Joins: gavin_ (gavin@74.103.208.221)
- # [20:26] * Quits: tH (Rob@87.102.84.23) (Ping timeout)
- # [20:26] * Joins: tH_ (Rob@87.102.10.97)
- # [20:26] * tH_ is now known as tH
- # [20:38] * Parts: hasather (hasather@81.235.209.174)
- # [20:38] * Joins: hasather (hasather@81.235.209.174)
- # [20:43] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Get thee behind me, satan.)
- # [21:04] * Quits: hasather (hasather@81.235.209.174) (Connection reset by peer)
- # [21:05] * Joins: hasather (hasather@81.235.209.174)
- # [21:10] * Joins: inimino (inimino@75.71.88.233)
- # [21:13] * Parts: hasather (hasather@81.235.209.174)
- # [21:25] * Joins: hasather (hasather@81.235.209.174)
- # [21:32] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
- # [21:33] * Parts: hasather (hasather@81.235.209.174)
- # [21:37] * Joins: gavin_ (gavin@74.103.208.221)
- # [21:48] * Joins: Flor1an (fvandillen@212.67.186.30)
- # [21:53] * Parts: Flor1an (fvandillen@212.67.186.30)
- # [21:53] * Joins: hasather (hasather@81.235.209.174)
- # [22:19] * Quits: tH (Rob@87.102.10.97) (Ping timeout)
- # [22:22] * Joins: tH (Rob@87.102.10.97)
- # [22:38] * Quits: ROBOd (robod@86.34.246.154) (Quit: http://www.robodesign.ro )
- # [23:02] * Joins: sbuluf (rtg@200.49.140.228)
- # Session Close: Sun May 27 00:00:01 2007
The end :)