[Standards] LAST CALL: XEP-0308 (Last Message Correction)

Gunnar Hellström gunnar.hellstrom at omnitor.se
Tue Aug 14 20:03:02 UTC 2012

On 2012-08-14 20:09, Kurt Zeilenga wrote:
> On Aug 14, 2012, at 8:57 AM, Peter Saint-Andre <stpeter at stpeter.im> wrote:
>> In fact, I'd argue that this spec is a technical solution to a social
>> problem, and thus is wrongheaded. It's easy enough to send a new
>> message correcting the old one (e.g., "sorry, I meant to say Y instead
>> of X" or "s/X/Y/"). People do that all the time. Magically and
>> invisibly changing a message after the fact strikes me as a really bad
>> idea.
> The more I think about it, the more I think I that this XEP is a bad idea.
> Basically, the kind of user confusion I illustrated in by prior post can really only be addressed if this extension is mandatory to implement (for all clients).  And, I don't think extensions should be mandatory to implement.  They should be truly optional.
> -- Kurt
I think this extension gets rid of one annoying characteristic of XMPP 
IM. - The message borderline being an artificial borderline for 

If an implementer really thinks it is risky, and that it can cause 
undetected modifications, then the implementation can require from the 
user that there is no text in the current message before the last 
message may be entered for correction.  So, if you have started to type 
the new current message, you need to erase it before entering the 
previous message for correcting it.  That will make corrections appear 
with care and only when it is felt efficient by the typing user. There 
is no difference in the protocol for this feature, so it does not need 
to influence the specification.

I hope that the current discussion, with some voices saying that 
correction of last message should not be allowed, will not influence the 
acceptance of XEP-0308.


More information about the Standards mailing list