October 7, 2006
@ 02:33 AM

Tim Bray has a blog post entitled On Comments where he writes

I’ve had comments running for a few days here now (I prefer to say “contributions”, but whatever). People are irritated at me because an ongoing fragment shows up as unread in their feed-reader whenever a new comment comes in. I’m not sure what the right thing to do is. This piece outlines a few options and asks the community for discussion.

This is one of the reasons I've given about disliking the atom:updated element in blog posts like Indicating Updated Items in RSS Bandit. It should be up to the user to decide what count as 'significant' updates that warrant marking the item as changed or new in the user interface, not the publisher. Tim thinks that new comments in a blog post should lead to the reader being notified by their aggregator, I think this should only be the case when the user has explicitly opted in for notifications on changed new comments. This doesn't extend to updates because the definition of what counts as a 'significant' update is going to vary from publisher to publisher and from user to user.

My advice to Tim is to  use the Atom threading extensions which provides explicit mechanisms for indicating changes to the number of comments and provides a way to link to comment feeds as opposed to hacks like changing the value of atom:updated or putting the comments into the atom:content of the entry. Those both sound like recipes for a negative user experience when reading his blog in many aggregators.

The title of this blog post is probably harsher than I intend. I think it is useful to have a last modified date in the form of atom:updated on items in a feed. What I disagree with is impacting the user experience based on changes to that element.


Saturday, October 7, 2006 4:14:46 AM (GMT Daylight Time, UTC+01:00)
It is impossible to tell programmatically whether a given change is a typo or more significant. atom:updated gives the the consuming application more information than it would have otherwise. Having more information is generally a good thing. Particularly with feeds like Tim's where he uses this judiciously.

But in the end, it is just information. One can argue, as you are doing, that if your application choses to always respect this over the wishes of the end user, then this is a bad choice for the application to make.

This issue isn't specific to updated. I know of feeds where guids can't be trusted.

That's why I provide an option in Venus to specify what information one is to "ignore_in_feed". This information can be specified globally (affecting all feeds) or on an individual feed basis.

Saturday, October 7, 2006 6:13:36 AM (GMT Daylight Time, UTC+01:00)
Dare, sorry this has nothing to do with the topic but why is your blog so slow? Or often fails to serve the page?

I find myself often avoiding coming here because of that.

- Sylvain
Monday, October 16, 2006 3:12:53 AM (GMT Daylight Time, UTC+01:00)
Comments are closed.