Then I slept on it. And in the morning, it started to make
some sense. It reminded me of another time, years ago...
The current version of
Apache
Tomcat is 5.5.15. Two prior versions are still
supported It wasn’t always that way. Six years
ago, there was a messy fight between those that were maintaining
Tomcat version 3 and those working on a version 4. This
stressful situation lead James
Duncan Davidson to pen
Rules for Revolutionaries. It talked
about how attempting to force differing approaches onto a numeric and
ordered scale increases social friction. We saw a similar situation
in the syndication space when RSS 0.94 was renamed 2.0.
The
IETF
AtomPub working group internalized that knowledge and respected
the roadmap,
leading to a
tentative
endorsement by Dave Winer (note: the effort was known as Echo
at that time). Also note that that endorsement merely acknowledged
that another effort was underway, it was emphatically not an
acknowledgment that the the effort was in any way be treated as a
successor to RSS 2.0.
The situation is
fractal. The issues that arise in a 3.0 vs 4.0 discussion
apply equally to a 2.0.1 vs 2.0.2 discussion.
To this, I say,
What Dave Said, though I will voice my
preference
on the two options that Dave puts on the table.
Footnote: I have a pet peeve concerning the use of the term ”RSS” without
further qualification. Depending on the content, it means non-RDF versions of RSS, or all syndication formats that call themselves RSS, or all popular syndication formats. In many cases, such ambiguous statements end up implying things that simply aren’t true.
I’m fumbling around for a style convention for how to talk about Really Simple Syndication without making people think I’m also talking about RDF Site Summary or Atom.
When I use “RSS”, I generally mean Really Simple Syndication only. The use of RSS in the media and blogs to describe Atom bugs me.
Yikes. This is turning into a soap opera (as opposed to a SOAP opera ... oh nevermind). I’m not sure I want to invest the time to separate the Players from the Story.
To offer my doubtless inaccurate sum up ...
Atom took forever, was tedious to follow, but did finally arrive with a clear definition and (more important) a clear set of unit tests.
RSS is a small zoo of slightly foggy beasts that are widely used and mostly work, except when they don’t - and that’s about as good as you can expect.
“People” commonly have used “RSS” both as referring to a whole bunch of things+ and as referring to one or other specific markup format (usually being RSS 2.0, at this point).
IMO, the power of RSS has always been around the former, not the latter--which is why “people” don’t care so much whether their RSS is 1 or 2 or Atom or whatever.
Sam Ruby: “Early drafts of the new spec referred to “its predecessor”. This language was removed, but the version number was bumped. Quite frankly, both are acts of social violence.” The problem with this RSS war is less about...
Let’s say I was developing an application from scratch which had the ability to produce feeds. Let’s say that I could only pick one feed format. Why would I pick RSS over Atom (is the only reason that RSS may have wider adoption)?
It seems like RSS is a mess, and that the Atom spec clearly defines what type of data can go where in the feed. Or am I mis-understanding?
I’m fumbling around for a style convention for how to talk about Really Simple Syndication without making people think I’m also talking about RDF Site Summary or Atom.
The only deployed specification that expands the RSS acronym out in that manner is RSS 2.0. In many cases (in particular in a description of non-Atom, non-RDF podcasting applications), describing it that way is both clear and precise.
Sam Ruby: Rogers makes the case that all podcasters relying on multiple enclosures will be publishing RSS feeds that don’t work for what is potentially their largest audience, and Dave pleads for Rogers to “decide whether you’re...
Elliot: most of what you see by in this post by Daniel Berlinger and the ensuing discussion between Daniel and Phil captures my sentiments on this discussion.
The only thing I would add is that at this point, in early 2006, I would recommend that you stick with one of the big three: RSS 1.0, RSS 2.0, and Atom 1.0.
I’ll come right out with a stronger recommendation: pick RSS 1.x if your internal data model is RDF anyway; in all other cases, use Atom.
RSS 0.9x/2.0 has the widest deployment in terms of published feeds, but both other choices are deployed so widely that no consumer application can afford to ignore them, so there is no reason to pick RSS 0.9x/2.0 just because of mind share. RSS 0.9x/2.0 is only advisable if you are targeting specific consumers who will have known and unfixable problems with the other choices.
Sam Ruby: “Early drafts of the new spec referred to “its predecessor”. This language was removed, but the version number was bumped. Quite frankly, both are acts of social violence.” The problem with this RSS war is less about...
Sam Ruby: Version Numbers An interesting view on software versioning, and the problems that can occur when a new version comes out with a bumped version number. “Attempting to force differing approaches onto a numeric and [...]...