Atom is done. Now the editorial processes grind away and eventually the official specification of the Atom Publishing Protocol will be an RFC substantially identical to draft-ietf-atompub-protocol-17; it’ll join RFC4287 as the official products of the IETF Atompub Working Group.

What’s Next? · Now we’ll find out who’s interested. The Atom feed format is a success; RSS isn’t going away, but a steadily-increasing proportion of the world’s new feeds are Atom 1.0.

I personally think the protocol’s going to be a big deal; here’s why. My track record as a prognosticator is good but not perfect. There’s one thing of which I’m confident though: if Atompub takes off, it’ll be in at least one area that makes no sense at all to me, seems completely crazy.

Atom and Me · I started getting interested in in Sam Ruby’s “Pie” project in 2003. The IETF process started in 2004. So that’s four years of my life, more or less.

I don’t regret a moment of it. I got to actually work, not just hang out, with Sam and Joe Gregorio and, well, if I start listing names, this entry will get out of control. Except for Paul Hoffman, who combined IETF process expertise with deft interpersonal skills, exercised not least in putting me on a leash when I wanted to deal with blockages by applying extreme violence. He’s become a friend, and you can’t have enough of those.

I should also say thanks to Sun for paying me while I spent part of my time working on this. Having said that, I think anything that makes the Net bigger and easier and more useful is good for us because after all we sell the iron; so I think it was a decent investment.

Plus, well, I think the Internet and the Web really need the Atom protocol.

I’m kind of hoping, though, that I can take a vacation from standards work for a while.

The IETF? Well, it ain’t perfect, but work gets done. It’s absolutely intolerable, though, that the name of our spec co-editor Bill de hÓra will appear mis-spelled on the RFC because of the IETF’s bone-headed ignorant backward-looking bigoted ASCII-only policy. Speaking personally, it’s a problem I can’t just live with, which means probably not much further work in the IETF. Oh well.

What Do We Call It? · The term “Atom” is hopelessly vague, and most people use it to refer to the feed format, which is fine. We could say “Atom Protocol” or “APP” or “Atompub”; let’s see what shakes out.


Comment feed for ongoing:Comments feed

From: Dustin Whitney (Jul 24 2007, at 13:48)

Where will Atompub's success make no sense to you at all? Or was that a statement that such a thing will happen, and you're not quite sure where, yet?


From: Nick Johnson (Jul 24 2007, at 14:00)

Excellent! Coincidentally, I've just started working on a forum package in Python that will use Atompub to allow programmatic reading and posting of the forum outside the HTML interface.


From: Scott Johnson (Jul 24 2007, at 18:21)

Congrats on a job well done! Those four years you gave to the project mean a lot to the web. We are all in your debt.


From: pkeane (Jul 24 2007, at 21:51)

Well Atom and APP could certainly be the next big thing, and from the look of the Google APIs, etc., some other folks thing so as well. The question (I think) is whether folks will be able to recognize it as the 'right tool for the job'. It can be a simpler Dublin Core (base line metadata schema), a simpler WebDAV (transfer protocol), a simpler OAI_PMH (protocol for metadata harvesting). As an application developer I need those protocols to tie together increasingly distributed systems. If the tools and libraries (mod_atom +1) become ubiquitous, it ought to work. If on the other hand, it is seen as simply something for reading and writing to blogs, perhaps not.


From: Mark (Jul 25 2007, at 06:18)

"Bigoted"? Chill out, dude.


From: Andrew Phoenix (Jul 25 2007, at 08:58)

"It’s absolutely intolerable, though, that the name of our spec co-editor Bill de hÓra will appear mis-spelled on the RFC because of the IETF’s bone-headed ignorant backward-looking bigoted ASCII-only policy."

Don't hold back, Tim - tell us how you really feel!

Kidding aside, I agree that doing something ASCII-only in an international setting is really... well, the word "stupid" comes to mind, and it's the only thing that makes sense to me. Especially if you are the "Internet Engineering Task Force".

I mean, c'mon; shouldn't they be paving the way to make it easier for everyone to adopt a character set that everyone can use?


From: Sean Gillies (Jul 25 2007, at 11:20)

Unforseen applications of APP? Geospatial/GIS, if I get my way.


From: Mark (Jul 25 2007, at 23:16)

"Doing something ASCII-only in an international setting is stupid."

ASCII's the lowest common denominator, backward compatible. It seems like a good choice.

What about contributors whose names use Greek, Cyrillic, kanji, Hangul? If the spec is in English, the character set might as well be in the core English character set.

I'm using Webkit 3.0 on the Mac, and de Hora is coming out with an accented *capital* O -- sort of proves the point that this stuff is not ready for prime time, I'd say.


From: Tim (Jul 25 2007, at 23:59)

Hey Mark, Bill's last name is "de hÓra". Yes (gasp) it's possible to have a complicated surname that may not have an upper-case letter on the front and may have diacritics on subsequent letters whatever the case. It's a big world. Deal with. If you can. The IETF can't. That sucks.


From: Bill McGonigle (Jul 26 2007, at 13:08)

First, thanks for all your work on this. It looks really nice.

Re: What to Call It? - I agree, 'Atom' has been tied to the feed format, but 'The Atom Protocol' or 'APP' sound rather pedestrian compared to 'Atom'. Perhaps you can come up with an (un)official official name based on the root concept, as a verb, like 'Accelerate', 'Fuse', 'Radiate', 'Split', 'Spin', or 'Smash'.

"OK, I've got to go smash some new blog entries up now."

"Did you spin that new PR piece yet?"

Yeah, yeah, I'm just thinking out loud here.


From: Aristotle Pagaltzis (Jul 26 2007, at 17:48)

> Perhaps you can come up with an (un)official official name based on the root concept

We already have a name. It’s called Atompub.


From: Damian Cugley (Jul 27 2007, at 01:12)

I appreciate your concern over Bill de hÓra’s name being mangled (and of course he won’t have been the first), but I can see why the IETF has been right to be conservative over formats permitted for RFCs. I’d hate it if RFCs from different years were in FrameMaker, compressed PostScript, DVI, LaTeX, Microsoft .DOC, PDF, DocBook, HTML, XHTML, and so on. Plain ASCII with CRLF has its limitations, but it is very ineroperable.

On the plus side, isn’t there a definition of ’Net-Unicode’ being slowly ratified by the IETF or some body or other ( I would hope that, after a suitable period of beard-stroking, the powers that be could allow for Net-Unicode in RFCs, perhaps with a rule that a parallel ASCII version be provided for the next 150 years while we wait for the last of the EBCDIC mainframes to die.

The main issue would be that Microsoft’s recent operating systems generate invalid UTF-8 (they prepend a bogus byte-order mark). I seem to remember that Java’s standard library also had a bad UTF-8 codec; I assume that has been fixed. And Net-Unicode uses CRLF not LF, which doubtless will annoy Unix people. But any browser can display peoplerly tagged UTF-8 character data, so once these small niggles are dealt with it should be plain sailing, no?


author · Dad · software · colophon · rights
picture of the day
July 24, 2007
· Technology (77 fragments)
· · Atom (91 more)

By .

I am an employee
of, but
the opinions expressed here
are my own, and no other party
necessarily agrees with them.

A full disclosure of my
professional interests is
on the author page.