Google Usenet search is knowledge management.

Last weekend I was working on a little utility to post to MovableType via email. And I got stuck figuring out how to include files that lived in different directories. The instructions in the book and the PythonWin IDE were confusing and insufficiently helpful. So I turned to the source of all earthly wisdom.
A quick search of the Google Usenet archives found questions from several people who’ve been confused by the same problem over the last decade, complete with helpful and instructive answers. I followed the instructions that Python guru Tim Peters provided a newbie in 2001: no need to mess with the Windows registry; simply include the reference to the desired path in the header of your program.
Google Usenet search is incredibly useful for this type of question. Books don’t have space in 300 pages, or even 1200 pages to cover every conceivable implementation decision and configuration nightmare. FAQs are effective precisely because a patient editor has distilled the sea of knowledge into an elixir of Questions asked Frequently. By contrast, Google Usenet search isn’t bounded by page count or the patience of human editors. Someone, somewhere has encountered the problem that has you climbing the walls, and someone, somewhere has answered it.
Many person-hours of labor and numerous PhD theses have been devoted to designing sophisticated knowledge management systems, incorporating text painstaking tagged and cleverly autosummarized; employing expert rules and meticulously built case repositories. But my guess is that a really good search engine and a deep database of human conversation can beat fancy knowledge management a lot of the time, and most of rest of the time, the Google-Usenet approach wins on price-performance.
Once again, the intelligence in the semantic web is largely human; a person who asked a question, and a person who answered it; the machine merely serves to connect today’s seeker with yesterday’s guru.
Of course, to succeed with this approach, as David Weinberger points out, you need to know how to phrase a query that will retrieve the right antique conversations. “PYTHONPATH Windows” succeeded instantly at finding the answer to my question last weekend. The skill of phrasing a search query ought to be taught in middle school, around the same time kids get old enough to figure out that a paragraph should have a main idea.

Movable Type, Mitch Kapor and the semantic web

I’ve been a little slow on reading for the last couple of weeks, working on code instead. Learned CSS last weekend, and installed Movable Type this weekend. The MT software lets you categorize entries, so people who want to catch up on my life don’t have to slog through essays on complex systems, and vice versa. The code is available, which raises all kinds of intriguing possibilities for nifty hacks, involving email and the Amazon API and comments.
The technology industry is in a depression, and the big boys and girls wonder if the days of innovation are behind us. But there’s plenty of creativity going on in the open source world.
A couple of years ago, open source hackers were working on OS kernel implementations, web servers, and development tools. Reliable, heavy-duty carpenter’s tools; software of, by, and for professional technologists, intent on improving the machine, and more power to them.
These days, there are also communities working on tools for publishing, collaboration, communication. Creative applications using the Amazon API, the Google API, RSS syndication. And I just read on SlashDot that Mitch Kapor and Andy Hertzfeld are working on an open source competitor to Outlook, using bits and pieces of Mozilla, Jabber, and Python. The current wave of open source software development is in tools and applications for people.
Some thoughts about this trend, in several different directions:
1) During the boom, Jerry Michalski, an industry visionary and highly decent human being, used to talk about how the internet would provide tools for people to communicate and collaborate. And he’d talk about the potential for Yahoo and Amazon and AOL to be new platforms. But the economy went south, companies slowed innovation, and focused understandably on paying the bills. The good thing is, there’s no reason to wait for a Yahoo or Amazon or Microsoft to provide the tools. People are coding happily away in kitchens and living rooms.
2) Despite the fact that Mitch Kapor’s project seems to attack Microsoft in an area of towering strength, his business isn’t as crazy as it sounds. IBM is building a big business implementing open source software; there are similar services opportunities downmarket of IBM. IBM would be quite happy to deploy armies of professional service people to deploy an open source messaging system.
The Kapor announcement is vaporware; it may or may not go anywhere. The niche might be filled by some other project, some other year. But open source poses a threat to Microsoft’s dominance of the email market, just as it does in operating systems.
3) Tim Berners-Lee and various other very smart people have described a vision of the “semantic web.” According to Berners-Lee’s view, the Semantic Web would be for machines what the World Wide Web is for people, a uniform way to see and use vast amounts of formerly hidden information. The classic example is a robot secretary that will scour the web and schedule your airfares, hotel rooms, and meetings, using metadata published according to standards, and discovered via automated search and publish/subscribe notification.
Open source hackers and software companies are building a semantic web today, and it’s different from Berners-Lee’s vision. In the robot version of the semantic web, the nodes of the network consist of information, nicely categorized according to standard XML taxonomies. The links consist of protocols and tools to traverse the network, and automated processes to make calculations and execute transactions; to find the shortest travel time at the lowest cost.
In the version of the semantic web exemplified by, Daypop and Google News, the nodes of the network are people. The links of the network are relationships among people; who are reading books, selecting stories to publish, selecting sites to link. Google News, which is marketed as a replacement for human editors, depends thoroughly on humans; editors and bloggers, who select the stories to cover to begin with, and readers around the world, who chose which stories to read. The semantic web doesn’t replace human intelligence, it multiplies it by connecting people.
Despite the Nasdaq, tech innovation surely isn’t done.

Tuning out the customer

It is nice to see the mainstream press discussing music industry policies as anti-customer rather than repeating the industry’s piracy message.
And it’s also great to read Jerry Michalski saying it, since he’s been thinking and talking about these issues for years now.
Perhaps is helping to change the terms of the debate. The recent Boucher and Lofgren bills describe their goals as protecting the rights of customers to traditional fair use of media.
The good thing about using the term “consumer” in this context is that an individual hears the word and thinks “that’s me”, and my rights to things that I have in my house are being taken away. It becomes an area where politicians can take a populist stand. It takes the discourse out of the realm of abstract and technical legal principles and rights. It’s great that there are lawyers fighting these issues in the courts, and more power to them. But the language of lawyers doesn’t get people to identify and take action.
What the term “consumer” leaves out is Jerry’s “co-participant” message, which the Fortune article quoted but didn’t seem to understand. Personal music sharing, fan sites, etc. are ways for individuals to participate in the creation and sharing of culture. People’s desire to contribute could be embraced into media business models, instead of repelled as invasions into the territory and property of the media industry.
It is also pretty weird to read the characterization of Jerry as a “cyberspace libertarian” — he just doesn’t fit that image of a scruffy maladjusted coder who rants in favor of guns and drugs and abolishing the government!

Handy PC Utility

FreeRAM XP Pro from YourWare Solutions. Automatically frees memory on Windows machines, enabling them to work somewhat longer before inevitably crashing.
Of course, I ought to upgrade to Windows XP or 2000. But that would mean setting aside the time to troubleshoot the upgrade; so I live with chronic Win98 memory leaks, and use FreeRam as a palliative for the symptoms.

The villain’s in the mirror

The music industry’s problems are caused by the changing needs and
tastes of customers, as reported by this informative story in Slate.
Industry mistakes include:
* Believing in teen idols. Buyers over the age of 40 account for
44 percent of the CD sales, up from 19.6 percent in 1992,
according to a recent RIAA survey. Older listeners have diverse
tastes, and it’s harder to reach a fragmented market. So the music
industry keeps trying to manufacture teen stars with mass
popularity, even though Britney’s audience is shrinking.
* Offering one color, as long as it’s black. Successful industries
differentiate products. Yet the music industry focuses on mass
distribution of a single product, at single level of quality.
* Blaming technological threat while ignoring customer boredom.
The last big lull in music sales occured in the 70s, when cassette
tapes were taking off. The industry blamed plastic. Then punk and
new wave killed disco, and sales rose again.

SpamKiller must die

Avoid SpamKiller, the anti-spam shareware utility published by McAfee. Whenever it captures a piece of spam, it utters a hideous, barking, crunching sound, as if Cerberus, the three headed demon guard dog of Hell, had just clamped its infernal jaws around the thighbone of an intruder. Even if you turn off the sound effects, it crashes frequently and hard.
Anyone have better suggestions?