[Imap-protocol] Optimistic incremental THREAD=REFERENCES

[Mark Crispin]

On Sat, 10 May 2008, Timo Sirainen wrote:

>> The subject-joining rule magnifies this. Personally I didn't imnplement the 

>> subject-joining bit.

> Me neither. That generates more complexity and I'd rather drop the subject 

> merging completely (my X-REFERENCES2 threading algorithm does that).


Hopefully this does not cause a flamewar.

I do not approve of the practice of an implementor deciding not to 
implement a portion of the specification, even if there seem to be good 
reasons to do so.

This leads us to a "do what we say, not what we do" situation when it 
comes to non-compliant or broken servers such as Exchange and Gmail.

"I didn't implement header searching.  It generates too much complexity."

"I didn't implement the delete-expunge model.  It generates too much 
complexity.  Setting \Deleted just moves to Trash."

"I don't care that I transmit bad syntax in BODYSTRUCTURE.  Outlook 
doesn't use it, so it doesn't matter."

Now, with that said, I do not object to the creation and publications of a 
new threading algorithm that is identical to REFERENCES but omits the 
subject consolidation step.  I would probably implement it, especially if 
it really is identical modulo that step.

Nor would I weep if REFERENCES2 (or whatever) ends up being the one that 
everybody uses and REFERENCES passes into obscurity.

All I ask is that if you implement REFERENCES, implement it according to 
the specification.

-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.