Re: A message to the lurkers on the list

[Aaron Swartz <aswartz@swartzfam.com>]

Dave Winer <dave@userland.com> wrote:

> 1. I too am interested in having the ability to add the full text of an
> article to the RSSization.

That's good to hear. I'm glad there's so much support for the content
module. It was originally introduced as the weblog module, to allow
functionality similar to that provided by the scriptingNews format, but I
quickly realized it had broad uses. This is a good example of the power of
extensibility, and how RSS is useful to all websites, not just those that
publish news.

> 2. The divorce I seek could be elegantly accomplished by choosing a
> different name for "RSS 1.0".

If we changed the name, would this resolve all of your issues with the 1.0
proposal?

> Even Dublin Core is not as simple as it appears.

Could you explain this statement? Why is that?

> That's the slippery slope issue,
> tomorrow the RDF working group could change their mind and invalidate RSS.
> Ooooops.

This doesn't make any sense to me at all. I don't understand how the WG
could "invalidate RSS" since it works just fine even without the existence
of RDF as an official standard. Even so, RDF is an official W3C
Recommendation, and, AFAIK, Recommendations cannot be invalidated, merely
deprecated. Even if it could be invalidated, the same could be said about
RSS, since it's based upon XML which is also a W3C Recommendation -- exactly
the same as RDF.

> 4. I've always felt that your ticker element was the canonical example of
> why RSS should upgrade incrementally. Having the docs for such an element in
> the baseline spec would give people ideas of the practicality of RSS. Like
> the Dial Phone command in MORE 1.0, few if any used it, but when they saw it
> there, it gave them an appreciation for the utility of the product. Further
> such examples might make the format even richer.

I agree, it's good to have such examples of power, like the Ticker element.
We try to do this in the RSS spec, with the examples at the end, and also, I
expect nearly all RSS tutorials/explanations will include examples of useful
modules. However, placing such features in the spec makes things more
difficult, since we'll need to come out with a new version of the spec every
time we want to add new functionality. IMO, this would be a version mess,
since most RSS applications would be able to read later versions just fine,
even if they didn't understand some of the new elements.

IMO, namespaces makes it easier for everyone, as well as allowing people to
extend RSS for personal uses, without having their extensions added as part
of the spec. Not to mention allowing people to route-around the group in
charge of the spec, in case it gets too bogged down.

-- 
Aaron Swartz | me@aaronsw.com | http://www.aaronsw.com