Thursday, June 13, 2002

Aggregation as a tools for the exterprise 5:06:08 PM • • comment [] • Google It! Paolo seems to agree. Good. John Robb seems to be thinking along the same lines as Paolo. Being able to easily set up tools to constantly monitor external information sources that interest you, pull the relevant data and categories it is something that interests many people. I have on my desk a piece of software written by Microsoft Research called Sideshow, which does some of these things. It could be ideal to actually view the data produced by an aggregator.

RDF as a LINK tag on steroids 4:43:06 PM • • comment [] • Google It! I very carefully read the RDF Primer yet I am still at a loss to explain what I could use it for. [BitWorking] Here's a simple explanation of what is RDF. I believe it's correct (if not, please let me know!) To explain what RDF is about, let's use a similar concept -- LINK tags in HTML. Lately Blogland had a LINK craze, and everybody started putting LINK tags on their weblogs. These LINK tags look something like this: <link rel="alternate" type="application/rss+xml" title=RSS href="http://radio.weblogs.com/0106548/rss.xml"> This link essentially says the following: "The resource http://radio.weblogs.com/0106548/rss.xml is associated with this page by way of it being its application/rss+xml." Or, in other words, what it says is: "if you want to find out (one of) the RSS channel of this page look at that url." This is exactly like RDF, with the following modifications: The LINK tag is placed in the HEAD of the HTML file to which it refers. Thus, the subject of the association is implicitly the page in which the association is mentioned. In RDF the association does not have to be embedded in the subject -- is can be external. The type of association is encoded as a mix of attributes (REL, TYPE, and TITLE). In RDF, the association type is more explicit. Also, RDF is built so that it is possible to create globally unique associations types (which is a constant problem with MIME types). Well, there are more differences in the "little print", but I think these are the really big issues.

Aggregators should bring more, not less, information 12:29:59 PM • • comment [] • Google It! Excerpts from [Paolo Valdemarin: Paolo's Weblog]: The point is that with tools like Radio, users can move a significant part of the content management workflow to their desktop, they can choose how to visualize it and define the flows of information. Most of all, they can decide how to visualize those contents. Indeed. However, Paolo later says: Besides, unlike email, where the message is only stored in your local mailbox, aggregators are just a way to be notified about updates, the actual data still sits on the New York Times site, your accounting software, intranet server, other weblogs, etc. With this I take issue. I think that the real power of aggregation chains (an information feed aggregator possibly preceeded by an RssDistiller-like stage that takes care of subscriptions, periodic pinging sources, scraping, etc) is that it bring all the relevant information to your client, not that it brings links. As an example, I'm subscribed to more than 60 channels. There seem to be two aspects to blogging people differ at: (1) whether to post in the RSS channel only a headline or the entire information capsule, and (2) when the post is about some other Web post, whether to quote the relevant post or just link to it. Invariably, the posts that are easiest to track are those where maximum information is provided "built in". Posts like Ernie's are best, because I don't have to break my attention by the act of opening another Web window to read the stuff he comments on. Rather, I get it all in one place. Feature request for Web aggregators: Automatically harvest commented-on articles and bring them along the original post. This could save considerable time, that could be invested in reading yet-more channels...

© Copyright 2002 Ziv Caspi.