It’s been kind of quiet around the blog, twitter and everything so here’s an update for you.
We’re working on changing the design/build mode and the way walls are put up. A new set of rules will be put in to effect and wall pieces will be placed in sections, sort of like tetris..
The new rules will not permit structures to be totally closed, or be made up of too long interconnected sections. Corridors will also be two tiles wide and every part of a design will be reachable.
It might sound like it would be boring to build under these rules, but this change is to make the game more fun, both when defending and attacking others. As much fun as building 8 layers of walls seem to be for some, it doesn’t really contribute to the game in a good way. Another upside with this new system is that an indestructible wall type can be used to build stuff. Fortresses will become more like mazes of death if this works out the way we hope!
You can try our simple mockup and get a feel for it (press space to flip pieces).
Enemy waves will be totally remade as well, to accomodate the three new enemies we will put in the first release (more will follow). Right now the initial difficulty is too high and will be lowered. Weapons will be more powerful and get a better reach, maybe cheaper as well.
But all this will also mean a wipe out will occur, and all progress will be lost. Only purchased lives and level sizes etc will remain.
We’re right in the middle of all this and can’t quite tell when it will be released, sorry!
Anim test of the four new enemies in the next update of Farm Fortress!
Motherbug, Flybeetle, Drillbot and Archeroid.
While working on all the new enemies, I also have pushed a few updates.
The main thing about them is that prices are lowered for lives and the starter pack is more value for money - you get 20 lives when buying it now. Price is the same. Also there’s an option to buy a very affordable 50 pack of lives for 8 bucks.
Fixed a couple of bugs, one was pretty severe and would occur when exiting the game while in the shop.
If you buy anything ever, you wont get that nag screen suggesting you to buy a starter pack.
We changed server configuration a little and scaled it down a little to halve the costs, but there’s still lots of capacity left. The real problem is to get people in the game, not scale it as it stands now.. Our hosting costs are around 100 USD / month now, and what the game pulls in does not quite cover it yet. But we still have money to run these machines for well over a 6 months now, so there’s no worry. As soon as we figure out a few good ways of getting traffic here we know it converts to money.
Feels like just a day a week is too little for this project, but I’ll keep working on those enemies and keep you updated!
Edit: A nice long thread about it produced on reddit! Check it out for a few answers.
Yesterday our Farm Fortress tweet was RT by mighty @notch, sending it past his some 800.000 followers. Today I thought I’d share what came of it and what it has meant to our project (well short term, since this was yesterday). Here are some numbers!
August 16-17 (24h)
Rather immediately, our web front end machine chocked out of low memory. This was discovered to be caused by two things.
Firstly, I had forgotten some (soon to be added) database request code on the front page. The generated queries were not even optimized yet, as the function wasn’t meant to be online.
Secondly, web server is hosted on a VPS which does not allow you to have swap space. This may have been the worst factor. Memory quickly ran out and I could not send a single command to it, or connect via ssh. This affected the game since it requires a http response when it loads a list of game servers of the web server. Obviously the page embedding the game could not load either. The game data was loaded from a CDN.
After a while, I found an option to temporarily boost memory for the machine on the fly, and that help getting the site responsive again. I could then also remove the unused but heavy database queries.
When notch retweets something, within seconds he sends a horde of 1000 or so users (and bots) to jump on your site and try to load it. This sudden wave should be manageable, but kills a lot of small sites.
Some really good news - it looks like I underestimated the load on the game servers, making the game a (potentially) better business. Technically the game can be compared to a very dynamic and persistent multiplayer 2d counter strike, for each player the server runs a full game simulation (even if player is alone in the fortress). This means grouping players together improves performance, but we can also handle a lot of “loners”. At most the game servers had 50-60 players on them, but causing very little load. A cheap AMD Operton will easily handle 150 instances by itself, maybe even over 200. Hopefully we’ll see that one day.
Having a small project and getting Notch’d is the equivalent of throwing a cup of gas on a little fire. It’ll burn real nice to start with, but pretty soon it will need some better fuel to keep burning.