Stepping back a moment

Stepping back a moment

Okay. Lets step back a moment. If I really wanted to cheat, I could spread my username and password far and wide. Indeed, I seem to recall that at one point, there was a subscribed "cypherpunks" login that did just that. Let us assume Mr. Corbet, rather than announcing the cypherpunks login on the front page, decided to ban all people who logged in more than a dozen times in a week. At that point, I could grab an entire issue of LWN with wget, and forward it by e-mail. Worse, I could stick it on a pirate web site, or even Kazaa. Mr. Corbet could then embed watermarks and infiltrate pirate networks to figure out who was cheating. I, in response, could subscribe from multiple accounts so that I could try to remove the watermarks... To make a long story short, if people really want to cheat, they will. Right now, I think most of the readership wants to help out Corbet. Indeed, many of us donated before subscriptions, and are now subscribing. I don't believe cheating will be a significant problem.

The problem LWN has, and this is pretty fundamental, is illustrated in this graph:

If those two lines tracked, Mr. Corbet would be well on his way to buying a small yacht. Sadly, they do not, since there is no convenient way for new readers to find out about LWN (as in, find out about the superbly high level of articles posted on a regular basis, rather than the mere existence of Yet Another Linux Site). While it may be possible to milk some additional money out of the current readership, that amount is pretty minimal. To be truly sustainable, LWN needs to find some way to draw in the additional readership from members who joined the Linux community in the past 3 years, as well as those who will join in the future.

I don't know the best strategy for this. A big part of it is to let people who are not members of the LWN community view some content. Sharing links is part of this. Another idea is a free trial subscription (on the facist end, this can be done by verifying credit card number; middle ground would be verifying non-free e-mail address; hippy end would consist of verifying any e-mail address, and having cheaters need to go through the hassle of getting a new e-mail address every few weeks).

Again, I do not know the best strategy. If I were Mr. Corbet, however, I would try opening up the site a lot. I'd probably go for the extreme -- links to articles can be viewed by anyone, and only the main pages are locked (so people cannot find the articles without subscribing or having someone forward a link). People can get a free 2 week subscription by verifying their (potentially free, potentially only non-free) e-mail address, but people cannot use the same e-mail address more than once a year. I would then monitor:

• The level of abuse. Referrers could point out patterns of cheaters in links being accumulated on some web page. Free subscribers using 26 different Yahoo e-mail addresses could be tracked with cookies.
• Growth vs. shrinkage of the number of subscriptions.

If the experiment fails, so be it. If it succeeds, Mr. Corbet gets his yacht (or more likely, a decent level of income).

(On a sidenote: I think the free subscriptions idea people suggested below is brilliant -- give out a free 3 month subscription to new FSF/EFF/LUG/Slashdot subscribers/etc. under what was listed as the drug dealer model...)