Atom AutoDiscovery
James Snell: In an attempt to wrap up the details of the long dormant Atom autodiscovery draft, I have taken over the role of editor and have resubmitted the draft to the IETF as an individual submission.
Cool. James has also expressed a willingness to work together with the RSS Advisory Board, which has begun work on their own document.
Update: A new draft of the RSS Feed Autodiscovery specification has been published. Top of the list of Major changes?
removed suggestion to use this spec with Atom
*sigh*
Update: Paul Querna: Less than two hours after making these changes, you announced that this vote ‘passed’ the board.
*sigh*
The RSS Board has been unresponsive to my offer to work with them on the Autodiscovery draft. I’m not too concerned. So long as both specs focus on standardizing what is currently deployed rather than on trying to invent something new, it’s unlikely that there will be significant inconsistencies between the two efforts. Regardless, the invitation is open and I suspect the process will be fairly quick.
Posted by James Snell at
Please don’t misread the intent of removing Atom references from the RSS autodiscovery specification. We did that because we didn’t want to step on your toes while the Atom autodiscovery spec is being drafted.
I’m going to look at your spec this week and help where I can.
Posted by Rogers Cadenhead atTowards that end, could you publish a link to the wiki where I’m supposed to comment on your draft spec?
Posted by Rogers Cadenhead at
Towards that end, could you publish a link to the wiki where I’m supposed to comment on your draft spec?
The Atom Wiki front page has a section called Notes for First Timers which points to HowToGetInvolved.
Short summary: comments to mailing list, proposals to the wiki.
Posted by Sam Ruby atDo you guys really not think there would be any objection if the RSS Board suddenly posted a spec purporting to document Atom autodiscovery when the Atom WG itself it still arguing about it?
Posted by James Holderness at
James,
I’m not sure if I really want it to be talking about atom when the Atom-WG itself is still arguing about it, but I am mostly unhappy with how the specification was revised and then ‘passed’ two hours later after significant changes were made to it (not just the dropping of mentioning atom).
I personally believe the Atom-WG will not depart significantly from what the rss-board spec is talking about, and what today is common practice. It makes sense to refine and define common practice, to improve interoperability, but coming up with a completely new method doesn’t seem to be something anyone is interested in.
-Paul
Posted by Paul Querna atDo you guys
What a way to start a comment. Exclude from the very beginning any possibility of collaboration. Sheesh.
The current state is that we have two draft documents, even though not all of the RSS Advisory board recognizes the one posted on their site as such. One only documents “Atom” autodiscovery, and focuses on compatibility with HTML, with the result being a document which tends to reflect the perspective of a consumer that wishes to handle all the cases that, say, Firefox or IE7 would naturally handle. The other only documents “RSS” autodiscovery, and severely limits variation in a number of cases, reflecting the perspective of a producer that wishes to insert autodiscovery tags in a manner that will maximize the number of consumers that will successfully recognize the autodiscovery elements as intended.
Both perspectives are valid and important.
There also are minor issues, like the way base is described in the RSS Autodiscovery document is neither complete (what about content-location headers? What about redirects?), nor does it reference the authoritative section of the HTML document which contains the relevant information.
A unification would benefit everybody. But it has to start with a genuine desire to collaborate.
I hope the RSS Advisory Board takes very seriously the input of Paul Querna. He gets it. I also happen to agree with his input on “no inventions”.
Push the reset button, and get a better attitude.
Posted by Sam Ruby at“Push the reset button, and get a better attitude.”
‘sigh’ is not free of attitude either Sam.
Posted by Bill de hOra at‘sigh’ is not free of attitude either Sam.
Acknowledged. I am frustrated. At the last possible minute a number of changes were made based on input that wasn’t on the public list (nor even on the exclusive and exclusionary board list as far as I can see), and the significantly revised proposal was hastily declared done. Meanwhile, James’s repeated offers of collaboration were ignored.
Frankly, that both stinks and sucks.
Posted by Sam Ruby atWhat a way to start a comment. Exclude from the very beginning any possibility of collaboration. Sheesh.
I’d love for there to be collaboration between the RSS Board and the Atom WG. The issue I was questioning was whether we should be including Atom in the spec without approval from the Atom WG. That seemed a tad presumptuous to me.
There also are minor issues, like the way base is described in the RSS Autodiscovery document is neither complete (what about content-location headers? What about redirects?), nor does it reference the authoritative section of the HTML document which contains the relevant information.
I had considered bringing this up. The problem is that none of the major browsers support the RFC2616 definition of base. Nothing we can say is going to change that, so I figure we might as well stick with something that at least reflects current usage. Posted by James Holderness at
I’d love for there to be collaboration between the RSS Board and the Atom WG. The issue I was questioning was whether we should be including Atom in the spec without approval from the Atom WG.
One way to handle that might have been to post a query to the Atom mailing list, soliciting input.
Another way is to post a draft which shows both true value and true leadership and then at the last possible moment strike out the portions that relate to potential collaboration and virtually immediately declare the results done.
Posted by Sam Ruby atAs already pointed out on atom-syntax: The latest HTML5 draft defines both
feed and alternate.
Posted by Henri Sivonen at
Paul Querna’s concerns about the publication of the RSS Advisory Board’s first autodiscovery spec are valid, but our charter requires a vote to be finished in seven days. We’ll work on our process.
As I’ve said, I’m eager to work with Atom on universalizing autodiscovery. I’m submitting some suggested edits to your draft this morning. If your spec is approved, I’ll do what I can to make sure that both function in the same manner. I also will suggest that references to Atom be added to our spec.
Atom guidance was dropped from our spec to avoid conflict, not demonstrate our desire to go it alone.
Posted by Rogers Cadenhead atLatest links
Sam Ruby: Atom AutoDiscovery Comments seem to indicate we may see a shared RSS and Atom spec, which would be good Atom Autodiscovery spec M. Pilgrim and J.Snell, Ed. RSS Autodiscovery spec 1.0 Rogers Cadenhead, James Holderness and Randy Charles...Excerpt from Blogging Roller at
You can put the blame on me for the removal. I asked Rogers to do that because James Snell was moving his own Atom spec forward and I wanted to avoid any confusion on which spec should be used with Atom. Hope that clears up any confusion.
Posted by Randy Charles Morin at
I’ve added two Atom autodiscovery proposals to the wiki, but I’ve run into trouble adding the third. When I submit the newly created page for preview, I get an “You are not allowed to access this!” error.
If there’s an existing account for rcade or RogersCadenhead in the wiki, please delete it so that I can create a new one and log in.
Posted by Rogers Cadenhead atIf there’s an existing account for rcade or RogersCadenhead in the wiki, please delete it
done
Posted by Sam Ruby atAtom and RSS Go Together Like Peanut Butter and Bananas
When Randy Charles Morin and I were trying to wrap up the RSS Autodiscovery specification, we removed references to Atom to avoid discord. Telling Atom publishers how to implement autodiscovery while they’re working on their own spec seemed like a...Excerpt from Workbench at
Another way of putting Henri’s comment is that there are currently three spec drafts which are trying to specify the same thing.
Personally, I’d love it if both the RSS Advisory Board and the people working on the resurrected Atom Autodiscovery spec helped the WHAT WG with their version of speccing this stuff — WHAT WG is neutral ground, syndication-format-wise.
Posted by Edward O'Connor atEdward, I’ll be happy with whatever the Atom-WG and RSS Board chooses to do so long as the decision is made openly and without any single person trying to bully the decision one direction or the other. Everyone who has responded to this particular post seems to have a honest desire to work together, which I think is cool.
Enough metadiscussion about it tho; let’s go talk about the technical specifics of the WHATWG stuff on the atom-syntax list and see if it really does cover the requirement.
Posted by James Snell atlet’s go talk about the technical specifics of the WHATWG stuff on the atom-syntax
Actually, the appropriate venue for WHAT-WG stuff is the WHAT-WG mailing list.
Posted by Robert Sayre at“Everyone who has responded to this particular post seems to have a honest desire to work together”
Well... most everyone.
Posted by James Snell atLazy web request: can anybody dig up a reference to the schedule proposed for HTML5 by Ian Hickson in response to Tim Berners-Lee and the W3C?
Posted by Sam Ruby at
I don’t see how that’s relevant, though. Lots of pieces from HTML5 have been pulled out and worked on separately by W3C groups.
Posted by Robert Sayre atAm I reading correctly that the WHAT-WG’s work towards HTML 5 has been proposed for final publication in the year 2022? We’re supposed to defer the authority to define autodiscovery links to a body that may not complete its work for 16 years?
I need to take better care of myself if I’m going to engage in specification work.
Posted by Rogers Cadenhead atAm I reading correctly that the WHAT-WG’s work towards HTML 5 has been proposed for final publication in the year 2022?
No. That’s a proposal for the W3C schedule.
Posted by Robert Sayre atIMHO, both specs need to trim some fat.
1. BASE and HREF - I don’t see a real need to mention BASE in the spec. LINK’s HREF will be seen as other uses of HREF and will be used as such regardless of what the spec say.
2. REL - I don’t see much need for flexibility nor multiplicity here for auto-discovery. Why expand a full page enumerating all the nutty variations?
I know most of you will disagree with me but I think it will be far more useful to developers to just provide a specific example of the auto-discovery LINK element and key variations and leave it at that, leaving fools to fall off unspecified edges.
Posted by Don Park atAutodécouverte des fils de syndication
Alors que tout le monde s’est enfin mis d’accord sur l’icône de représentation d’un fil de syndication et que la cohabitation des formats Atom 1.0 et RSS 2.0 semble s’installer, on pouvait s’attendre à ce que les efforts se portent sur l’adoption...Excerpt from just call me pep at
In the RSS document, I wonder if that should be read as “You MUST want to be the next Om Malik” or “You SHOULD want to be the next Om Malik”.
Posted by dave glasser at