How does someone add something to RSS 0.9x?

Last week I spent some time talking with Ben Hammersley, the British tech journalist who's working on O'Reilly's book on XML-based content syndication with RSS and other formats. Yesterday we emailed back and forth several times. I think we're getting started on something that can work. My last email to Ben yesterday attempted to answer this question, how does RSS 0.9x evolve?

How does someone add something to RSS 0.9x? 

Shortly after 0.91 came out, Mark Kennedy of Motley Fool posted an RFC about adding a <cusip> sub-element to <item>. No consensus was reached (we had no process after Netscape withdrew) so he just added it to his feeds. Nothing broke. Basically, it seems to me, that anyone can add elements to RSS 0.9x and nothing breaks.

So, instead of making the base format more complex, why not just allow what's been happening to continue to happen (in fact there's nothing we can do to stop it, one of the fallacies of mail lists is that they have any authority on what people can or can't do).

We could also do something truly revolutionary -- work together in a collegial way. I've been lucky to be part of several processes in my career that worked this way, and the results were spectacular. This is the real issue for RSS, much more important than the names of individual elements, or whether or not we use namespaces (I think we should, but only after we get some working-together energy going, without that you just move the problem out one level, you pay the complexity tax and get nothing in return).

Emphatically, I am just a note-taker in the informal RSS process. I've been careful to build on other people's ideas, and when people address each other with respect, we've been able to make a lot of progress. Look at the delta betw 0.91 and 0.92. Lots of useful features were added. We can add more. The three which we're exploring now are all very important missing functionality in RSS. Why can't we have them? We can, imho.

Look at the example of how the Blogger API came into being. Evan proposed a blogging API. UserLand adopted it, with no revisions. When something arrives that works the best answer is yes, and not piss over minor distinctions. I wrote a piece on Sat where I explained this. The guiding philosophy is "If possible I'm going to do it the way you want to do it."

Ben, I'm hoping your work can contribute to getting a collegial community to form in place of the unproductive mail lists. One good way would be to carefully comment on the three new elements I wrote about, after your jetlag clears, and do it in public and be conspicuous in respect, and let's set the tone for something that's moving, not paralyzed.

BTW, the <ttl> feature is important in defending the Internet against the US Congress and the RIAA and MPAA. I want to get tons of great RSS content to flow through Gnutella. Makes it harder to shut it down, and further, makes us more decentralized, in case we have to deal with another 9/11.