PenguinTutor Linux Website - One week old

2 June 2006

It's now a week since I first started getting visitors to my latest web site: www.penguintutor.com.

I actually made the site available two weeks ago, but only through a link on this web site. I publicised the site by posting to the LPI email distribution list last Friday.

Using the email distribution list was pretty much a baptism of fire for the web site. Although the number of hits to the site was not huge a lot of these were within a short period of time from sending the e-mail out. I believe this is a result of the website arriving into individual email addresses, rather than forums which people will visit at different times of the day. The website did suffer some slow response, and the occasional page timeout, although when I tried I could just click again and it worked.

I was careful to not spam anyone, and just chosen a specific mailing list, which was relevant for the web site. I've also added posts to a couple of forums, but I've replied directly to relevant posts rather than randomly posting to the forums. Forums can be a good way of publicising a website, but only if they are used responsibly and are used to contribute to the aims of the forum.

There was one other minor problem, that I caught fairly quickly. One of the questions was causing a problem. I had tested the question with a separate test function, but as this didn't go as far as caching my answer it did not catch the error.

The reason I was able to identify this problem so quickly is that I have included email notification of all errors in the program that I wrote.
There is a downside to this in that if there are a lot of errors then I can get quite a number of emails, but I can change the threshold. So for example if I suspect an error is because of a session timeout, or because a user entered invalid data, then these could be screened out. Ideally it would be better to have something that checks the error logs and ensures it only sends an email once per type of error, or something that aggregates the errors over a period of time, but that's something for the future.