[standards-jig] JEPs and Jabber Adoption

Tijl Houtbeckers thoutbeckers at splendo.com
Sun Jun 29 14:54:31 UTC 2003

Nathan Walp <faceprint at think.faceprint.com> wrote on 29-6-2003 16:25:10:
>I'm not even gonna quote a specific reply here, it's been said about
>twelve times, so just imagine I'm replying to someone who is comparing
>JEPs to the linux kernel ;-)
>If I implement a new VM for the kernel (for example), and it SUCKS, has
>some horribly major flaw, the WORST that it can do is hurt the machine
>that is running it.  That is all.
>If I write a new JEP, it goes to Draft, gets implemented, and a huge
>flaw is discovered, the WORST it can do is hurt the entire XMPP/Jabber
>community, because now that implementation is in the wild.

Why would *any* of the client authors do such a thing? Are they people 
outthere that want to hurt Jabber that badly? What are you saying 
anyway? That we don't implement anything cause we might risk "hurting 
the entire community"? 

I'm getting tired of these "don't implement anything cause we'll end up 
in interoperatability hell" arguments cause there is nothing to back 
them up! And that's *still* besides the point, cause time and time 
again I've made this point clear: JEPs are being blocked because of 
competing ideas, because there are different solutions to the same 
problem the entire process becomes *stuck*. Nowhere, and I do mean 
nowhere, did I suggest we advance JEPs to DRAFT that have horrible 
flaws in them. 

>Does everyone see the difference?  The spaghetti-wall method of working
>(keep throwing stuff against the wall until something sticks) just 
>can't work for Jabber, because we all have to live with whatever 
>decisions are made for a long time.

And this is simply not true. Just because a JEP is DRAFT does not, not 
at all even, mean that any of you have to work with it unless you like 
the what the JEP is trying to do. The whole point of an implementation 
of this idea (just like with linux) is that you can see the idea be put 
in to practise. Then you have at least a solid base to critisize it if 
you think your idea is so much better. 

People here are suggesting it's impossible to write a good JEP, take 
the flaws out, and keep it stable and backwards compatible if any 
updates were to be released, unless 100% of the community and council 
is 100% happy with the idea before it's ever used. What drives you to 
say that? Are people convinced that an idea MUST be bad just because 
they don't like it at first sight? 

No JEPs that have horrible technical flaws in them will go to DRAFT, 
nor any JEPs that claim to do something they do not. I'm not suggesting 
the community should go stupid. I'm suggesting we advance JEPs that are 
technically ok, with all the flaws worked out by all who want to 
contribute to it, and that do what they say (solve the problem they are 
trying to tackle). We're surthenly not doing that now. But honestly, 
that doesn't sound like such a strange idea to me. But when I hear you 
people talking it sounds like all hell will break loose if we'd start 
doing that. :) 

>Linus et. all could care less about the VM in the
>2.0 series of kernels.

Well, I can't argue with that..

Tijl Houtbeckers
Software Engineer @ Splendo
The Netherlands

More information about the Standards mailing list