Sorry. I was confused about what was being proposed... Disregard earlier
comments. I'll have more relevant comments shortly in a new message.

bob wyman

On Sun, Nov 20, 2011 at 4:49 PM, Julien Genestoux <
[email protected]> wrote:

> Bob, I think Alexis comment should have cleared up any doubt... but just
> in case, here is how I view things.
>
> The "consumers" (I believe you are describing the subscriber in the
> pubsubhubbub dialect), are actually never aware of how the hub got the
> content it pushes to them.
>
> As you pointed, it would be no different from how it is done right now.
> However it is not what the spec says. The spec currently specifies how the
> publisher should ping the hub... and well, at least for Superfeedr's hub,
> it is rarely how things actually happen. I believe it is not how things
> happen for other hubs hosted directly by the publishers (wordpress...).
>
> As for your last paragraph, I think there is a misconception. Right now, a
> hub never advertises the feeds that are available thru it... it's the
> publisher (thru the feeds) that advertises the hub. Which means that in the
> context of delegation, the feed can just simply point to several hubs (and
> many feeds do already).
> Finally, in the context of chaining, once again the subscriber don't
> really need to know how the hub "learnt" about the update. It is actually a
> big chunk of Superfeedr's business : some susbcriber subscribe to
> Superfeedr's hub for any feed, including the ones that have a hub in them.
> When that happens, superfeedr subscribes to the actual hub (as pointed to
> by the publisher), and then pushes the updates forward to the subscriber.
>
> I hope this helps,
> julien
>
>
>
>
>
> On Sun, Nov 20, 2011 at 10:34 PM, Alexis Richardson 
> <[email protected]>wrote:
>
>> Publishers could be seen as posting content to a local topic which
>> consumers express interest in by talking to the publisher.  Then, a
>> hub is just a consumer that republishes using the same topic name.
>>
>>
>> On Sun, Nov 20, 2011 at 9:28 PM, Bob Wyman <[email protected]> wrote:
>> > If the publisher->hub part of the spec is removed, how will consumers of
>> > feeds express their interest in receiving updates? How would that
>> process be
>> > different from what is currently specified for subscribing to hubs?
>> Would
>> > there protocol for subscribing to a single feed publisher be distinctly
>> > different from subscribing to a publisher that handles multiple feeds?
>> > If, as Monica suggests, it is clarified that a publisher can serve as
>> their
>> > own hub, isn't that sufficient to accomplish the goals you've specified?
>> > It is likely that some consumers will prefer to receive updates
>> indirectly
>> > via a small set of hubs even if publishers are able to send those
>> updates
>> > directly. In such cases, it would be desirable for a publisher to
>> specify
>> > not only that it serves as its own hub but also delegate to one or more
>> > multi-publisher hubs. (Note: A hub that serves many publishers should be
>> > able to achieve a variety of efficiencies that would not be achievable
>> if
>> > all feeds were served by their publishers. e.g. various aggregation
>> options
>> > as well as reuse of existing connections.)
>> > bob wyman
>> >
>> > On Sun, Nov 20, 2011 at 10:59 AM, Julien Genestoux
>> > <[email protected]> wrote:
>> >>
>> >> All,
>> >> I believe that the Publisher -> Hub part of the protocol should be left
>> >> out of the "core" spec.
>> >> The motivation behind it is that many hubs out there (Wordpress,
>> >> Superfeedr's... etc) have different ways of handling this without
>> affacting
>> >> much of how the general "idea" of the protocol works.
>> >> I believe the Section 7.1 should be set as optional, as well as section
>> >> 7.2.
>> >> Something like this would work well, according to me :
>> >> Once a publisher designates a hub he wants to use, he then SHOULD use
>> any
>> >> of the supported method by this hub to notify new content. 7.1 and 7.2
>> >> should certainly be put as an annex/example of how this can be done.
>> >> What do you think?
>> >> Julien
>> >
>>
>
>

Reply via email to