Brian Crecente has posted what I have been saying for a long time — there is a fine balance between timely and accurate reviews.
However, LittleBigPlanet has galvanized the point so clearly. It's been three days since LittleBigPlanet's official launch and the servers are still down, meaning we cannot get online, play online or publish levels. In other words, it doesn't work and EVERY LittleBigPlanet review out now is bullshit - not worth the digital pixels they're virtually printed on.
I could have written a review by now; I was lucky enough to nab the game on Oct. 11 and have played it to death by now. I thought of writing an offline-only review, but do not like that idea at all. So I have waited ... and waited. I will write a LittleBigPlanet review for Blogcritics eventually, but I sure as hell won't write one review when I cannot even test out the online features. My first steps online in the game were not good.
Brian's article brings up a very important point: often writers review by beta code, or test on limited servers. Of course, the experience on a server that is not in a production environment is going to fair well, as you do not have throngs of people hammering it.
This brings me to a point I would like to make about developers: Few seem to stress test anymore and if they do, they're not doing a good enough job. I felt pretty good about LittleBigPlanet because it had a beta test period. This gives the developer time to see how the game fares with large amounts of people connected to it in a real world setting, not a closed network that has little to no latency. Still, a limited beta is not an accurate test of a production environment and the game has not fared well at launch.