RSS in the dog days of summer

It's the dog days of summer , a time when energy, enthusiasm, and news are all supposedly at a low ebb. But my RSS newsreader this morning delivered a flood of items that got me thinking in a dozen directions.

Expect More

Sam Ruby's Expect More is a powerful meditation on extensibility in an era of abundant web services. Reacting in part to Clemens Vasters , who has recently argued for more interface strictness, Sam appeals to ~~Tim Berners-Lee's~~ Jon Postel's dictum, "Be liberal in what you accept, and conservative in what you send." ¹

⁽¹ Postscript from Sam: "I actually was aware that Jon Postel made the comment, but I wanted to refer to T B-L's principles of design, which makes that quote without attribution." )

Clemens acks this Greg Reinacker response:

I think you should only extend the request in places where the schema specifically allows it; in the case above, if there was a xml:any extensibility element, you'd be fine, and you could put your extensions there. If you want to omit parameters, you can only omit the ones allowed by the schema.

Sam's article, however, also approvingly quoted this from the Hailstorm spec's intro:

Anywhere you see {any}, read this as a location where you can extend the schema with your own freeform, namespace-qualified XML.

So is there really a disagreement here? I'm not sure there is. The universe of web services is vast, and it will contain a range of use cases. Some will prefer tolerance, others strictness. These are not mutually exclusive strategies. The {any} escape hatch creates a zone of experimentation, and ensures evolvability. As in nature, species once evolved cease -- by definition -- to be able to mate with other species. Limits to tolerance, in other words, can naturally evolve.

Too much collaboration?

Another take on evolution comes via Jeroen Bekkers , who cites Michael Helfrich's lament that IT still doesn't get collaboration.

Sharing things isn't the problem and yes, intellectual property needs to be secured, but most enterprise-class technology frameworks have ways to facilitate this. At least those that want to stay in business.

Prehistoric, glass house, armpit-scratching folks still exist in many IT organizations. And they are stifling your ability to innovate.

Order from randomness

I love the fact that the O'Reilly Network publishes articles like Owen Densmore's Examining Random Events , the first in a series on the subject of complexity. Writes Densmore:

Randomness leads from unordered to ordered solutions.

Why should you care? Because computing is starting to use the techniques of complex adaptive systems to build extremely robust systems, composed of many independent but communicating components.

The same theme of order arising from randomness occurs in Werner Lowenstein's The Touchstone of Life , which I found by way of Sam Ruby's Neurotransmitters essay, and which led me to write another of those speculative and interdisciplinary pieces that the O'Reilly Network likes to publish.

Client-side vs. server-side scripting

When Tony Bowden read my recent post on client-side versus server-side scripting, he thought, aha!

I turned SSI on in Apache for my blog directory, and wrote an Apache Logging Handler that throws the referrer information into a MySQL database. Then I wrote a simple CGI script that tallies up the referrers for a given page, and outputs them as HTML.

Then in my Day Template, I added the line:



Now when Radio generates my page, it will include the SSI command in the static page, which in turn will get called at page request time, calculating the HTML to insert in the final document. [ Tony Bowden ]

Nice! I hope to rehost my Radio blog on an Apache server where I can do the same kinds of things. But of course, while SSI Perl scripting is something that people like Tony and I do in our sleep, it's not for average folks. I foresee two future directions here. Centralized blog services will increasingly offer the kinds of new features that the blog scripters are inventing. And so will decentralized services like Radio. The browser is still mainly a client/server creature, but going forward, technologies like .NET and Flash blur the boundary, offering the same kinds of programmability on both ends of the pipe. Where to deploy code will become less an architectural constraint, and more an architectural choice.

Privacy Digest RSS feed now readable

I almost unsubscribed from the Privacy Digest weblog because its full-bore feed was overwhelming my newsreader. But I noticed today that it has trimmed itself down to nice tidy titles and blurbs. Thanks!

To circle back to Sam's extensibility theme, by the way, I think we're overdue for an optional RSS unique ID element that's been talked about on and off for a while. Another source of RSS clutter for me is repetition of slightly-modified versions of items. I'd like my reader to be able to deal intelligently with this repetition.

Wi-Fi honeyspots

Among the items that Privacy Digest picked up on today, in addition to Steve Gillmor's priceless .Net Insecurity Day , was this SecurityFocus story on wireless honeypots:

Peter Shipley, the security researcher who coined the term "war driving" over a year ago to describe the practice of cruising city streets in search of wireless networks, says he thinks wireless honeypots can produce interesting results, but that it could prove impossible to accurately differentiate between deliberate intruders and ordinary users accidentally dropping into the network. "The statistics are not going to be black and white" says Shipley. "They're going to be iffy and there's going to be a lot of speculation involved."

This one came to me twice, by the way. I get the SecurityFocus feed directly. Privacy Digest got the item through Slashdot. Triangulation at work. Further triangulation came from Jon Schull who cited this Seattle Times story:

Sarah Eder, a Denver-based spokeswoman for AT&T Broadband, which provides high-speed cable modem service in Washington and 14 other states, was blunt: "We consider this theft of service."

Still, Eder said AT&T is treading carefully. It has a vested interest in spreading broadband, so doesn't want to discourage awareness of the technology, she said.

(Note that this item came yesterday from 802.11b News , whose feed I receive. It didn't cross my threshold yesterday but did today thanks to Jon's triangulation, and to a cluster of related items in today's feed.)

DRM-based content protection runs into the same problems, of course. AT&T Broadband, for example, knows that a certain amount of "pass-around" subscribership will actually help it to thrive. How much is too much? It's a dilemma! Those darned bits keep on acting like they don't think the laws of atoms apply to them. We're led, as a result, down slippery philosophical slopes like this one:

For example, as Russ Nelson asks on the BAWUG mailing list, "So if I park within range and open up my Win/XP laptop and it DHCPs an address with no intervention on my part, am I guilty of a crime?" In other words, how does one judge intent? [ David Sifry , via 802.11b News ]

News in the dog days

Slow news day? Not hardly. I might need to take a vacation from RSS so I can get some work done!

Former URL: http://weblog.infoworld.com/udell/2002/07/30.html#a360