Am wondering:

Is there a format for lists of iCalendar file subscriptions - seems like an ideal thing to use OPML for?

What’s the best way to include links and images in syndicated iCal.

Blogrolling has opened up its own ping server, instead of reading Weblogs.com to alert updates.

I think this is unfortunate, if ping servers become a Balkanized mess this will cause confusion. One solution may be to federate the Weblogs.com server much like DNS, i.e. have updates propagate through a network but have Weblogs.com become the top of the tree.

I can’t see any objection to this since weblogs.com is open and allows xml access.
The Weblogs.com ping server is potentially a crucially important piece of the web’s infrastructure.

Blogrolling allows pings

Update - Dave writes that Weblogs.com can indeed be federated.

Here is the first stage of something I have been collaborating on with Ian Davis:

BIO: A vocabulary for biographical information

The web works with hyperlinks - and hyperlinks that have some explanation work best. For news, the age-old headline provides perfect text for a link and headlines are specifically created to whet the appetite for more information.

One of the central problems that developers will need to work around for blogging multimedia files is how to create meaningful links.

For example the Audblog system that lets you dial a number, record a message and post it to your weblog leaves links that don’t say anything about the content, and for good reason, this is very difficult.

There are 3 solutions:

Automatic text summarization of a sound file (Autonomy can do this - but the results are unreliable).

Prompted voice recognition of a spoken title (and these systems ‘just love my accent’).

Keying in a title (difficult on a cellphone).

Despite the difficulties, a clean headline is the basis for all syndicatable content (e.g. RSS is basically a headline a link and some other stuff at its most simple) and Moblogging will need to solve this problem.

Aha - so MMS uses SMIL.

“This is a multi-part message in MIME format.
——<200303132002.UAA09403@mmscfe1.mms.mnc033.mcc234.gprs>
Content-type: multipart/related;
boundary=”Boundary_(ID_p5GD6C2ojYJObmQ605bMrg)”; type=”application/smil”;
start=”<941366072>”
”

azeem.azhar.co.uk:

Seems I was wrong about XFML, Peter Van Dijk: “XFML allows for unlimited depth in the hierarchies”

Where Next for RSS?

Tim Bray, co-creator of XML sees the future of RSS consumption as moving from standalone clients to being built directly into browsers.

“this stuff just belongs in the browser”

I see the issue as being client v. server side. Although RSS readers are currently very useful, they are akin to client side Usenet readers that were built into browsers and email clients. As the volume of RSS grows, I would rather use a web based interface to RSS in the same way that I now use Google to read newsgroups.

Jon Udell:

“Service advertisement techniques such as UDDI are not likely to pass the View Source test “.

XML and Web Services in particular are designed for machine to machine communication, but, as Jon rightly points out, their success ironically depends on human readability.

Udell: Services and links

Faceted metadata is very interesting. The notion of dynamic taxonomies and adaptive search criteria is very important for news databases where topics and attributes change much more rapidly than general search engines. For Moreover this is a crucial area.

I am less sure about XFML led by Peter Van Dijk, where the serialization of topics and facets seems to only allow 2 levels of hierarchy i.e. don’t see how you can have subfacets of facets or subtopics of topics. Now it is true that you don’t stricly need hierarchy beyond facet -> topics, but you get bloat, rather like flat file databases. Someone correct me if I am wrong.

dive into mark: XFML

XFML home