Henri Sivonen wrote:
On Mar 1, 2010, at 11:58, Julian Reschke wrote:
which reminds me that the Atom WG successfully assigned the final
namespace name only *after* the spec was approved, thereby avoiding
problems with compatibility problems of early implementations.
It's not clear to me that the Atom WG was "successful" on this point.
There are still a lot of "Atom 0.3" feeds around. To the point that
feed consuming apps probably need to support "Atom 0.3" to be
competitive.
I do not believe that to be the case.
I see how this could be seen as a versioning success (you can
distinguish 0.3 and 1.0), but I see it as a failure to eradicate
versioned draft-based deployments because versioning allows you to
support both. That is, versioning didn't remove the problem of old
stuff sticking around. Instead, to avoid the problem of having to
support multiple versions in the future, there needs to be a solution
other that distinguishable versions.
- Sam Ruby
