On Saturday, February 5, 2005, at 02:07 PM, Robert Sayre wrote:
Antone Roundy wrote:
On Saturday, February 5, 2005, at 11:48 AM, Robert Sayre wrote:
Part of our charter is to define a format suitable for archiving feeds.


Right, and breaking the feed format isn't the way to do it. Since you're advocating versioning, what are your plans for versioning the state of the feed itself? What if the title changes? It would be better to archive a series of feed documents.

Okay, I don't imagine you're going to like this, but here's a possibility:
<aggregation>
<feed>
<head> ...

You're right! I think it's vile :). It would be reasonable to want a title for the root. Perhaps the location of an HTML version would be good to know as well... oops.

I left that out of the example, but sure (<aggregation> has a <head> in PaceAggregationDocument):


<aggregation>
        <head>
                <title>Archive of feed XYZ</title>
                <updated>...</updated>
                <link .... />
        </head>
        <feed>
                ...

How about this. A series of feeds is probably overkill... I'd just use subversion. So, let's take a look at what Bob said:

Bob Wyman wrote:
As long as multiple instances/versions of an entry are permitted to
exist in a single atom document while sharing the same atom:id, the current
Atom document format provides a useable "archive format."

This is clearly a non-starter for a syndication format, since ignoramus apps that don't keep a history will show more entries in a feed than ones that do. hmm...


Also, there's mnot's suggestion that there should be an archiving "profile". I don't really like all the baggage with profiles and mustUnderstand and all that. Why don't we just define an element called <archive> that acts exactly like <feed> except for the id restriction. This approach also happens to be "mustUnderstand". Could you live with that?

This would not address your question of tracking changes to the feed's metadata. I'd be perfectly happy with not allowing atom:id to be repeated in a <feed> or the hypothetical <aggregation> if we had an <archive> element which acts exactly like <aggregation> except that atom:id may be repeated.



Reply via email to