|« September 2014|
I've been amazed at how well Google's news service works. But no matter how good the technology is, occasional mistakes are inevitable when computer programs try to compile and correlate news headlines from lots of different sources. Today, some of the seams really showed.
In the "Top Stories" section at the top of the page, there's always a subsection called "In the News" that contains a short list of topical links: topics that seem to be getting a lot of coverage, but haven't made it to the prestige positions that include headlines and pictures. The links in "In the News" aren't headlines. Instead, they're topic keywords that have been extracted from the headlines by the Google software. Things like "Tel Aviv," "Harry Potter," and "NATO Summit."
Today I noticed three topics in particular: "Our Man Flint," "Magnificent Seven," and "Academy Award." It's clear what's going on there -- news outlets are writing about the death of James Coburn, and Google is picking up on references to his achievements and most famous films in the headlines. But when I clicked on the topics, things got even more interesting.
But I was really surprised when I clicked the link for Magnificent Seven. Just looking at the first ten hits, I learned:
I don't mean to take anything away from what Google has achieved. All things considered, it works amazingly well. And quite frankly, occasional strange juxtapositions like this can be good -- they add an element of the serendipity that's present in a real newspaper, where you can occasionally run across a fascinating article that you never would have looked for.
Think about it. I'll probably watch at least one of those Coburn movies on TCM Sunday night. The story about the horses was interesting, and I was surprised to learn that five years have gone by since the McCaughey septuplets were born. And it's interesting that the producer of the film and one of its stars died during the same week.
I learned one more thing, too. All of those stories included the words "Magnificent Seven" -- most of them in the headline. The name of the film has entered our language. That, in itself, says something about the legacies of James Coburn and Marvin Mirisch.
Part of the answer can be seen in the way I started blogging. This entry flowed from my fingertips one morning, almost fully formed, and I thought it was fairly good ... but so far as I could tell it wasn't sellable. Many of the things I write are that way: they're very small pieces that would be difficult to turn into full-fledged articles (and I wouldn't have time to do it anyway). Thinking about that led me to approach O'Reilly about a blogging spot on their site.
Another part of the answer lies in the heading on my "blogroll" over in the right margin. A few weeks ago, discussing blogging with Mike Clark, I described the blogging community as "like Usenet, but with only the interesting people." I like that phrase, and it's a pretty apt description. I recall that even in the heyday of Usenet (i.e., before AOL and spam) I tended to gravitate toward posts by people who were interesting. (Of course, dull people blog, too. But the point is that you can choose which blogs you read.)
Communities like that are important. So important that people have to keep reinventing them. After Usenet degraded, mailing lists experienced something of a resurgence, as did IRC. For a while, in the software development world, Ward's Wiki filled the void. These days, much of that community seems to be rebuilding itself using weblogs.
All of these things are just different ways of extending our communities of interaction, finding other people to be a part of them, and increasing the opportunities for sharing ideas. Five years in a row now, at JavaOne, Duncan and I have made it a point to have at least one extended meal together. It's a great opportunity to share ideas (baked or not), things we've learned, things we're working on, and things we're thinking about. Each of us always comes away with a head spinning with new things. Dave and I have lunch every two or three weeks, for similar reasons. Greg Vaughn and I regularly have extended IM chats that help keep us energized with new ideas. Dave, Chris Morris, Joe Tatem, and I recently organized the Dallas Pragmatic Practitioners, and it's the best "user group" I've ever been a part of -- largely because there's so much opportunity for informal discussion, where you learn so much from the ideas and experiences of others. (I can think of several other people or groups I could put in this list, but I think you've got the idea.)
Blogging is one of the ways I contribute to this little "idea ecology" that I'm a part of.
Here's an immediate example: last Thursday, Mike Clark wrote some interesting things about EJBs, and Duncan responded. The combination of those two posts resonated with a discussion we'd had at the Dallas Practitioners meeting two nights prior. (Synergy!) Until today, I've been too busy to write my own response knitting the three together, but you'll see it next.
It does seem that they're using something at least vaguely similar to PageRank -- there's definitely a feedback process involved to assess the importance of various stories. One of the nice things about this is that Google frequently features stories that are off the map for mainstream American media. I'm sure this is partly due to Google's inclusion of non-U.S. news services, but I suspect it also reflects an American population that is changing faster than the newsrooms are.
Today, I was treated to not one, not two, but three stories about Cricket on the Google News front page. What a nice surprise. (And hurray for Australia! With Shane Warne looking more powerful than ever, they're in great shape for the Ashes next month.)
Come on, people. Google's ranking system is based on heuristics, and sometimes those heuristics will give good results, sometimes bad. Google is a huge improvement on what went before, but the difference is just that their heuristics are better (and, incidentally, much more expensive to calculate).
Furthermore: it has to be that way. Given the weak semantics of HTML, the different ways of writing and structuring web pages, and the very different things that different people think are important, there's no "one true way"; heuristic ranking is the only thing that makes sense.
Nokia has one of the best products around for building that future: Nokia Rooftop. It's tailor made for community networks, with good management features and exactly the kind of mesh routing support that Negroponte talks about.
Sadly, though, just five days before the Negroponte article hit, the BBC was reporting that Nokia had come out with this statement: "anyone using bandwidth without the permission of the person paying for it is simply stealing."
I can see their point. But for Negroponte's vision to work -- and I sincerely hope it does -- the granting of permission has to be easy and automatic. Negroponte doesn't mind anyone accessing his wireless access point, and he certainly doesn't want them pestering him to ask for permission. But right now there's no way, short of knocking on the door and asking, to find out whether you're allowed to use an access point or not.
Instead of attaching the "criminal" label to a lot of people who mean no harm, Nokia should instead be active in efforts to enhance the security of wireless access points, with the aim of making publicly-accessible lily pads a legal reality, while giving those who want to restrict their networks easier ways to do so.
In the process, they would help build the market for one of their best products.