26 August 2005

The 100+ Year Old RSS Virgin

Better late than never.

An RSS 2.0 feed of AT&T News is now available from the home page or direct from the source.

Many thanks to one Ben Hammersley for writing a most excellent and comprehensive book on how to build them.


Whither Atom and RSS 1.0? Good question. Perhaps we should contemplate RSS 1.0 (which can contain a lot more info within) and use Feedburner.

What about the content type? Right now we’re examining the Accept header. If it sports application/xml then we send back application/rss+xml. Otherwise we fall back to <gasp!> text/xml.

I’m not sure if this is entirely wise, looking for application/xml only to send back application/rss+xml. Perhaps we should just call it as we see it, so to speak. Then again, are any user agents sending application/rss+xml in the first place? Firefox doesn’t appear to be. If I understand correctly, it’s not 100% official as far as MIME types go. As for text/xml, isn’t that going the way of the dodo? Yet a lot of browsers support it nowadays. Even Dave Winer sez use it.

Ahh, the conservative (production) vs. the liberal (acceptance). Whaddaya gonna do.

What else. No <guid>? No, not at the moment. I tried using that in conjunction with <link>, and in the latter case tacking on ?from=rss (a la Slashdot). Somehow, adding query string data like this to a <guid> permalink “seems wrong” in terms of wanting to keep the permalink comparatively “pure” but that too could be unfounded. As it is, I noticed that Bloglines will point the title to the <link> value and the “Link” at the end of the post to the <guid> permalink value, so the tracking ability is somewhat lost depending on what link folks follow. Then again this behavior could be entirely correct on Bloglines’ part.

Over time I’m hopeful we’ll add more feeds (not all of them news and whitepapers, hint-hint), an RSS index for easy one-stop feeding, even some OPML (once we have more feeds to actually put in it).

Posted by joe at 02:15 PM

Trackback Pings

TrackBack URL for this entry:


Dan of Dan Century writes:

I like the approach where you default to text/xml, because it’s then open to any tool that can grab xml over an http connection, not just one that’s coded to specifically handle rss+xml. That would save someone a little time and money. It’s very considerate of you. My wack, generic xml parsers and I thank you. :)

BTW, your feed is sent to Firefox as text/xml, not application/rss+xml:

HTTP/1.x 200 OK
Server: Netscape-Enterprise/4.1
Date: Sat, 27 Aug 2005 15:45:58 GMT
Content-Type: text/xml; charset=utf-8
Connection: close

despite the fact application/xml appears in the Accept header:

Accept: text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,

Please don’t work on this over the weekend.

# 27 August 2005, 11:53 AM -05:00

Joe writes:

Hmm. That’s no good. We checked it with Firefox and Fiddler - seemed to do the right thing. I’ll check it again though. On Monday. :)

# 27 August 2005, 01:34 PM -05:00

Joe writes:

OK. Turns out I was mistaken. We had it checking for application/rss+xml outright, never application/xml.

So it was a 1:1 match after all. Right. Should we ever check for or send back application/xml? Or just application/rss+xml with a default of text/xml? The mind reels.

# 29 August 2005, 09:59 AM -05:00

Post a comment

(will not be published)

Remember Me?

(you may use HTML tags for style)