[Standards] pubsub/pep auto-creation

Ian Paterson ian.paterson at clientside.co.uk
Sat Mar 24 13:20:58 UTC 2007


Rachel Blackman wrote:
>> We should make user interfaces as easy as possible - even if they are 
>> rarely used. The flow you describe forces the user to provide input 
>> in a second dialog that they probably weren't expecting to pop-up on 
>> their screen after the round trip to the server. This is more 
>> irritating and less efficient for the user than if a "Publish 
>> Publicly" checkbox had been included on the first dialog (completely 
>> eliminating the need for the second dialog).
>
> I'm actually going to disagree on this one; the protocol and the 
> dialogs do not need to directly coincide.  It's perfectly possible to 
> have the 'Publish Publicly' checkbox in your client's 
> publish-information form, and still do the round-trip 
> check-config-and-replace-if-necessary.

I agree it is possible, but as I pointed out in the previous post 
(http://mail.jabber.org/pipermail/standards/2007-March/014417.html), not 
only is it sub-optimal, it is vulnerable to race conditions. Although 
errors would typically be very rare in practice, IMHO, the existance of 
the vulnerability alone makes it hard to justify the 
"check-config-and-replace-if-necessary" approach when implementing this 
user interface.

> I still think publish+configure makes things simpler, as it means a 
> PEP publish stanza is sort of 'fire and forget,' no need to do 
> multiple exchanges with the server, so I'm still in the pro camp there.

Yes, it seems we have a consensus of all (client) developers bar one. 
Although he is one of the PEP authors, and probably neither camp is 
"wrong" ;-)

- Ian




More information about the Standards mailing list