I think that Typepad may have just switched it’s Atom feeds from .3 to 1.0. How do I know? Because two blogs I read just popped up with 10 new entries (none were new) and each one was broken in Bloglines. (Which is the single largest aggregator in the world, at least according to WordPress.com feed stats.) Here is Seth Godin’s as viewed by the feed validator. This is a bold move, but I certainly wouldn’t want to be their support department tomorrow. This could also just be my misunderstanding, as some feeds like this one from Marginal Revolutions (one of my favorite blogs) seems to be on Atom 0.2.
Certainly possible, as I saw a few of TypePad feeds *jump* late last night. They note an upgrade at about that time:
http://www.sixapart.com/typepad/news/2006/04/captcha_on_comm.html
It’s pretty major though, you think they would note it in that post.
They most likely missed it. Seeing that they use Subversion they cannot easily run cvs2cl so people tend to keep track of changes to a NEWS.txt like Zend do for the Zend Framework.
The feed that you link to for Marginal Revolution seems to be really outdated (last update in 2003), and it’s not the feed that they offer to syndicate on their site. Unless I’m confusing something.
I just tried the standard Atom URL on Typepad, figuring that would be the most up-to-date version.
I’m subscribed to six Typepad feeds. Two of those are still using Atom 0.3; four are Atom 1.0. I didn’t get any new messages that weren’t actually new, so maybe they’ve always been using Atom 1.0.
As for Bloglines, it’s not exactly famous for its flawless support of any of the syndication formats. I don’t think it’s fair to blame Typepad for that.
I’m not blaming them for anything, just noting what I experienced.
Just curious, Matt, how would you have wanted it to be handled? In general, I think that tools should be aggressive about implementing standards to spec, even if it causes a slight inconvenience it benefits over time.
Implementing standards halfway is arguably worse than not implementing them at all.
Marginal Revolutions feed generates “Error parsing RSS XML: Unsupported Atom version: 0.2 ” on my aggregator (SharpReader).
It’s a tough problem, we’re facing the same thing. Arguably
/atom/in WordPress should be Atom 1.0 even though it’s currently 0.3. We’re basically still sitting out the switch to give the market more time to catch up with the format.If our bloggers update WordPress and their feeds which worked yesterday stop working, they blame us, not hte aggregator or the format. I guess that’s why I’m surprised about Typepad’s move, because I’m sure that their users are going to assume Typepad broke something, not that they got an upgrade, especially since functionally to the end-user both version of Atom are identical.
You have a fair point, Matt – I upgraded to Atom 1.0 some time ago and then had to make changes to ensure that my feed worked in Bloglines after people started complaining to me that it was screwed up. But hopefully having a large hosted service like this migrating will mean that Bloglines et al get around to making their aggregators fully compatible with Atom 1.0. It’s not like it’s brand-spanking new either – the Atom 1.0 spec was agreed in July last year.
I should have said they “may be more inclined to get around to making their aggregators fully compatible with Atom 1.0.”
True, but XHTML 1.1 was released in 2001, and still has horrible support.
>XHTML 1.1 was released in 2001, and still has horrible support.
Indeed.
IMHO:
In short: because XHTML 1.0 Strict (or Basic) is good enough for nearly everyone who cares to use XHTML instead of HTML4.
Longer: XHTML 1.1 offers no additional practical(*) benefits to web publishers/authors over XHTML 1.0 Strict, and requires use of the application/xhtml+xml content-type, which then requires that the content be treated as XML, so you got weirdnesses like *either* you have to wait until your browser has loaded and well-formedness-checked the entire document before you see it (no progressive rendering), *OR* (and this is stranger IMHO), you see parts of a document display progressively, and then if there is an XML error (like a missing close tag) at the end of the document, the browser is supposed to not show the document and give you an error instead. This makes no sense to the user.
The analogy does not apply to Atom v1.0 vs. Atom v0.3 vs. RSS 2.0. Atom 1.0 is an official IETF standard, Atom 0.3, even by the people who wrote it, is not a standard. RSS 2.0 is both a defacto standard, and an official standard as declared by those who wrote it.
Tantek
(*)XHTML 1.1 was supposed to allow more easy valid (and validatable I think) semantic extensions to XHTML via “modularization” (above and beyond XHTML 1.0), but in practice that has been a sad joke. In the past year plus, another method for easy valid semantic extensions to XHTML has shown far more promise.
TypePad actually deployed support for Atom 1.0 several months ago, albeit with little fan fare. I mean honestly, how many people really care what version their feed is. Aside from us geeks?
There are, however, a minority of blogs on TypePad that use TypePad’s advanced templating system – for those users, their Atom feeds are likely to still be using the 0.3 templates. So technically it is still possible to find blogs on TypePad with 0.3 Atom feeds.
What’s interesting to me is that feed readers are just plain dumb. Seriously. The other issue mentioned above, the one where Bloglines detected all new content in a feed, when in fact there was none, was the result of an errant XML element that made its way into our Atom feeds. We we quickly rolled back that change ( bug), but for the brief period of time that the bug manifested itself, feed readers interpretted those feeds as being completely new. No other aspect of the feeds changed, especially the most critical element: the element, which should signal to a smart reader which elements are actually new.
Upon brief investigation I learned that many some feed readers maintain checksums for entries within a feed. And if the checksum changes, the entry is determined to be new. So the addition of an element that has no bearing on the content at all results in some (what I consider to be) unexpected behavior in some readers.
Thanks for the inside scoop Byrne. 🙂 Assuming you’re in SoMA, we should grab a bite sometime.
This is true: WordPress will remain at the mercy of Bloglines probably only as long as it helps Bloglines remain complacent. Much like IE vs. lax Web authors—it’s an codependency of regressiveness.