Curiouser and Curiouser!

ENT enabled RSS. ENT enabled RSS: No idea if its correctly implemented. Maybe a ENT validator is in order. Update: Some errors corrected.... [Daily Bytes]

Congratulations to André for having the first ENT enabled MovableType feed!

This is really, really, fantastic news.

Easy News Topics. Easy News Topics: Finally someone is working on blog categorization, just as I suggested some time ago. Paolo and... [Daily Bytes]

Congratulations to André for having the first ENT enabled MovableType feed!

This is really, really, fantastic news.

Not so easy?.

As requested, David Sifry reviews Easy News Topics 1.0.

The upside:

As a format, ENT is easy to understand, easy for application developers to implement, and pretty easy to parse. Kudos to Matt and Paolo for coming up with a design that is simple but extensible.

The downside isn't so easily excerpted. Read the link for more.

[The Doc Searls Weblog]

Without wishing to seem oversensitive I think it is fair to say that David's issues are not with ENT per se, but with the idea of people cataloguing their own items with meta data.  His problems are social and the same argument could be made about RSS1.0 and mod_taxo.

David reflects upon what happened to meta data in web pages and wonders if the same thing won't come about here too.  Here is how I responded in a comment on David's blog:

Thanks for the comment. I think you're broadly right that there will be a capacity for abuse here but it scales differently.

When Google was indexing billions of pages the metadata became meaningless even without the abuse. However with RSS I think the game is different.

I'm subscribed to 108 feeds right now. I add perhaps 1 a week on average. But I trust that the authors of those feeds would use topics reliably or not at all.

Any application I have that can understand my subscription list can usefully utilize those topics whatever I think about anyone else.

In a social software context somone else might trust my confidence in the indexing of those feeds. Etc...

I think we can find ways of making this work. I also think that automatic annotation of feeds with topics is another possibility.

To pick an example from my head. Syndic8 could have a topic engine that creates an annotated feed for every feed they monitor. If you trust Syndic8 you could use that feed instead and get the benefit of their indexing. And so on.

All Categories are Local. Dave Sifry comments about the draft specification of Easy News Topics 1.0 (ENT), proposed by Paolo Valdermarin and Matt Mower.... [BookBlog]

Adina's feed doesn't carry the whole post so:

Dave is concerned that a category standard would fall prey to the problems of ambiguity and scamming that killed HTML META tags.

As I noted in comments on his post, Dave is absolutely right at the scale of the web or the blogosphere.

However, I think that categories will be much more valuable at the community level. For example, Austin has a meta-blog, aggregating posts related to Austin. People in other cities are starting to do the same. If we could map sub-categories, we would be able to create a cross-regional directory. There are local editors who keep the system from being spammed, and make decisions about how to map categories.

So, I think that the system can work in the context of defined groups and defined applications.

The only thing that ENT is missing is a way to alias categories -- Austin's "music events" maps to Ann Arbor's "concerts." Presumably this could be implemented at the application level.

Aliasing topics is a powerful, and sometimes complicated, concept so we choose to leave it out of the ENT spec.  We took the view that this issue was best resolved by using a proper topic map behind the ENT cloud.  However if the functionality is needed and we can find a simple way to do it, we're open to that - especially while the spec is still in at the draft stage.

The XFML standard, for example, provides a tag which can be used to explicitly link a topic in one document to the same topic in another document.  ENT could just duplicate this facility.  But what would it be connecting to?

In XFML the answer is easy.  An XFML topics in one document connects to an XFML topic in another document.  Similarly for XTM the implication is that you make associations with other XTM topics.  But ENT topics are transient things that live for a short while in an RSS feed.  So logically you wouldn't connect an instance of an ENT topic in one feed to an instance of an ENT topic in another feed.  Would you?  Apart from the fact that I know of no way to address an arbitrary element in an RSS feed I think it far more likely that you want to alias the use of the topic to a fixed point of reference, e.g.

When you see "music event" in the Austin feed you should read it as the same thing that Ann Arbor folks mean by "concert."

There may be no current item in the Ann Arbor feed using the concert topic so we definitely need a fixed point of reference.  What is that?  One answer might be a topic map and topic maps are certainly good for this.  How else could we do it?

It would be easy enough to add an XFML style  element to the ENT spec.  An example might be:

  http://annarbor.example.org/topics.xml#concert  

However ENT also allows topics without the backing of a topic map, in which case someone may just want to specify a topic ID, e.g.

  concert  

But the problem with this approach is that it means that music_event is now synonymous with anyone's definition of concert.  We could add an href attribute to connect to differentiate but I don't particularly like that either.

Maybe someone else can come up with a simple alternative that befits the style of ENT?  I'm still leaning towards this as a problem best solved by topic maps.

[Update] I've been thinking about this a little more and feel that the best way forwards is to work with groups like SocialText and the SSA to make sure the right tools are in place to solve this kind of problem using topic maps.  Adina?