[Standards-JIG] RE: Standards-JIG Digest, Vol 28, Issue 52

Jean-Louis Seguineau jean-louis.seguineau at laposte.net
Thu May 18 09:43:30 UTC 2006


> You cannot have optimized distribution without handing out a little bit of
> control. 

Delegation always imply relinquishing some direct control, this is true. It
does mean 1/ this  should be mentioned in a disclaimer mentioning the
collateral risks incurred 2/ the proposal should be amended to propose
viable solutions to minimize this risks.

In short, it is not enough to say we can decrease traffic by sending a
single stanza over the wire. Everybody will agree with this. But looking at
the reactions on the list, this 'simple' addition has opened a much bigger
Pandora box. And you are not credible by rejecting the legitimate concerns
expressed on the list because you have build your original thoughts on many
'assumptions'. 

> Also the proposal to use JEP-0033 will pass control over to the
> other servers, and even the current situation is trivial to abuse for any
> user as I said before. I'm sorry you are so emotional about this, but all
> we can do is fix the holes in the system.

I am sorry to say this is not a valid statement. For sure JEP-0033 delegate
routing of stanzas to other compatible multicasters. But only routing is
involved. There is no remote distribution list involved as the actual list
is always contained in the stanza extension. And therefore, the process does
not create privacy and security concerns per-se. In a routing scenario such
as this one, the possible non-delivery of a stanza has been identified as an
issue, and proposals put forward to ensure a better delivery. It does not
entirely solve all the incurred issues, though. Yes you may have forked or
lost traffic. But at least this is known, documented and work is in progress
to address part of it.

Now I agree that applying JEP-0033 to presence in this known context will
not bring the expected decrease of traffic if we also have to ensure
delivery through receipts for example. 

> The 'to' is completely redundant in this case, and there is no break other
> than making the RFC no longer up to date, as fippo has shown. But yes,
> if it makes you happy, we will add a placebo 'to' element to the JEP.
> We're not the ones running the big servers, to whom this extra element
> does make a traffic difference. So if you had access to the current
> version of the proto-JEP you could see that we are acknowledging your
> need to maintain the XMPP traditions.

This way of posting use an unnecessary inflammatory and sarcastic tone,
which is not really appropriate on this list. The all point is not about
pleasing or making anyone 'happy'. You cannot just walk in and bash at other
people practices. In other times kings have been beheaded for less than
this... A little respect for different opinions, a slightly less arrogant
stand would certainly play better in your favor ;)   

The point brought forward on the list is not about the evidence that sending
one message is always better than several. People have express concerns
about the effect of using this approach. They are telling you the advantage
of a decrease in traffic does not compensate for opening security/privacy
breaches. You cannot just dismiss those concerns. You cannot just say the
problems existed, we are just exposing them, go and fix them. This is not a
responsible approach. The least you could do at this point would be to go
back to the drawing board, compile the entire list of concerns exposed on
the list without omission, and propose ways to address (not dismiss) them.
It would show that beyond assumptions you can also make valid analysis. 






More information about the Standards mailing list