What
 · Technology
 · · Atom

PubHubSunday · Which is to say, It’s Sun­day and I just wired up my lit­tle pub­lish­ing em­pire here to the new hot­ness in Web syn­di­ca­tion tech­nol­o­gy, PubSubHub­bub. If you’re run­ning a hub and you’re not evil, let me know and I’ll ping you ...
[12 comments]  
Serious Syndication · This was turned up by Aaron Swartz: The hundreds-of-billion-dollars-big glob of US stim­u­lus leg­is­la­tion im­pos­es a bunch of guide­lines on re­cip­i­ents, which seems sen­si­ble. Espe­cial­ly sen­si­ble is the one say­ing that those re­ceiv­ing the fund­ing have to do sub­stan­tial re­port­ing, and that the key re­port­ing da­ta be avail­able in a syn­di­ca­tion feed, and I quote: “preferred: Atom 1.0, ac­cept­able: RSS”. I can’t think of any­thing to ad­d.
[4 comments]  
How Fast is Fast? · In prepa­ra­tion for my mod-atom pre­sen­ta­tion at ApacheCon next week, I’ve been run­ning some per­for­mance tests with the help of a gag­gle of client ma­chines rus­tled up by some good peo­ple in the Sun engineering-lab group. The first num­bers are trick­ling in, and I’m a bit at a loss both on what to mea­sure and how to eval­u­ate the re­sult­s. Is 180 POSTs/sec­ond on a T2000 good?
[Up­date: Make that 320/sec­ond; have some bet­ter da­ta pre­sent­ed in a table.]
 ...
[7 comments]  
Inside mod_atom · This is a lengthy note to my­self. I ini­tial­ly want­ed to cap­ture the think­ing that went in­to the con­struc­tion of mod­_atom while it was still fresh in my mind, and dumped out the first dozen or so sec­tion­s. Then as I ex­pand­ed and refac­tored the code, I find that I’m keep­ing this up to date. This most­ly by way of putting it in a place where I won’t lose it. I can write stuff for on­go­ing faster than for any oth­er medi­um, and “On the Net” is a good place not to lose stuff. If mod­_atom even­tu­al­ly gets picked up and used, this may be use­ful to me or any­one else who’s main­tain­ing it; and if it doesn’t, there’ll still even­tu­al­ly be an AtomPub serv­er mod­ule for Apache, and this might be use­ful to who­ev­er builds it. But this is not de­signed to be en­ter­tain­ing or ped­a­gog­i­cal; among oth­er things, it’s in es­sen­tial­ly ran­dom or­der ...
 
Testing REST · I’ve been us­ing RSpec in a way that’s prob­a­bly in­ap­pro­pri­ate, but has got me think­ing about Test-Driven Devel­op­ment and REST ...
[13 comments]  
Build One to Throw Away · This is a max­im from Fred Brooks’ The Myth­i­cal Man-Month. Th­ese days I’m think­ing it’s the sin­gle most im­por­tant les­son there is about soft­ware. It’s been brought rude­ly home to me by my re­cent work on mod­_atom, whose de­sign is ter­ri­bly sim­ple; but I still got the first cut wrong in im­por­tant ways ...
[9 comments]  
It’s Called AtomPub · Re­cent­ly, I was asked for feed­back on some tech­nol­o­gy be­ing built in­side Sun which was said to re­ly on “Atom Pub/Sub”. In re­lat­ed con­fus­ing news, more than one big com­pa­ny has talked about “Rolling out APP”. Brand­ing mat­ter­s. So we took it up on the Atom Pro­to­col mail­ing list and, for what it’s worth, the com­mu­ni­ty of im­ple­men­tors has agreed that we’re all go­ing to re­fer to the pro­to­col spec­i­fied in RFC 5023 as “AtomPub” and noth­ing else. Please co-operate ...
[1 comment]  
Atomic Monday · Here­with some ev­i­dence, for the gen­er­al tech pub­lic, that Atom­pub is a big deal, and for the Atomis­tas, some in­ter­est­ing de­vel­op­ments ...
[6 comments]  
USB and Atom · The last lit­tle while, I’ve been do­ing a lot of in­fras­truc­ture shift­ing, and I’ve re­al­ized that in the world of com­put­er­s, USB is maybe the great­est stan­dard ev­er. I’ve heard ker­nel en­gi­neers sneer­ing and scoff­in­g, but here’s the high­est com­pli­ment I can pay, speak­ing as a fair­ly hard-core com­put­er tech type: I have no idea how USB work­s, and I don’t think I’ll ev­er have to learn. If the plug fits in­to the sock­et, what­ev­er I’m plug­ging in will do what­ev­er I’m ex­pect­ing it to do. I think syn­di­ca­tion ought to be like that, which is why Eric Gar­ri­do (via Bill de hÓra) is right: please stop an­nounc­ing mul­ti­ple syndication-file ver­sion­s, be­cause no­body cares. Pick one and run with it. I’d say pick one that’s stan­dard­ized and sta­ble and de­bugged and mod­ern, but what­ev­er.
[7 comments]  
It’s an Atompub World · To­day, via In­foQ, we see that Mi­crosoft is bet­ting on Atom­pub for Win­dows Live. Since Google and WordPress and big chunks of Ja­va mid­dle­ware are al­ready do­ing it too, this re­al­ly does feel like a band­wag­on. For the last fif­teen years, HTTP has be­come the dom­i­nan­t, over­whelm­ing­ly dom­i­nan­t, ve­hi­cle by which peo­ple and pro­grams get things from the Net. With a lit­tle Atom­pub sea­son­ing, it’ll quite like­ly be the way most things get put back, too. There is a fly in the oint­men­t: tons of server­s, not that many clients. Hel­lo Mo­to? And Nokia and Ap­ple and Sam­sung and all the rest of you? There a kazil­lion, and grow­ing, pub­lish­ing gate­ways out there wait­ing for some­one to start ship­ping hand­helds with an Atompub-powered “Publish” but­ton.
[Up­date:] The Google Con­tacts Da­ta API - Atom­pub based, of course.
[Up­date:] Writ­ing Your First AtomPub Ser­vice with Ab­dera.
[Up­date:] Joe Gre­go­ri­o: Ap­pClien­tTest Up­date and Ap­pClien­tTest - now with unit test­ing good­ness; I par­tic­u­lar­ly like the “drama in HTTP” anal­o­gy.

[3 comments]  
Bejeweled Ape · I am de­light­ed to an­nounce that my col­lab­o­ra­tors (now in­clud­ing Bran­don Mitchell) have been wran­gling the Ape and it is now a Ru­by Gem. Wow.
 
Atomic News · Here­with a few ran­dom Atom­pub point­er­s. Each month that goes by, I’m hap­pi­er with the way Atom­pub came out ...
[5 comments]  
Ape News · There’s been some New-year Ape tin­ker­ing. A few bug-fixes from me and David Calav­era, and al­so I fi­nal­ly rolled in Joe Gregorio’s patch to make it work with Google’s semi-proprietary au­thent voodoo. Last and best, we have a new com­mit­ter, Si­mon Rozet, whose first patch was a Mon­grel adapter, so you can type ruby go-mongrel.rb and there’s your Ape on port 4000. Si­mon will prob­a­bly have checked that in by the time you read this.
 
End of a Chapter · The Atom Pub­lish­ing Pro­to­col, RFC5023, has been pub­lished. The HTML is here, and there’s a Ja­panese trans­la­tion by Ya­mamo­to Yo­hei (who’s ap­peared in this space be­fore). The Work­ing Group has been closed. Good-bye. This piece is some­thing like the bor­ing bit at the end of a con­fer­ence where they say all the thank-yous, but I do have some news and rant­ing to spice things up ...
[4 comments]  
Tab Sweep — Tech · This goes back weeks and week­s; I’ve been wide-finding and do­ing Sun stuff and the Web-watching has suf­fered ...
[6 comments]  
Atomic · Today’s big Atom news is from Joseph Scot­t, who was on the long-weekend shift with Pete Lacey and Sam Ru­by and I, pound­ing WP2.3’s APP-server code in­to shape. He has writ­ten HTTP Ba­sic Authen­ti­ca­tion, A Tale of AtomPub, WordPress, PHP, Apache, CGI and SSL/TLS; the ti­tle is ap­pro­pri­ate­ly long, as this is a very meaty piece that I bet will be read many times in the near to medi­um fu­ture by some­one puz­zled and frus­trat­ed by some com­bi­na­tion of Apache, PHP, and au­then­ti­ca­tion re­quire­ments. Wel­l, yes, the prob­lems popped up in the con­text of Atom­pub and WordPress, but there’s noth­ing spe­cif­ic to the pro­to­col or the pro­duc­t; it’s a big messy ug­ly cor­ner of Web tech­nol­o­gy ...
[1 comment]  
Bad, Feed Readers, Bad! · Piles of junk, I say. Par­don me, but I’m feel­ing grumpy. After much more work than it should have been, mod­_atom is now gen­er­at­ing rea­son­ably co­her­ent (not done yet, but get­ting there) HTML out­put and human-oriented (as op­posed to APP-oriented) Atom feed­s. It’s slight­ly id­iosyn­crat­ic XML, with lots of names­pace pre­fix­es. The Feed Val­ida­tor says it’s OK and I think it’s OK. But none of NetNewsWire or Vi­en­na or Blog­lines [Up­date: or Blog­bridge or Sa­far­i, or My Ya­hoo!, or Sage] can read it cor­rect­ly. I fart in their gen­er­al di­rec­tion. [Up­date: Google Read­er, Plan­et Venus, Snar­fer, Sim­plePie, Lifer­ea, Awa­su, Shrook, and Flock get it right! Good on ya, guys.] [Ah, Brent sent me a point­er to the lat­est be­ta of NetNewsWire 3.1, and it’s fine. I know oth­er peo­ple rave about GoogleRead­er and Vi­en­na and so on, but for me NNW is still way ahead of the pack in let­ting me scan a whole lot of news in al­most no time at al­l.] Are there any oth­er feed-reader im­ple­men­tors out there who think they can, you know, read XML cor­rect­ly?!?! If so, get in touch, and if you pro­cess my lit­tle bun­dle of joy prop­er­ly, I’ll lav­ish praise and links. Or if there’s a bug in the feed that nei­ther I nor the val­ida­tor can see, I’ll apol­o­gize humbly to the whole world. In any case, I’m go­ing to have to go back and patch up the code so it doesn’t emit any of those nasty colons and rel­a­tive URI ref­er­ences that ap­par­ent­ly hurt implementors’ frag­ile feel­ings. This does not im­prove my mood. [Up­date: Just to be clear, I’m not talk­ing about the on­go­ing feed; if you want to test your feed read­er, con­tact me and I’ll point you at the test feed.]
[17 comments]  
Autumnal mod_atom · I just did a mas­sive check-in on mod­_atom, and it’s now not just an Atom Store, it’s al­so a ba­sic blog pub­lish­er. This frag­ment is about how it work­s, and in­cludes a con­fes­sion; I did one fair­ly aw­ful thing along the way ...
[6 comments]  
The Sunday Contests · Who needs the NFL? The 0.00001% of the pop­u­la­tion who think Syn­di­ca­tion Se­man­tics and Web Ar­chi­tec­ture are all about fun ought to cruise by Sam Ruby’s space and take in the dis­cus­sion around One More Step For­ward?
 
Long-Weekend Fun · So, hey, it’s the last week­end be­fore school starts and ev­ery­thing gets re­al again. So we should be kick­ing back, right? As in, work­ing on Atom Pro­to­col and WordPress ...
[4 comments]  
Tab Sweep — Tech · There were a few here that I want­ed to do a whole piece about, but I just got­ta clear some space on these deck­s. To­day we have push­ing and pulling and queue­ing and Ruby.next and Ja­va hate and PHP-vs.-Rails. What’s not to like? ...
[9 comments]  
Branding the Ape · The Ape has a new lo­go, cour­tesy of Greg Boren­stein. Ain’t it cute? ...
[1 comment]  
Tidying HTML · I’ve de­cid­ed that mod­_atom re­al­ly needs to be a blog-publishing sys­tem, not just an Atom Store. And fur­ther­more, based most­ly on the com­ments to that San­i­ta­tion piece, I’ve made two de­sign de­ci­sion­s. First, the san­i­tiz­ing hap­pens on­ly on the HTML out­put; the Atom-store part will per­sist the da­ta as close as pos­si­ble to the way it was sent up­stream. Se­cond, I’m go­ing to try us­ing the TidyLib pars­er to pick apart type="html" text con­structs so I can clean ’em up ...
[4 comments]  
Welcome David · The Ape (runnable, source) has a new com­mit­ter: David Calav­era, who’s from Spain. He wrote me out of the blue say­ing “Here are some patch­es that are the be­gin­ning of an RFC2617 framework.” David need­ed the Ape to talk WSSE for his own work ...
[3 comments]  
Sanitation · It’s amaz­ing how is­sues float to the top of mul­ti­ple minds in­de­pen­dent­ly. I’ve been spend­ing a lot of time think­ing about how to san­i­tize to-be-published data. Then Rob Sayre wrote In­ter­op­er­abil­i­ty and XSS Mit­i­ga­tion; XSS stands for “cross-site scripting”, the main threat that you san­i­tize to avoid. Sam Ru­by no­ticed got ac­tive: In­ter­op­er­abil­i­ty and XSS Mit­i­ga­tion an­nounced the San­i­ti­za­tion rules wiki-space. Microsoft’s Joe Cheng is wor­ry­ing, too ...
[12 comments]  
Tab Sweep — Tech · To­day we have some Atom­ic Ap­ple love, iPhone Web friend­li­ness, Re­laxNG praise, and JVM Lan­guage widen­ing ...
[6 comments]  
1½ Days of APP Interop · Check the re­sults. This makes me very glad, as did the hours on the IRC, try­ing code, patch­ing code, peer­ing at log­files, laugh­ing at each oth­er... mak­ing it work. This is how the Net gets bet­ter. Like I said last time, Win­dows Live Writ­er is da bom­b, oth­er peo­ple who are in the blogging-client biz bet­ter watch out ...
[8 comments]  
Atomic News · We had a lit­tle on­line Atom Pro­to­col in­terop ses­sion to­day, via IRC and servers scat­tered here and there about the globe. I got mod­_atom to work with Apache Ab­dera and Win­dows Live Writ­er. With a cou­ple of lit­tle glitch­es in each case, fixed on the spot (that’s what the ses­sion is for), we had 100% squeaky-clean in­terop. Earth to client tool­mak­er­s: pay at­ten­tion, or you’re gonna be eat­ing WLW’s dust ...
 
Ape Rev · I just post­ed a rev for the Ape (runnable, source), as fol­lows: ...
[6 comments]  
mod_atom Status · Peo­ple who are in­ter­est­ed in the soft­ware shouldn’t have to read the acres of prose in the mod­_atom in­tro, so I’ll just keep this one up to date ...
[3 comments]  
That’s All, Then · 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 ...
[12 comments]  
mod_atom · This is a stripped-down im­ple­men­ta­tion of the serv­er side of the Atom Pub­lish­ing Pro­to­col as an Apache mod­ule, im­ple­ment­ed in C. It felt like some­thing that need­ed to ex­ist and I am better-qualified for this par­tic­u­lar chore than your av­er­age geek; hav­ing said that, I have no idea if any­one ac­tu­al­ly needs such a thing. mod­_atom ac­tiv­i­ty can be tracked on this blog, for now, here. If any in­ter­est de­vel­op­s, then I’ll trans­fer dis­cus­sion to a blog at mod-atom.net which will be driven, of course, by mod­_atom ...
[17 comments]  
Trouble for Atom · Yep, ladies and gen­tle­men, it looks like there’s trou­ble on the hori­zon. On the RFC4287 syndication-format fron­t, it may have been sta­ble since 2005 and wide­ly de­ployed, but watch out, there’s a new ver­sion of RSS 2.0! (2.0.9, to be pre­cise). RSS 2.0 is sort of RFC4287’s main com­pe­ti­tion, and if there are two dif­fer­ent spec­s, I guess that must mean it’s twice as good. On the Atom-Protocol side, Google’s John Panz­er has made a shock­ing dis­cov­ery, and I quote: “There seems to be a com­plaint that out­side of the tiny cor­ner of the Web com­prised of web pages, news sto­ries, ar­ti­cles, blog post­s, com­ments, lists of links, pod­cast­s, on­line pho­to al­bum­s, video al­bum­s, di­rec­to­ry list­ings, search re­sult­s, ... Atom doesn't match some da­ta models.” Wel­l, it was fun while it last­ed.
[8 comments]  
Interop Impressions and Pix · We worked from noon till six Mon­day, and from 8:30 till the mid­dle of Tues­day; a lit­tle over twen­ty peo­ple in the room. I got the Ape talk­ing to a few and failed with a cou­ple of oth­er­s, but in those cas­es the prob­lems were im­ple­men­ta­tion glitch­es, not the pro­to­col. Sur­pris­es? I saw a cou­ple of servers that didn’t ac­cept Atom en­tries, just var­i­ous kinds of me­dia ob­ject­s. OK, I guess. Pleas­ant sur­pris­es: get­ting pret­ty well com­plete in­ter­op­er­a­tion with (on al­most the first try) Nikunj Mehta’s code and (after a bit of work) Kevin Beyer’s and James Snell’s. I’m run­ning a few shots of the event just be­cause I like tak­ing pic­tures of peo­ple. I’ll write an­oth­er piece draw­ing some tech­ni­cal con­clu­sion­s ...
[1 comment]  
Atompub Interop Lessons · The re­sults are sum­ma­rized on the Wi­ki; what do they mean? Ob­vi­ous­ly, good news: The fact that peo­ple from this many places, most of whom had nev­er met be­fore, got to­geth­er and were able to put that many check-marks on the grid, based on a pro­to­col whose de­sign is not quite frozen, verges on the mirac­u­lous. Now let’s dive a lit­tle deep­er; what does each check-mark mean, and what can im­ple­men­tors bet on, and when? ...
[2 comments]  
Atom Publishing Protocol Interop! · Mark your cal­en­dar: April 16-17 at Google. Every­body is in­vit­ed, pro­vid­ed they bring along an APP im­ple­men­ta­tion, client or server. This was just an­nounced a cou­ple of days ago, and as I write this there are al­ready six thir­teen client and sev­en fif­teen serv­er im­ple­men­ta­tions signed up to be there and try to fill in the grid. Let’s drop some names, in al­pha­bet­i­cal or­der: AOL, Flock, Google, IBM, Lo­tus, Mi­crosoft, NTT, Or­a­cle, O’Reilly, Six Apart, Sun, WordPress. Um, have I men­tioned that the APP is go­ing to be huge?
[6 comments]  
Rewriting my Gravestone · If I had fall­en over dead any time be­tween about 1999 and last year, my head­stone and any obit­u­ar­ies would have said “Computer geek who was in the room when XML was invented.” I’m be­gin­ning to think that the Atom Pub­lish­ing Pro­to­col (APP) has a chance to crowd XML off the head­stone ...
[2 comments]  
Aping · To­day Joe Gre­go­rio point­ed the Ape at his “apptestsite” and found some prob­lem­s. And the Ape found some prob­lems with his stuff too. So we both fixed ’em and the Ape’s much hap­pier. This is fun! ...
[14 comments]  
Ape News · It’s been lev­eled up to the Protocol-13 draft lev­el and the source is avail­able. While I’m still mas­sive­ly un­sat­is­fied, the Ape as it stands to­day is ac­tu­al­ly pret­ty use­ful ...
[5 comments]  
APP-12 · That would be the twelfth draft of the Atom Pub­lish­ing Pro­to­col spec. We’ll be ask­ing for IETF last call on this draft. I’m pret­ty sure that the fi­nal prod­uct will look about like this. I wouldn’t be sur­prised if the IETF pro­cess patched a cou­ple of se­cu­ri­ty over­sights or un­cov­ered a cou­ple of cor­ner cas­es; nor would I be sur­prised if it didn’t. But I think that im­ple­men­tors who run with this will be pret­ty safe; mind you, there are a lot out there who didn’t wait this long; and they de­serve our thanks. As do the ed­i­tors and the good peo­ple in the Work­ing Group; this has been most­ly a pret­ty good trip. [Up­date: Some Atom-protocol news I had hang­ing around wait­ing to blog: Dave John­son links to sev­er­al im­ple­men­ta­tions, and Elias Tor­res tells of a quick­ie.]
[3 comments]  
Atom News · We’re up to Draft 10 of the Atom Pub­lish­ing Pro­to­col, and it’s very close to what the fi­nal prod­uct will look like. The APP is get­ting more ear­ly de­vel­op­er sup­port and in­terop test­ing than any stan­dards project I’ve ev­er been in­volved with, and that in­cludes XML. DeWitt Clin­ton cov­ers some of the bases, but Dave John­son makes it nice and sim­ple: “Atom pro­to­col can do ev­ery­thing that Me­taWe­blog API can do, and much more.” On the sub­ject of the Atom Feed For­mat, I don’t think much more needs to be said, but the U.S. In­tel­li­gence com­mu­ni­ty (which is quite feed-heavy in­ter­nal­ly) com­mis­sioned a study: RSS and Atom Con­sid­er­a­tions; it seems very thor­ough to me. In read­ing it, Wikipedia’s ar­ti­cle on In­telink may be a use­ful ref­er­ence.
 
“Publish” Everywhere · Atom, es­pe­cial­ly the Pub­lish­ing Pro­to­col, is mov­ing in­to the evan­ge­lism stage, since the com­mit­tee work’s al­most done and there are plen­ty of im­ple­men­ta­tion­s. Here’s why I think Atom is im­por­tan­t. If you look at Dave Sifry’s num­bers, the growth of the bl­o­go­sphere is as­tound­ing; look­ing past all the fluff and chaf­f, there are some­where be­tween five and ten mil­lion peo­ple out there who care enough to con­tribute to the Net once a week or bet­ter. That may sound like a lot, but I don’t think it’s near­ly enough. Here at Sun, in a blogging-friendly tech-savvy cul­ture, maybe 5% of the peo­ple post reg­u­lar­ly. So I look at the num­ber of peo­ple us­ing the Net and I won­der: “Why there aren’t 50 mil­lion, in­stead of five mil­lion, peo­ple con­tribut­ing ev­ery week?” The an­swer: “Because it’s too hard”. We can fix that. Here’s the Atom dream: A “Publish” but­ton on ev­ery­thing. On ev­ery word pro­ces­sor and email read­er and web brows­er and cell­phone and PDA and spread­sheet and photo-editor and digi­cam and out­lin­er and sales-force track­er. Real­ly, ev­ery­where. If it doesn’t have a “Publish” but­ton, it’s bro­ken.
 
Johnson on Feeds · Dave John­son gave a talk this morn­ing at a lo­cal XML in­ter­est group. His slides (PDF) are the sin­gle best in­tro­duc­tion and overview I’ve ev­er seen about feeds and syn­di­ca­tion and RSS and Atom and all that stuff.
 
Upcoming Gig: OSCON · What with the re­cent fam­i­ly en­large­ment, it wasn’t clear that I was go­ing to be able to get to OSCON. But the girl’s set­tling down pret­ty well, so I de­cid­ed to go. I re­viewed the sched­ule and was hor­ri­fied to see noth­ing about the Atom Pub­lish­ing Pro­to­col, which I’m in­creas­ing­ly con­vinced is hot stuff. So I sent a note off to Nat Tork­ing­ton say­ing “WTF, no APP?!?” and he wrote back say­ing “So sub­mit a pa­per already” and I did and now I’m on the pro­gram. I think I can on­ly go for two days, but Portland’s a great town for beer and book­stores, which there’s more to life than, but on­ly mod­er­ate­ly. [Up­date: I’m ar­riv­ing at PDX at 4PM on Tue 25th. I can’t be the on­ly one, shoot me a line and let’s or­ga­nize some taxi shar­ing, it’s a long bor­ing ride in from the air­port.]
 
Feed Format Kitten Fight · Like your syn­di­ca­tion pol­i­tics tasty and fresh? Head over to DeWitt Clinton’s Un­to.net and read On RSS and Atom. Clinton’s at A9, Amazon.com’s Sil­ly Val­ley search-wonk caul­dron, and his stuff keeps com­ing across my radar in re­cent week­s. Any­how, he has what seems to me a clear-eyed and dis­pas­sion­ate eval­u­a­tion of the feed-format choic­es fac­ing im­ple­men­tors these days. There is one place he gets it back­ward, say­ing: “I’ve been con­sis­tent­ly im­pressed with how well the au­thors of the Atom syn­di­ca­tion for­mat an­tic­i­pat­ed the needs of the ad­vanced con­tent syn­di­ca­tion community.” No; Atom’s de­sign re­flects the back­ward-looking ex­pe­ri­ence we got in the last few years of work­ing with RSS; it turns out that the fu­ture is some­what like the past. But don’t stop when you get to the end of DeWitt’s piece, there are dozens of com­ments, most of them in­struc­tive, com­ing at the is­sue from all sorts of di­rec­tion­s. Scoble pushed back at length, fol­low the point­er from his com­men­t. Some­one who signs him­self “Raja” has an aw­ful­ly fa­mil­iar style. And a fi­nal note: when Mr. Clin­ton talks about XML, for ex­am­ple an RSS <description> el­e­men­t, he says <description/>. Now, that’s the kind of pedantry I can re­late to.
 
Boundaries · “This town was named af­ter a mi­nor Dos­toyevsky character...” is the be­gin­ning. The Land­scape — Mar­fa, Texas Pt. 1 is by David Byrne and it’s re­al­ly a must-read, rang­ing through land­scape and mu­sic and bound­aries of var­i­ous kind­s. By the way, Mr. Byrne needs an Atom feed; re­cent­ly some­thing changed and in my news­read­er, his pieces are full of raw HTML markup and sans im­ages.
 
With Bloglines to Atom · A few days back I not­ed ap­prov­ing­ly that Blog­lines was work­ing on their long-broken Atom 1.0 han­dler; and that there were a still a few relative-URI prob­lem­s. I got a puz­zled, po­lite email from a Blog­lines en­gi­neer say­ing “Uh, are you sure? I don’t see that, might have been switchover artifacts.” I went and looked and sure enough, they were gone. I did see one lit­tle Keith-and-the-roaches bug with a stray “&amp;” so I wrote back say­ing “fantastic, great work, oh BTW you’ve got a stray ampersand”. Within a few min­utes he wrote back “Just fixed the &amp; is­sue. It might not show up on pro­duc­tion un­til later.” This is the way the Web is sup­posed to work. (Sam Ru­by tells me there are some lin­ger­ing corner-case bugs; re­port ’em and I bet they’ll fix ’em.) As of now, I am ab­so­lute­ly rec­om­mend­ing Blog­lines to all news­read­er new­bies as a good place to start; and al­so to any­one, new­bie or not, who doesn’t want to deal with the fuss and both­er of a sep­a­rate news­read­ing pro­gram. And with Bloglines’ switch, ev­ery ma­jor piece of in­fras­truc­ture that I know of is now Atom 1.0-capable. So I just per­ma­nent­ly redi­rect­ed my RSS feed to the Atom 1.0 ver­sion. If this looks weird in your news­read­er, please do let me know; and more im­por­tan­t, file a bug so your read­er gets fixed. Once this set­tles down, I look for­ward to tak­ing the axe to a whole bunch of double-escaping and RSS-writing code.
 
Atom Newsreel · I work on lots of dif­fer­ent things, but Atom is the most im­por­tan­t, don’t you for­get it. Item: Atom 1.0 isn’t even a year old and it al­ready has its first stan­dard­ized ex­ten­sion, for feed thread­ing; an­nounce­ment chez James Snell, who did most of the work. Item: Blog­lines has re­sumed work on their Atom 1.0 parser, and it’s con­sid­er­ably less bro­ken than it used to be. You can ac­tu­al­ly use it to read my Atom feed and it no longer mess­es up the white space. The relative-link han­dling still needs work, though; my pic­tures don’t show up, and some in­ter­nal links are bro­ken (not al­l... puz­zling). Any­how, good on ya guys, this is the right di­rec­tion. Item: The work­ing group just pub­lished Draft 9 of the pro­to­col (of­fi­cial and HTML ver­sion­s). My opin­ion is that noth­ing in here will change much, and there will be a cou­ple of very small ad­di­tion­s, and then we’ll go for IETF last cal­l. For Atom­ic pedants, the change here is what hap­pens when you post some­thing that isn’t an En­try, like for ex­am­ple a movie or a pic­ture.
 
Title Pain · James Hold­er­ness, a guy who re­al­ly knows his shit about syn­di­ca­tion tech, has been do­ing some torture-testing; see En­cod­ing RSS Ti­tles, which shows that if you want to do some­thing as ob­vi­ous as men­tion­ing “AT&T” in your ti­tle, you’re in deep RSS doo-doo. (Did I say tor­ture test? James blogs at www.詹姆斯.com; the boy’s got at­ti­tude.) Any­how, James es­tab­lish­es that there’s es­sen­tial­ly no safe way to do this. Quot­ing him: “Clearly if you want to sup­port Fire­fox or In­ter­net Ex­plor­er you’ve got no choice but to use the sin­gle en­cod­ing op­tion. For cer­tain strings, though, that would mean los­ing sup­port for at least twen­ty oth­er aggregators.” Yow. So I emailed James, ask­ing “Would it be over­sim­plis­tic to say: ‘Thus, use Atom 1.0?’” He wrote back “Somewhat. While Atom doesn't have the am­bi­gu­i­ties of the RSS spec, it has all the same prob­lems with bug­gy clients.” Fair enough. But I think that James proved that, with RSS, you can’t solve the prob­lem even in prin­ci­ple. With Atom, you can. Which seems like a de­ci­sive ar­gu­men­t, to me. [Up­date: Oh hel­l, James’ Chi­nese URI broke some­thing in the on­go­ing front-page gen­er­a­tor... un­til I’ve fixed it, use this.]
 
Blogging Cam: Almost Right · Via Niall Kennedy (who’s been very good late­ly): Mi­crosoft cam­eras that have one-button pub­lish­ing to MSN Spaces. This is so wrong. I don’t want to see a fu­ture in which your cam­era is LiveJournal-enabled or Facebook-ready. When I get a com­put­er or a mo­bile de­vice, it’s ei­ther Web-ready or not, it doesn’t have to be MSN-enabled or Yahoo-blessed. This is why we need the Atom pro­to­col. If your phone or your cam­era or your any­thing is Atom-enabled, then it will work with any pub­lish­er who sup­ports the pro­to­col, no spe­cial deals re­quired. Com­ing soon.
 
Atom Newsreel · I’ve been ac­cu­mu­lat­ing things Atom­ic to write about for a while, so here goes. Item: You’ll be able to blog from in­side Mi­crosoft Word 2007 via the Atom Pub­lish­ing Pro­to­col. Item: Sam Ru­by has wran­gled Plan­et to the point where it han­dles Atom 1.0 prop­er­ly. Item: Along the way, Sam re­port­ed a com­mon bug in Atom 1.0 han­dling, and his com­ments show it be­ing fixed all over (Plan­et, MSN, and Google Read­er, but not Blog­lines of course); the Kei­th ref­er­ence in Sam’s ti­tle is to this. [Up­date: Gor­don Weak­liem ex­tir­pates an­oth­er com­mon bug from the NewsGa­tor uni­verse.] Item: The Mov­able Type Feed Man­ag­er is based on James Snell’s pro­posed Thread­ing Ex­ten­sions to Atom 1.0; Byrne Reese seems to think that par­tic­u­lar ex­ten­sion is hot stuff. Item: Na­ture mag­a­zine is ex­tend­ing Atom 1.0 for their Open Text Min­ing In­ter­face. Item: The Google Da­ta APIs are old news now, but it looks like they’re do­ing Atom 1.0 and play­ing by the rules. Last Item: Over in the Atom Work­ing Group, we’re get­ting very close to declar­ing vic­to­ry and go­ing for IETF last call on the Pro­to­col doc­u­men­t.
 
It’s Odd · I’m quot­ing Stephen Dun­can Jr, de­scrib­ing the fact that Blog­lines would rather try to fool peo­ple who sub­scribe to my Atom feed by switch­ing in my RSS feed, as op­posed to just fix­ing their stupid­ly bro­ken Atom 1.0 han­dling. “Odd” is one word for it. Coun­ter­mea­sures are ap­pro­pri­ate and if I have to I’ll take them, but wouldn’t it be so much bet­ter to just, like, you know, im­ple­ment the In­ter­net Stan­dards? Nobody’s ask­ing any­one to stop pro­cess­ing RSS, but the world (un­like Blog­li­nes) didn’t freeze sol­id in 2003.
 
Atomic Monday · First of al­l, im­ple­men­tors of any­thing Atom-related need to spend some time chez Jac­ques Distler; in par­tic­u­lar, the con­ver­sa­tion that plays out in the com­ments. Se­cond, there’s this piece of soft­ware called Plan­et Plan­et that al­lows you to make an ag­gre­gate web page by read­ing lots of feed­s; for ex­am­ple, see Plan­et Apache or Plan­et Sun. Sam Ru­by de­cid­ed that its Atom sup­port need­ed some work, so he did it. Now, here’s the ex­cit­ing part: he pinged me over the week­end and said “Hey, look at this” want­ing to show me his cleverly-Atomized Plan­et In­ter­twing­ly feed. I looked at it in NetNewsWire and was puz­zled for a mo­men­t; some but not all of the things in the feed were high­light­ed as un­read, even though this was the first time I’d seen it. Then the light went on. This is Atom do­ing ex­act­ly what we went to all that trou­ble to make it do. NetNewsWire has good Atom sup­port and, be­cause Atom en­tries all have unique IDs and times­tamp­s, it can tell that it’s seen lots of those en­tries be­fore in oth­er feeds that I sub­scribe to. That’s how I found Jacques’ piece. This is huge; any­one who us­es syn­thet­ic or ag­gre­gat­ed feeds knows that dupes are a big prob­lem, show­ing up all over the place. No longer, Atom makes that prob­lem go away.
 
Atomic Google Hacks · Check out Mi­hai Parparita’s Google Read­er Tid­bits, about how he used Google Read­er hacks to do a bunch of clever feed splic­ing. The ar­ti­cle is in­ter­est­ing, and I think Atom is go­ing to en­able a bunch of feed-mashup cre­ativ­i­ty that I’m not smart enough to in­ven­t. But I want­ed to do a deep-dive on the ac­tu­al Atom feed he gen­er­at­ed, which is prob­a­bly of in­ter­est on­ly to ob­ses­sive Atom 1.0 fetishist­s ...
 
Atom as a Case Study · This is adapt­ed from my talk of the same name at ETech 2006. The talk’s sec­tions were en­ti­tled Why?, How?, What?, and Les­son­s?; I’ve left out What?, the de­scrip­tion of what Atom is, since we’ve had plen­ty of that around here. That leaves Why we built it, How we built it, and what Les­sons you might want take away from the ex­pe­ri­ence ...
 
APP Test Suite · No spec­i­fi­ca­tion or stan­dard is re­al­ly ready for prime time un­til it has a test suit­e. The Atom Pub­lish­ing Pro­to­col al­ready has one, and it hasn’t even shipped yet. Way to go!
 
Validated · For some time, the Feed Val­ida­tor has been mut­ter­ing warn­ings about my Atom 1.0 feed, hav­ing to do with my ag­gres­sive use of xm­l:base and rel­a­tive URIs. I fixed that up, and man­aged to con­vince Sam Ru­by that one of the warn­ings (due to a redi­rect from tbray.org to www.t­bray.org) was bo­gus. So the validator’s hap­py. I al­so took a lit­tle ex­tra trou­ble to pret­ti­fy the feed, and now I think it’s prob­a­bly use­ful as an ex­am­ple of a fair­ly full-featured and ultra-squeaky-clean Atom 1.0 feed; some­what human-readable, even. Have a look at my feed and Sam’s side-by-side, his is very read­able too, there are some points where the dif­fer­ence be­tween Sam’s choic­es and mine are in­struc­tive. Any­one have a sug­ges­tion on how my feed might be im­proved, to serve even bet­ter as an ex­am­ple?
 
Atom Bits & Pieces · FeedBurner’s FeedFlare API looks in­ter­est­ing. (Ac­tu­al­ly, there are a ton of Web APIs out there look­ing in­ter­est­ing, no­body could keep up­.) But fol­low that link above and search for “Atom”; they’re ab­stract­ing the uni­verse of feed for­mats us­ing Atom 1.0 to give you some­thing con­sis­tent and use­ful to write XPaths again­st. Microsoft’s do­ing the same thing in the IE7 API, but they’re ab­stract­ing to a bas­tardized not-quite-RSS2.0 for­mat they made up them­selves, which is kind of a con­fus­ing thing to do and will make the world learn yet an­oth­er mod­el. Atom’s pop­ping up over in Ning-land too, they’re us­ing the for­mat to ex­pose their ap­pli­ca­tions and data, and are work­ing on mak­ing it all write­able via the Atom Pub­lish­ing Pro­to­col (APP). And at ETech I heard about a cou­ple more gueril­la APP im­ple­men­ta­tion­s, at big smart well-known out­fit­s, not com­pre­hen­sive or in­tend­ed for ex­po­sure (yet) to the world, just cooked up in a hur­ry to get some­thing done that need­ed do­ing. My spidey sense is tin­gling.
 
What Dave Said · I’d just like to high­light Dave Winer’s re­marks to­day on the RSS 2.0 roadmap. No­body can say Dave hasn’t been 100% crystal-clear about this. Those of us who thought there was ba­sic, im­por­tant work that still need­ed do­ing in the area of syn­di­ca­tion for­mats had three choic­es; RDF-wrangling in the RSS 1.0 con­tex­t, namespace-wrangling in the RSS 2.0 con­tex­t, and putting a new name on it; to use Dave’s word­s, “make a new for­mat as an evolution”. Thus, Atom. We don’t have to agree about ev­ery­thing, but if the In­ter­net de­pends on any­thing, it de­pends on stan­dards that are sta­ble, and the frame­work of trust around that sta­bil­i­ty.
 
Why We Need Atom Now · Check out Mozil­la Bug #313441. Lots of juicy stuff: se­cu­ri­ty risks, open source good­ness, RSS 2.0 am­bi­gu­i­ties bleed­ing down in­to RSS 1.0. Blog­lines be­ing, uh, a lit­tle slow to catch up. And Atom be­ing the so­lu­tion. My fa­vorite quote: “If you need to use the char­ac­ter ‘<’ in a feed ti­tle, which Bugzil­la ab­so­lute­ly does, you have ex­act­ly three choic­es: be in­valid and work, be valid and fail, or, the *on­ly* re­al choice, use Atom instead.” It works for some oth­er peo­ple who re­al­ly care about se­cu­ri­ty, too. But maybe the security’s just a sideshow; the re­al ben­e­fit of mov­ing to Atom would be to avoid the an­nu­al RSS food-fight.
 
He’s Back! · Mr. Safe, I mean. I in­tro­duced him in 2003 (scroll down to “Houston, We Have a Problem”) and lots of oth­ers seemed to en­joy invit­ing him to their blogs for a chat about syn­di­ca­tion pol­i­tic­s. For those for­tu­nate enough not to have no­ticed, there’s a sil­ly brouha­ha go­ing on in the RSS world. Check out Rogers Caden­head, Shel­ley Pow­ers, and Dave Win­er. Sam Ru­by had al­ready pub­lished Toss­ing Ros­es, which I thought was kind of po­lite and help­ful but re­al­ly ir­ri­tat­ed Steve Gill­mor. What a mess... clear­ly, the man for the job is Mr. Safe, so Sam brought him back.
 
Atomic MSDN · Hey, MSDN blogs have start­ed gen­er­at­ing Atom 1.0; here’s the IEBlog feed. Good stuff! There are a cou­ple of lit­tle glitch­es: they use rel="self" in­stead of rel="alternate", and they pro­vide pub­lished but not up­dat­ed times­tamp­s. Both to­tal­ly for­giv­able in a 1.0 re­lease, and pre­sum­ably easy to fix. Take-away: we need to have bet­ter tu­to­ri­al ma­te­ri­al (others have made the same self/al­ter­nate mis­take), and to do bet­ter at telling the world about the Feed Val­ida­tor.
 
Hey Thanks, Firefox! · Via Rob Sayre, word that Fire­fox 1.5.0.1 (will auto-update) will con­tain the fix for Bug 262222, which kept Atom 1.0 feeds that use xm­l:base and rel­a­tive links from work­ing, no­tably in­clud­ing this one. We do make pro­gress. Now if Blog­lines would just get a clue...
 
RFC 4287 · I hadn’t seen the an­nounce­men­t, but this looks like a sta­ble of­fi­cial IETF link to RFC 4287, The Atom Syn­di­ca­tion For­mat. A lit­tle more work and we’ll have the pub­lish­ing pro­to­col done and I can re­turn to my plow (or equiv­a­lent). The work of the WG and ed­i­tors was just out­stand­ing, and the IETF did, as ad­ver­tised, pro­vide a use­ful quality-control pro­cess with­out un­du­ly get­ting in the way. Thanks ev­ery­one. The world now has a general-purpose syn­di­ca­tion for­mat that is smal­l, sta­ble, based on the last decade’s lesson­s, clean, and wide­ly im­ple­ment­ed. I feel hap­py.
 
Atom Status · The Atom Syn­di­ca­tion For­mat is done, cast in stone, will get an RFC num­ber as soon as the appallingly-slow RFC-Editor pro­cess con­cludes. The Atom Pub­lish­ing Pro­to­col is very close to done; here­with an overview of how it works and what still needs to be set­tled ...
 
Millions · Tex­tPat­tern, Mov­able Type have been do­ing Atom 1.0 for a while; now it’s time to wel­come LiveJour­nal on board. We’re get­ting in­to some big num­bers here. More to come.
 
Protocol Day One · It was a mild­ly his­toric day; a few of us got to­geth­er on IRC and tried to in­ter­op­er­ate some ear­ly Atom Pub­lish­ing Pro­to­col im­ple­men­ta­tion­s. The on­ly serv­er on the air was from Dave John­son. We got two dif­fer­ent clients to talk to it; one was a Big Se­cret Pro­ject from a Big Fa­mous Com­pa­ny based on all sorts of slick in­fras­truc­ture. Mine was curl. I man­aged to in­tro­spect the server, list the en­tries, cre­ate a new en­try, up­date it, and delete it. I screwed up the mes­sag­ing a few times, and Dave’s serv­er on­ly blew chunks about half of them. Those who know what curl is are prob­a­bly snick­er­ing now. But I think the fact that you can de­bug a non­triv­ial ap­pli­ca­tion with curl -X -i -d -H is a sig­nif­i­cant weapon in the quiver of RESTa­far­i­an­s. Let’s see ya do that with your SOAP + WSDL + WS-Policy + WS-Addressing + WS-MetadataExchange + WS-ReliableMessaging ap­p. Ac­tu­al­ly, the big take-away isn’t that, it’s that the Atom pro­to­col is sim­ple and easy to im­ple­ment and ro­bust. The world needs some­thing like this. Later: Hey, more Atom Pro­to­col stuff from Joe Gre­go­rio over at XML.­com.
 
Unatomic Bloglines · Sev­er­al peo­ple have writ­ten to tell me that the on­go­ing Atom 1.0 feed is thor­ough­ly borked in blog­lines, most vis­i­bly in white-space and link han­dling. I sent them a note a cou­ple of weeks ago, but no fix yet. Could some­one at blog­lines please have a look?
 
Atom 1.0 · It’s cooked and ready to serve. It doesn’t have an RFC num­ber yet, but this is of­fi­cial­ly Atom 1.0 (HTML here). Here’s a com­par­i­son of RSS 2.0 and Atom 1.0. The im­ple­men­tors are hard at work. I’ll up­date this page reg­u­lar­ly with Atom-related news and point­er­s, send word if you want yours in­clud­ed. Read on for crunchy Atom­ic good­ness. [Up­date: NewsGa­tor On­line has 1.0 sup­port­.] ...
 
Pop the Champagne! · The Atom 1.0 da­ta for­mat is an IETF stan­dard. A “proposed standard” ac­tu­al­ly, but pro­posed stan­dards are first-class IETF cit­i­zens. In a few month­s, the slow-grinding wheels of the RFC Edi­tor will emit an ac­tu­al RFC num­ber, but the of­fi­cial text is echoed (in HTML, too) here. The fu­ture is un­fold­ing.
 
Atom API Sketches · I’m think­ing about Atom 1.0 from the coder’s point of view. I’m not think­ing about the Pub­lish­ing Pro­to­col, I’m think­ing about how you, the pro­gram­mer, should go about in­hal­ing and ex­hal­ing the stuff. I’ve nev­er be­lieved in One True API for XML, it’s just too broad-spectrum, but Atom’s pret­ty tight­ly con­strained. Ob­vi­ous­ly, you can use some­thing gener­ic like SAX or one of the many DOM-style APIs, or one of the mod­ern pull APIs. Maybe for Atom we could use some­thing sim­pler and more nat­u­ral. I’m think­ing out loud in this space, this is far from fin­ished, not even a pro­pos­al yet. But, I bet there are oth­er peo­ple out there who care ...
 
Atomic RSS · Sup­pose you’re gen­er­at­ing an RSS feed, or you’re think­ing about gen­er­at­ing an RSS feed, and you’re won­der­ing how Atom fits in­to the pic­ture. The fu­ture of tech­nol­o­gy is hard to pre­dic­t, but there’s a good way to hedge your bet­s. You can gen­er­ate an RSS feed and, by fol­low­ing a few sim­ple rules, be re­al­ly sure that there’s a 100%-equivalent Atom 1.0 feed, so that if you’re gen­er­at­ing both, they’ll be in sync, and if you need to switch back and forth, it’s just a mat­ter of chang­ing a few strings. Let’s call this future-proofed fla­vor “Atomic RSS”. It turns out that us­ing Atom­ic RSS is a Good Thing any­how, be­cause it will help soft­ware in gen­er­al and news ag­gre­ga­tors in par­tic­u­lar pro­duce bet­ter re­sult­s. [Ed. note: This is a first cut, and prob­a­bly has er­rors; cor­rec­tions grate­ful­ly re­ceived. Thanks to Sam Ru­by for sub­stan­tial con­tri­bu­tion­s.] ...
 
The Atomic Tribe · I just want to say, the Atom­pub work­ing group has been out­stand­ing. Not al­ways po­lite, but in­tense and hard-working and in­sight­ful. Not self­ish, ei­ther; we had to go prompt the mailing-list con­trib­u­tors in­di­vid­u­al­ly ask­ing them if they want­ed to be in the ac­knowl­edg­ments sec­tion. The Net owes a thank-you to: Dan­ny Ay­ers, James Aylett, Roger Ben­ning­field, Arve Bersvend­sen, Dan Brick­ley, Thomas Broy­er, Robin Cover, Bill de hÓra, Martin Duerst, Roy Field­ing, Joe Gre­go­rio, Bjo­ern Hoehrmann, Paul Hoff­man, Anne van Kesteren, Brett Lind­s­ley, Dare Obasan­jo, David Or­chard, Aris­to­tle Pa­galtzis, John Panz­er, Gra­ham Parks, Dave Paw­son, Mark Pil­grim, David Pow­ell, Ju­lian Reschke, Phil Ring­nal­da, An­tone Roundy, Sam Ru­by, Eric Scheid, Brent Sim­mons, Hen­ri Sivo­nen, Ray Slakin­s­ki, James Snell, Hen­ry Sto­ry, Asbjørn Uls­berg, Wal­ter Un­der­wood, Nor­man Walsh, Dave Win­er, and Bob Wy­man.
 
1997 and 2005 · Back in 1997, when the shape of XML start­ed to sta­bi­lize, I sort of ap­point­ed my­self Chief Mar­ket­ing Of­fi­cer and gave the XML pitch to ev­ery con­fer­ence au­di­ence and jour­nal­ist and in­ter­est­ed geek I could find. It was re­mark­able; pret­ty well ev­ery­one I pitched to said “Oh yeah, I could use some­thing like that for...” and then men­tioned an app that was al­most nev­er the kind of structured-document thing we thought we’d de­signed XML for. Flash for­ward eight years, and I find my­self pitch­ing Atom to all sorts of dif­fer­ent peo­ple. What’s dif­fer­ent is, in­stead of me push­ing at them, I’m get­ting asked “Tell me about this Atom stuff.” What’s the same is that pret­ty well ev­ery­one I pitch to says “I could use some­thing like that for...” and what comes af­ter the dots ei­ther isn’t blog­gy at al­l, or is some weird syn­di­ca­tion an­gle I hadn’t thought of. I’m not say­ing his­to­ry al­ways re­peats it­self; just telling you what I’m hear­ing.
 
The Last Atom Issue · The Atom Work­ing Group has sailed through our IETF last call on the da­ta for­mat with a fair­ly small num­ber of com­ments, none of them rais­ing par­tic­u­lar­ly trou­bling is­sues. This leaves us with a tee­ny lit­tle set of is­sues to set­tle, of which on­ly one is re­al­ly ma­te­ri­al: can the same en­try show up more than once in a feed? It’s trick­i­er than you’d think, but I can see the in­di­vid­u­al fibers in the fin­ish line ...
 
IETF Last Call on the Atom Format · This is a mile­stone: the IETF Atom­pub Work­ing Group thinks we’re most­ly fin­ished with the Atom Syn­di­ca­tion For­mat, and our Area Direc­tor has put the draft out for last call IETF-wide. It’s not set in stone, the WG is chew­ing on a cou­ple of last lit­tle change pro­pos­als that might get con­sen­sus, and it’s quite like­ly that some of the smart IETFers will spot prob­lems and ask for changes. We think Atom does a rea­son­ably good job of co­a­lesc­ing all the years of ex­pe­ri­ence with all the RSSes, and I’m pret­ty con­vinced it will make a dif­fer­ence. We’ve tried to err on the side of omis­sion; there are lots of things that could have been added that weren’t. We on­ly felt con­fi­dent enough cov­er the ar­eas where there’s a lot of pri­or art and the ar­gu­ments are most­ly over. Any­how, Atom’s ex­ten­si­ble; I’m pret­ty sure the mar­ket is cre­ative enough to con­verge on good ex­ten­sions to ad­dress any im­por­tant things we left out. So please have a look ei­ther at the IETF ASCII or nice mod­ern HTML ver­sions of the draft, and see if you think we’ve missed any­thing.
 
Why We Need Atom Now · We’re get­ting re­al, re­al close to send­ing the Atom data-format draft off for gen­er­al IETF re­view; the rest of the pro­cess can’t hap­pen too fast for me, be­cause there are two big prob­lems that bite me ev­ery day that Atom will give the en­gi­neers the tools to fix ...
 
Hats Off to the Atomics · What hap­pened was, we’re get­ting close to done on Atom, so we tried to close down our last fif­teen is­sues. The good news is that the de­bate has been deep and in­tense but (al­most al­l) po­lite and grown-up. To­day, a Sun­day for­sooth, we had over 120 mes­sages to the mail­ing list. The bad news is that the last call jarred loose a few new is­sues, as last calls will, but we’re still very close. The work­ing group is do­ing great, I love ’em.
 
One-Click Subscription · Re­cent­ly, Dave Win­er point­ed out that there’s a prob­lem in how peo­ple go about sub­scrib­ing to to RSS feed­s, and made a pro­pos­al to ad­dress it. Dave’s idea—essentially One Great Big sub­scrip­tion dispatcher—should work, near as I can tel­l, so any prob­lems would be around busi­ness and pol­i­tic­s, not tech­nol­o­gy. John Robb does some more think­ing, as does Phil Wind­ley. If we can’t find the busi­ness cre­ativ­i­ty, Atom has an­oth­er so­lu­tion that RSS could maybe bor­row ...
 
Atom Format: Final Flurry · The Atom Da­ta for­mat (whose most re­cent draft spec is avail­able in ASCII and HTML) is about ready to move on to its next phase, con­sid­er­a­tion by the whole IETF. We’ve ganged up the last 15 open is­sues for the Work­ing Group’s fi­nal yes-or-no. The dis­cus­sion is here; if you care, speak up now.
 
The Atom End-Game · I re­cent­ly pro­posed to the IETF Atom Work­ing Group that we might be near­ly fin­ished. Some peo­ple think that’s a mis­take be­cause, as they point out, Atom doesn’t have much more in the way of fea­tures than RSS. Here’s why I dis­agree ...
 
Wanna Be Atomic? · Drop by the Atom Hackathon at the XML 2004 con­fer­ence. Sam and I will be there for sure, and maybe we can make some real-time pro­gress.
 
How About a Date? · Over in Atom-land, we have all been ed­u­cat­ing each oth­er at in­cred­i­ble length (365 email mes­sages so far) on the sub­ject of the dates you at­tach to the elec­tron­ic what-nots you pub­lish. It has, like most ed­u­ca­tion­al pro­cess­es, been painful. I frankly can’t imag­ine who out there would care about this, but I need to get it out of my sys­tem. Plus it has a hu­mor­ous rhyming cou­plet ...
 
Tribal Drumbeat · Some­times it feels like a fam­i­ly. What hap­pened was, we’re hav­ing all this angst over in Atom-land about well-formedness and me­dia types and the fact that RSS is usu­al­ly served “wrong” and the prob­lems this caus­es. Wel­l, be­fore too much longer, there are go­ing to be a lot of Web re­sources named this.atom, that.atom, and the-other.atom be­ing dished out by Web Servers ev­ery­where, and by de­fault those servers are gonna look at the names and say “Dot-atom what? Yer tex­t/­plain, punk.” So I ap­pealed to Greg Stein of Apache and Google, and he had a pow-wow and re­port­ed back I've gone ahead and done this: the ap­pli­ca­tion/atom+xml (for .atom) type will ap­pear in our next re­leas­es (A­pache 1.3.32 and Apache 2.0.51), when­ev­er those come out. Wel­l, Apache’s not the on­ly serv­er out there, so I wrote off to Obasan­jo and Scoble and said “Here’s the prob­lem, how about IIS?”. So Scoble did some dig­ging and got rout­ed to Thomas Dem­l, lead pro­gram man­ag­er on IIS, and I saw a for­ward­ed email say­ing The change goes in­to Win2K3, SP1. Now if we could sort out the rest of the Internet’s is­sues that smooth­ly... any­how, thanks guys.
 
Atomic Heartbeat · The IETF AtomPub Work­ing Group for­mal­ly buck­led down to work on June 23rd, and it’s been more or less won­der­ful since then. Here­with a few words of ap­pre­ci­a­tion ...
 
Atom News · Yes, there will be an Atom IETF Work­ing Group. No, there is no “standards war” ...
 
Atom Community Meeting Report · Here­with notes and re­port­ing from the Atom Com­mu­ni­ty Meet­ing, June 2nd at Sun’s San­ta Clara cam­pus. I’ll up­date this frag­ment all day and see how that goes. IRC is ir­c://ir­c.freen­ode.net#atom, lots of peo­ple there. I’m writ­ing this in chrono­log­i­cal or­der, not blog or­der, so the new stuff is at the bot­tom. Here are Wal­ter Underwood’s notes. [Up­date: 4:40PM] ...
 
Keith and Angle Brackets · This, by Dave Walk­er, shows up fif­teen mes­sages or so in­to the com­ments on a Dive Win­er piece, and I can’t link to it di­rect­ly (noth­ing pur­ple) [Up­date: yes there is, and I can], but it’s way too good to pass up: When the world end­s, the on­ly things left will be cock­roach­es, rat­s, Kei­th Richard­s, and man­gled text that has been es­caped one-too-many or one-too-few times.
 
Identifiers · Mark Pil­grim has a good piece to­day on how to choose per­ma­nent iden­ti­fier­s. I’m go­ing to pay him the com­pli­ment of dis­agree­ing at length ...
 
Vive les Masturbateurs · I can’t say I’m on board with the mes­sage, but you got­ta love the ti­tle. Some­one trans­late this one for Dave Win­er, he’ll like it (se­ri­ous­ly, not kid­ding). [Up­date: Di­di­er Bar­bas did.] And while on the sub­ject of good names, the tortoise’s blogroll has a point­er to some­thing called Une ron­delle de saucis­son et l'addition; I can re­late to that.
 
Atom Community Meeting · We’re go­ing to have an Atom Com­mu­ni­ty Meet­ing June 2nd at Sun’s Silicon-Valley of­fices. It’s wide­ly known that Sam Ru­by (with some help from me) has been try­ing to find Atom a home over in the IETF. I’m per­son­al­ly op­ti­mistic that that’ll work, but even if it does, the next time we could get the Atom com­mu­ni­ty in a room would be at the Au­gust IETF, and it’s been a while; I’ve nev­er been to an Atom meet­ing, my­self. Check out the meet­ing and come if you want to con­tribute. Read on for a quick note on my po­si­tion, and Sun’s, vis-a-vis Atom ...
 
Atom@IETF.org · Sam Ru­by is charg­ing ahead with the idea of tak­ing Atom to the IETF, and I gen­er­al­ly ap­prove. Sun thinks it would be a good idea for me to put some cy­cles in­to syn­di­ca­tion tech­nol­o­gy, so I told Sam that and he came back in about fif­teen sec­onds say­ing “Wanna co-chair?” ...
 
author · Dad · software · colophon · rights
Random image, linked to its containing fragment

By .

I am an employee
of Amazon.com, 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.