tags:

views:

41

answers:

1

Purely in terms of its conceptual model, is the purpose of Atom (and RSS) only to provide a time-sequential series of frequently-updated items, such as "most recent blog posts" or "last twenty SVN commits," or can Atom be legitimately used to represent static and/or non-time-sequential listings/indices?

As an example, "index of files under this directory", "dog breeds" or "music genres".

Even if there's a date associated with the items, like a file's last modified date, what if you don't necessarily want time to be the primary consideration when you represent that model to your users?

The context for this is passing around (generating and consuming) lists of things in a REST-ful environment, hopefully using a well-understood format, where "date something was created/updated" is a pertinent detail, but not the primary consideration. I realize there's probably no right answer, but wanted to get some perspectives.

Thanks.

+2  A: 

Sure, it's a great way of doing non-sequential objects.

Google use it for most of their APIs, here's the API for contacts: http://code.google.com/apis/contacts/

Whilst contacts may have a traditional order (alphabetical) they're not sequential.

Just to keep things in balance, here's an article from Microsoft on why they're adopting Atom for their ADO.Net library:

http://blogs.msdn.com/astoriateam/archive/2008/02/13/atompub-support-in-the-ado-net-data-services-framework.aspx

So, that's not just supporting Atom for blog-type entries, but across all database access.

In terms of recognising what these two heavyweights are doing, I believe it is a 'right answer' in this context.

Good luck with the adoption, Ian

ianmayo