« February 2011 | Main | May 2011 »

March 24, 2011

Is there a limit to network-effect benefits?

In recent months I’ve been finding Stack Overflow and the related Stack Exchange network sites to be a tremendously valuable resource for resolving technology problems. It’s not that they can always answer my question, it’s that frequently someone else has had the same question before, and I can piggy-back on the answers they got. Searching Stack Overflow, in other words, is often more useful than asking Stack Overflow.

The value of Stack Overflow as a Q&A site is the huge number of people using it. For any given Ruby on Rails question, for example, there’s a pretty good bet someone among the thousands of users scanning those questions will have an answer. Things get a little thinner when you get to very new technologies like SproutCore (for a while I was among the top 20 answerers for SproutCore, which says more about the traffic in that tag than it does about me).

However, as Stack Overflow grows, the number of questions seems to be overwhelming the number of answers. I’ve posted two questions in the last two days, and as of this writing neither has been “viewed” by as many as ten Stack Overflow users. This isn’t because the questions are unanswerable, I think; it’s that there are so many other questions to answer, mine have been buried almost immediately.

We always say the value of a network grows with the size of the network. But Stack Overflow is suggesting to me that there might be a limit to that rule. If the network becomes big enough that messages get lost, the value of the network may begin to fall as it gets larger.

The Area 51 site where new Stack Exchange sites are suggested, debated, and spawned seems to aim at building a critical mass of users to make each new site valuable and useful. There isn’t an internal control for sites which get too big and therefore lose value; I wonder how that could be created? Is this a big enough problem to bother?

Posted by pjm at 8:01 AM | Comments (0)

March 10, 2011

State by State

There are a lot of things I could be writing about right now. This is warm-up (although the main event may not be published here).

At some point when I was in college, I learned on the Dead Runners’ Society listserv about SEXY-LU points. I don’t remember now what the abbreviation stands for, other than that the S stood for the last name of the person explaining this scoring game to the list. (He shall remain googlenonymous here, but Ed, I do remember.)

A runner would accumulate points for a given span of time, generally a calendar year, although I imagine one could accumulate lifetime points as well. Each state and/or foreign country in which one ran in that time counted for at least one point. The actual score for each depended on how much one ran in those states, basically boiling down to the exponent needed when expressing the number of miles in scientific notation:

  • 1-9 miles: 1 point
  • 10-99 miles: 2 points
  • 100-999 miles: 3 points
  • 1000-9999 miles: 4 points

A serious runner would generally pick up four points for their home state (I doubt anyone has scored five) and one or two for most places they visited. A good year for me when I was doing a lot of track traveling would be in the mid-20s. In 2010 I think I hit 16 or 17, and I may have done as well in 2009. (I don’t recall if different Canadian provinces counted for extra points, but that hasn’t mattered for me since 2001, the only year I visited more than one.) I haven’t tried to calculate my lifetime score, but I started thinking about it when I read this post by Scott Douglas.

The drawback to this program is that a weekend in Japan counts for the same number of points as ten miles in Connecticut, but on the other hand, a quick run in the Denver airport can pick up extra points as well.

Posted by pjm at 8:18 PM | Comments (0)