[Standards] XEP-0138 vs. TLS compression

Boyd Fletcher boyd.fletcher at je.jfcom.mil
Thu Aug 30 20:05:44 UTC 2007




On 8/30/07 1:50 PM, "Peter Saint-Andre" <stpeter at stpeter.im> wrote:

> Boyd Fletcher wrote:
>> I strongly disagree. I think that sends a confusing message to developers.
> 
> How so? We have a registry. Refer to the registry. Do you think the XMPP
> Registrar creates these things for fun? ;-)
> 

I don't have a problem with the registry. I just had a problem with XEP-138
listing LZW and not EXI.


>> Perhaps, we should remove all the compression algorithm references XEP-138
>> and have it only specify the core approach. Then create new XEPs for
>> specific implementations: ZLIB, LZW, EXI, etc.... That is analogous to how
>> the IETF does it - look at how they handled RTP and the different packet
>> formats.
> 
> Right. I have no objections to that. But ZLIB is mandatory to implement.
> 

should that be numbered like XEP-138a or a totally new XEP.

>> Or, we make XEP-138 obsolete and create a new one that adds EXI. IETF also
>> does this approach.
> 
> That seems unnecessary.
> 
>> But having a separate XEP for just EXI for stream compression would I think
>> be sending a confusing message to developers.
> 
> As far as I understand it, Efficient XML is more than just a compression
> algorithm. In order to prevent developer confusion, I think it might be
> helpful to write a XEP that describes how Efficient XML is to be used in
> the context of XMPP applications. That XEP could also register an item

probably a good idea, but I think that is different than how it can be used
for stream compression.

> to be added to the compression algorithms registry.
> 


+1 for adding to the registry but we should remove LZW from XEP-138 and move
its description text to registry.


> /psa
> 



More information about the Standards mailing list