[Standards] About stream namespaces

Thomas Charron twaffle at gmail.com
Sat Mar 17 15:36:03 UTC 2007


On 3/17/07, Carlo v. Loesch <CvL at mail.symlynx.com> wrote:
> Peter Saint-Andre typeth:
> | Where were you in 1999 when we were defining XML streams? Where were you
> | in 2002-2003 when we formalized XML streams at the IETF? Obviously we
> | sure could have used your help...
> Who gave a damn when around 1999 we pointed out that Jabber (IMPP/XMPP
> whatever) was not going to scale well if it ignored the multiple
> recipient problem?

  No one.  Just as no one posts a link to it on slashdot when after 11
years, a 1.0 is released.

> Who said back then, that multicast presence is not necessary, because
> the average user would only have ten people in his roster? Why does
> XMPP have 60% presence overhead¹ today?

  Because an individuals presence changes 100 times a day, while they
may only have 2 or 3 chat conversations.

> If XMPP is suboptimal today, there is a flaw in the process of
> decision-making.

  Suboptimal by your definition.  BTW, HTTP is also suboptimal.

> We were there, willing to bring our 12 years of headstart over Jabber
> into the design of XMPP, to make the best of it. In fact we are still
> here, trying to make the world better. May sound cheesy, but it's as
> simple as that.

  You HARDLY have a 12 year head start over Jabber.  I know, this
entire message is a flamed response to a flame pst to begin with, but
this is ridiculous.

> So given the circumstances, excuse us if we are occasionally a little
> angry.

  I'd bet you a beelion dollars that none of your servers is coming
close to the traffic that..  You know what, it isn't even worth the
conversation.

  Physced was built to be a replacement for IRC.

> P.S. Concerning the choice of XML as a syntax in messaging protocols, enjoy
> the dozen good reasons why that was a bad idea on http://about.psyc.eu/XML

  All of your reasons why XML is a bad idea are not why XML is bad,
but why rfc822 is good.  I should also point out that when used in
conjunction with an SSL layer, your reasons are entirely flawed
because the entire stream gets compressed along the way.  And XML is
HIGHLY compressible.

  Oh, and XMPP is fragmented XML.  Which is perfectly valid.  Quoting
conversations from your own chat rooms is hardly reputable evidence to
the contrary.

-- 
-- Thomas



More information about the Standards mailing list