Comments on James Clark's Random Thoughts: A tour of the open standards used by Google Buzz

Thanks for good summary! I've tried making tra...

2010-12-26T12:22:50.468+07:00

Thanks for good summary!
I've tried making translation into Japanese. [http://kshigeru.blogspot.com/2010/12/tour-of-open-standards-used-by-google.html]
Is there any license term?

Hey, James. How's that New Year's resolut...

2010-08-27T23:38:53.817+07:00

Hey, James. How's that New Year's resolution holdin' up?

James, Thanks for the comments. If you're in...

2010-02-27T13:32:18.862+07:00

James,

Thanks for the comments. If you're interested in filing some issues against the Salmon or Magic Signatures specs, I've opened up an issue tracker at http://code.google.com/p/salmon-protocol/issues/list and tracking issues raised there.

I assume you mean RFC3447 for RSAPublicKey -- it's possible that ASN.1 DER encoding w/base64url would work, it's just very hard to verify that by reading the ASN.1 spec itself. If there's a reasonably way to specify this as a profile of ASN.1 DER (just for the set of integer fields needed for keys) without pulling in all of ASN.1 as a dependency it could be reasonable to use it.

Salmon has a specific problem in that it needs its signatures to survive not only transit but storage in hostile environments (e.g., SQL) prior to re-syndication. Thus the armoring provided by Magic Signatures.

is this all ? not more ? :-)))) look our "tou...

2010-02-22T21:29:30.768+07:00

is this all ? not more ? :-))))
look our "tour" for yiid.com

http://blog.yiid.org/2010/02/22/openyiid-ein-update/

AFAICT the list of names on the OWF site shouldn&#...

2010-02-22T15:28:04.709+07:00

AFAICT the list of names on the OWF site shouldn't be read as an endorsement, but rather an effort by folks to give guidance on what looked like an interesting endeavour.

It's been around for a while, and it's still hard to even say what the OWF is, much less what it does. Beyond the license, of course.

Dare, I take your point about using the term &quo...

2010-02-22T11:16:41.455+07:00

Dare,

I take your point about using the term "standards". I guess I should have said something like "draft specifications that could plausibly evolve into open standards".

Dare -- right. That's why I immediately left ...

2010-02-22T04:19:23.025+07:00

Dare -- right. That's why I immediately left a comment clarifying the difference between standards and specifications the moment I saw James' post. My comment is right above yours.

And to be clear, nearly all of these specs _are_ standards. The ones that aren't, like Activity Streams and Salmon, are still being worked on in the open and no standards body would accept them in the current form anyway. What should I call them instead?

If there's something you'd like to add about what Google or I should be doing differently in building Buzz, I'm all ears. In fact, we'd love it if you'd contribute.

And of course, James knows better than anyone the difference between standards and non-standards. He's written more of them than the rest of us combined. : )

Dewitt, Standards mean a very specific thing in...

2010-02-22T04:03:03.909+07:00

Dewitt,
Standards mean a very specific thing in our industry whether they are de facto or de jure standards. Misusing the term is bad for the industry because it allows companies to sling FUD while using what are effective proprietary technologies in their products and it misleads customers.

This is orthogonal to the merits of the underlying technologies or whether Microsoft employees are involved in OWF or Atom-related spec work.

Dare, PuSubhubbub was indeed developed by Google ...

2010-02-22T03:45:41.370+07:00

Dare,

PuSubhubbub was indeed developed by Google employees, but as a side project and the spec has been released for anyone to implement. No licensing issues.

Webfinger is an effort by a number of folks, most influencially Eran Hammer-Lahav (currently of Yahoo). Yes, Googlers have participated in the definition, as have folks from Yahoo, Six Apart, and any number of indie voices.

Dare -- actually, I'm very very careful not to...

2010-02-22T03:45:01.972+07:00

Dare -- actually, I'm very very careful not to ever use the word "standard" to refer to anything that hasn't gone through the IEFT, OASIS, W3C, ISO, etc. And I ask the people I work with to do the same.

You'll notice I make that very point in my comment above.

You've made it crystal clear that you don't like that Google is pushing for these specs. But what's odd is that Google is building on the very same specs that Microsoft authored or is advocating for -- Atom, AtomPub, Activity Streams, OAuth, etc. Moreover, the lead counsel writing the OWF license agreement is a Microsoft attorney, (and a very good one at that), who is doing it in part so that Microsoft has a publicly-vetted open licensing model to use.

Try not to throw the good parts out—the parts that your own colleagues are writing—just because Google is also championing for them. To be sure, we're not turning them away just because Microsoft had a hand in creating them.

I'm confused as to how a bunch of these things...

2010-02-22T02:57:52.103+07:00

I'm confused as to how a bunch of these things are being called "standards" especially by someone like James who's been around the block in this space.

Most of these so-called standards are specs written entirely by Google employees and primarily implemented by Google products (e.g. WebFinger, Salmon, PubSubHubub). In that case, why don't we call Twitter's API or the Facebook platform standards? After all, their specs are online and there a lot more apps that implement them than any of the above.

I find it extremely troublesome that one company has started a trend of mislabelling proprietary technologies as standards and has gotten members of the press and tech elite repeating their spin.

2010-02-22T00:47:54.609+07:00

This comment has been removed by the author.

Thanks for writing this up, James! We're wor...

2010-02-22T00:31:22.006+07:00

Thanks for writing this up, James!

We're working on the APIs out in the open over at http://groups.google.com/group/google-buzz-api if you or your readers want to join us there.

Regarding the OWF, the OWF isn't a standards body of course. But the OWF offers the OWFa, a permissive license that can be used for open specifications. Many of the specs we're using are licensed under the OWFa. This can happen even before a spec goes on to be standardized in a formal setting.

Orthogonally, many of these protocols are standards by way of the IETF and OASIS, some with their own permissive licenses.