Logs of Lag ready to go

My Logs of Lag site is ready to launch. It works well, a clean and simple way to drop a League of Legends logfile onto a web page and see some statistics.

Not sure how to properly “launch” this yet. Success would be some visibility on the League of Legends Reddit, I’ll probably post it on Monday but it’s a crapshoot of the first 15 minutes whether it gets upvotes or not. I don’t expect an amazing response, this kind of tool is really something more useful to hang around and be available for months or years when people need it.

I’m pretty happy with how the final project came together. I was worried for a long time about the visual style. I still don’t think it’s great, but adding a basic background image texture and switching the fonts from Helvetica/Arial made a big difference in making it feel “real”. And I don’t want to do the over-produced glitzy graphics thing that’s popular with most gaming sites, I find that a useless distraction.

Ideas for future features:

  • Highlight “bad” data. Ie; if the loss/second is greater than some threshold, make that show up in big angry red warning font. I want to collect some logfiles from users to see what is typical before building a classifier.
  • Parse broken logs. In particular the logfile gets wonky if the game disconnects, would be nice to parse that correctly.
  • An analyzer for a whole directory of files. This is the way lolparse.com works, presenting summary statistics for gameplay in many games. It’d be useful to characterize lag for many games as well. Totally feasible given all the work is client-side, it could be fast. A histogram showing the distribution of median ping and median packets/s loss would be a good start.

And ideas for the server:

  • Replace the Python CGI with a real server. I have a Node version but I’m not sure I like Node. Porting the CGI to Gunicorn would suffice for scaling.
  • Or.. Remove the server entirely. Just make the client Javascript upload stuff to S3.
  • Or.. commit to a real server, and add logins and user data management.