[standards-jig] JEP-0102

David Banes dbanes at ozemail.com.au
Tue Jul 1 12:09:27 UTC 2003


Hi all,

I'm going to make a couple of observations as someone on the edge of the 
Jabber developer community that's trying hard to develop Jabber clients. 

I'm not taking a swipe at anyone, far from it, I want to help, 25 years 
coding and 15 running dev teams should get me some street cred. :-)

I apologise in advance if any of the following is not quite accurate, 
but you'll get the idea.

1) We (www.cleartext.com.au) looked at Jabber in 1999/2000 and decided 
to wait until it was a more established.  The assumption being that the 
market was so fragmented that another IM solution stood as good a chance 
as any. In the meantime we developed our won proprietary IM complete 
with public/private key crypto (RSA).
--- Well it(Jabber) is established, and even better there are people 
using Delphi to implement Jabber applications.:)

2) One reason for not supporting Jabber in 2000 was lack of file 
transfer. In the meantime we developed our own proprietary IM with file 
transfer, complete with crypto and digital signatures(RSASSA).
--- Where is the 'ratified' interoperable file transfer? All I see is a 
bunch of confusing schemes and various implementations, some are cool 
and they work. Why arent we using these?

3) Another reason for not supporting Jabber was lack of clear standards 
to code to.
--- On the whole this is much better, particularly the earlier JEP's 
documenting base functionality. Later specs read like rfc's designed by 
committees and we'd need an army of documentation people and analysts 
just to interpret them for the developers. :(

4) Seeing the vague support for OpenPGP was a supporting factor in the 
decision to come back to Jabber. This would fit in nicely with our own 
encryption support scheme in v1 of our product.
--- However, it's not clearly documented, which leads me on to JEP-0102 (
http://www.jabber.org/jeps/jep-0102.html). Who other than the author (
good job bye the way Jean-Louis) and someone like myself who's spent the 
last three years working with crypto will be able to implement this in a 
reasonable time frame?

So, to summarise, whilst we(Cleartext) have not been widely successful (
and are actually poorer for the journey in $ terms!) in developing a 
world beating IM solution, we did at least manage to deliver a reliable 
secure IM system, with messaging, presence notification and file 
transfer; with two developers and many late nights.

This as all been achieved within the time frame that the Jabber 
community has been growing. So my question is, why? 

Why don't we have file transfer? 
Why don't we have end to end crypto? 
Why don't we have ...

There are many, many good (great) things about Jabber but I can't help 
thinking the ratio of discussion/documentation to agreement/coding is 
weighted the wrong way.

Someone on one of these lists hit the nail on the head a few days ago; 
what we need are simple easy to understand and implement base line 
features that at least get us to first base on the competitive journey, 
we can go back at a later date and improve on these.

Again, Cleartext may have to re-consider our decision to tread the 
Jabber path and come back in a couple of years. I hope not, with MS, AIM 
and Y! still unable to agree on a common architecture I really do think 
Jabber can take a shot at being a real standard, the standard.

Best Regards, and with good intent,

David Banes
Cleartext Pty Ltd




More information about the Standards mailing list