- Wednesday, November 12, 2014

mif@jabber.ietf.org

Wednesday, November 12, 2014< ^ >

Dave Thaler has set the subject to: IETF-90 Toronto meeting in progress

Room Configuration

Room Occupants

GMT+0
[19:01:13] dmitry.anipko joins the room
[19:13:34] dmitry.anipko leaves the room
[19:16:08] dmitry.anipko joins the room
[19:22:34] dmitry.anipko leaves the room
[22:56:05] Meetecho joins the room
[22:57:46] dmitry.anipko joins the room
[23:03:21] Dmitry Anipko joins the room
[23:10:08] sarikaya2012 joins the room
[23:11:36] dmitry.anipko leaves the room
[23:15:17] <sarikaya2012> Suresh is presenting
[23:15:19] <sarikaya2012> http://tools.ietf.org/html?draft=draft-ietf-mif-mpvd-id-00
[23:15:22] dmitry.anipko joins the room
[23:16:29] Ted Lemon joins the room
[23:26:27] Dave Thaler joins the room
[23:27:58] <Dave Thaler> http://tools.ietf.org/html/rfc2776#section-3.3 had a similar problem and chose the IP address syntax to guarantee there didn't need to be something else to configure
[23:28:08] <Dave Thaler> similarly routing protocols chose IP router ID
[23:30:36] dmitry.anipko leaves the room
[23:31:45] dmitry.anipko joins the room
[23:37:00] 410fd10042cc10d3 joins the room
[23:37:06] 410fd10042cc10d3 leaves the room
[23:37:25] Alex Petrescu joins the room
[23:38:12] <Alex Petrescu> UTF-8 is better than ASCII 7bits but isnt UTF-16 even betteR?
[23:39:57] <Ted Lemon> no
[23:40:24] <Ted Lemon> well, maybe "it's complicated."
[23:41:00] <Ted Lemon> utf8 does have the problem that it favors roman scripts.   but it's also the encoding that's used most commonly in other protocols, whereas UTF-16 is very rarely used as an encoding format.
[23:41:30] <Dave Thaler> UTF-8 is almost always better in network protocols... it's more compact and has no byte-endian issues.   See the plenary slides we did in the Hiroshima plenary
[23:42:46] <Alex Petrescu> endianness - we have this network order and host order, recent platforms have this capability deal with both...
[23:42:59] <Dave Thaler> so the IETF policy is to use UTF-8 (RFC number is one I have to look up)
[23:43:07] <Dave Thaler> RFC 2277
[23:44:26] <Alex Petrescu> by its rfc number I can guess the recommendation UTF-8 is probably as an evolution from ASCII-7, or Ill have to read it.
[23:44:26] dmitry.anipko leaves the room
[23:45:44] <Dave Thaler> it remains the recommendation (again, see the slides from the hiroshima plenary at http://www.ietf.org/proceedings/76/slides/plenaryt-1.pdf for more info)
[23:45:55] Matt Mathis joins the room
[23:47:22] Matt Mathis leaves the room
[23:48:26] Dan Wing joins the room
[23:51:34] <Dave Thaler> anything you can do in UTF-8 you can do in UTF-16 and vice versa.  The differences are just in how many bytes it takes and whether endianness is an issue
[23:52:20] <sarikaya2012> Or if you have something like Chineese alphabet with so many characters?
[23:54:30] dmitry.anipko joins the room
[23:55:48] <Dave Thaler> as Ted says, there are some cases where the UTF-16 encoding might be more compact than the UTF-8 encoding but usually it's the other way around.  The IETF policy (RFC 2277) is that ietf protocols must support UTF-8.  They may also support other encodings too, but UTF-8 is the one that's mandatory to implement.
[23:57:14] <Dave Thaler> the iab I18N program has an action item to propose updates to RFC 2277 to say additional things, but there's no plan to change the recommendations it does say.