I am really curious why OPML hasn't caught as many peoples' imaginations as it warrants and these goofy little twitter lists are now all the rage. My theories so far are that twitter makes it all the easier to publish updates, it's tied to individuals, and it's easy for people to understand lists of individuals that publish easily. I tell people all the time "just open that OPML file in a text editor and look at it - it's super simple!" but I think Twitter and its lists made it over the UX hump in a unique way. Still trying to figure it out though.
dave
· 1 month ago
My belief is that it's content that drives the apps. You need something or someone to go first. With RSS it was Wired, Red Herring, Motley Fool and Salon then the early blogs then the NY Times and it blasted off.
With podcasting it was IT Conversations, the Gillmor Gang, Morning Coffee Notes, Daily Sourcecode, the community, then NPR and it blasted off.
This confluence has not (yet) happened for directory structures. It's not immediately obvious who the big drivers are going to be, but if they're out there, the Twitter lists feature is getting them to think about this stuff. I don't doubt that OPML will be part of the bootstrap and that people will quickly want to make lists that include resources that are not Twitter users or lists of Twitter users.
In other words, this is the most promising moment for OPML directories that's come so far.
anildash
· 1 month ago
I agree -- this is a watershed moment for public editing of lists/outlines, whether people know that OPML enables the interoperability or not. If Twitter just makes a great UI (a web-based one, instead of having to download/install the OPML editor), then that's a valuable opportunity for bootstrapping this stuff, especially since it's going to be very easy for people to make apps to export Twitter lists as OPML files.
As I was writing about this today, it really does feel like a moment where we can talk about there being two ways to discover the web, by humans or by machines. And thus far we've spent this decade only using machine-based methods.
Jay Cuthrell
· 1 month ago
As one of the writers I follow put it, sometimes you want to follow the list of those that don't follow a list of certain lists. While tongue in cheek, it's actually an invaluable and (perhaps?) established metric for various dating/match-maker services as possible ephemera for what you like as much as what you don't like. When you consider the rapid stratification possible on any given polarizing issues list in these approaching times of proximity like interaction online, the ability to know what you want to avoid, the ability to squelch, and all of the finer grained aspects -- may become as or more desirable that merely finding what you Like [1].
[1] overt FriendFeed/Facebook reference
scott
· 1 month ago
Coming from the perspective of a developer, the problem with OPML is that it is focused on outlines and not on distributed lists. There are already two standards for distributed lists; RSS and Atom. The moment you introduce hierarchy in a single document format you lose the distributed nature that most of the use cases rely upon and you open the door for content getting out of sync which introduces a new problem that is very difficult to solve and decreases the usability of such solutions. An RSS enclosure or an Atom link can reference any content type, not just a media file. If the entries of a feed contained links that simply referenced other feeds then developers could leverage existing technologies and paradigms as well as provide a better fit for the common use cases of distributed lists.
dave
· 1 month ago
Look deeper, OPML inclusion takes care of this.
scott
· 1 month ago
OPML inclusion is a good example of the problem I am trying to illustrate. In the context of lists of lists I would need to convert all of my feeds to OPML. The only value I see in doing this would be to enable my content to be editable in your OPML Editor thus making my content incompatible with all the tools and infrastructure that currently exists for RSS and Atom. I can't see why any developer would want to do this or how this proposition would benefit the domain of open content distribution.
dave
· 1 month ago
No, actually you wouldn't.
Maybe you should go study the OPML 2.0 spec, before you explain OPML to me.
Geez Louise. A little knowledge could make you dangerous. :-)
scott
· 1 month ago
I have a feed that has entries which each link to other third party feeds. Please explain how OPML can add value to this list of lists without me having to convert my feed to OPML. My list needs to be dynamic therefore I cannot model it as an OPML subscription list because aggregators do not support updates to OPML subscription lists. They do support updates to feeds.
lmorchard
· 1 month ago
Once upon a time, I wanted to see more OPML on Delicious. This is as far as I got:
That's probably as far as it will get. It did interesting things in my OPML Editor at one point.
Doug
· 1 month ago
Hi Dave
Lets talk about how we could match Twitter and OLMP... Are there some standard (defacto or otherwise) type field values for this? I have come across "rss" and "link" in the documentation and examples. I ask in particular reference to twitter users. Perhaps "twitter" would be a good type? Its more telling than a generic "rss" or "link" tag as it implies the entire twitter service with friends,etc. You could then have a field "user" (equivalent to "xmlURL" for a feed). My goal is to integrate OPML to our feed application but it does more than just feeds and weblinks.. if there is a precedence for twitter I'd love to know.
Another approach would be type="service" service="twitter" account="davewiner"
All Comments
With podcasting it was IT Conversations, the Gillmor Gang, Morning Coffee Notes, Daily Sourcecode, the community, then NPR and it blasted off.
This confluence has not (yet) happened for directory structures. It's not immediately obvious who the big drivers are going to be, but if they're out there, the Twitter lists feature is getting them to think about this stuff. I don't doubt that OPML will be part of the bootstrap and that people will quickly want to make lists that include resources that are not Twitter users or lists of Twitter users.
In other words, this is the most promising moment for OPML directories that's come so far.
As I was writing about this today, it really does feel like a moment where we can talk about there being two ways to discover the web, by humans or by machines. And thus far we've spent this decade only using machine-based methods.
[1] overt FriendFeed/Facebook reference
Maybe you should go study the OPML 2.0 spec, before you explain OPML to me.
Geez Louise. A little knowledge could make you dangerous. :-)
http://feeds.delicious.com/v2/opml/tags/deusx
http://feeds.delicious.com/v2/opml/deusx/writing
http://feeds.delicious.com/v2/opml/deusx/filety...
That's probably as far as it will get. It did interesting things in my OPML Editor at one point.
Lets talk about how we could match Twitter and OLMP... Are there some standard (defacto or otherwise) type field values for this? I have come across "rss" and "link" in the documentation and examples. I ask in particular reference to twitter users. Perhaps "twitter" would be a good type? Its more telling than a generic "rss" or "link" tag as it implies the entire twitter service with friends,etc. You could then have a field "user" (equivalent to "xmlURL" for a feed). My goal is to integrate OPML to our feed application but it does more than just feeds and weblinks.. if there is a precedence for twitter I'd love to know.
Another approach would be type="service" service="twitter" account="davewiner"
thoughts?
It'll take a few days at least...
But these are good ideas.
Dave