hit counter

I'm currently working on a DPhil in HCT at the University of Sussex. This section of the website is for an on-going 'learning diary', for me to write my thoughts and notes on various courses and my thesis.

Entries in omg (2)


Back above the parapet

After our lab group play session the work went kind of crazy for a bit there. I am slowly surfacing, coming out from the bunker, and preparing to stick my head back above the parapet and share what's been going on. 

Quite apart from the small(!!) matter of attending CHI and getting an application and subsequent paper in for DiGRA (more soon on both, promise), we had a substantial milestone for the game and project last Sunday (19th May). Last year we offered an extra-curricular activity in the middle of the STEPS summer school programme, where particpants could come and play the Green Revolution game. This year we offered the same activity, except this year we were using our game - African Farmer. 

Of course, this meant that all of the interface problems found by the lab group had to be reduced or (preferably) eliminated. Any back-end bugs had to be squashed with extreme prejudice. As much of the functionality that could be implemented had to be, and had to work and be straightforward. The very worst bugs we could tolerate needed to have a work-around. 

We had a major interface redesign after the first lab session, with money put aside to allow us to employ asilia to give us a consistent and much lovelier look and feel. That went brilliantly, but of course meant a reasonable chunk of time to put the new assets into the game. My model/view/controller separation was good, but nothing is going to protect you from a total asset change! On top of that, the bug-fixing and testing cycle was intense, particularly for the last two weeks before Sunday. As we cleared the most obvious bugs more insidious underlying issues became apparent, and slowly, slowly we got through those too. It wasn't a case of late nights and massive hours, just constantly working at a fairly intense level for the 9.30-5.30.

(I also got a cold after CHI. I had it for 2 weeks - a definite sign that I was pushing my system a little harder than comfortable!)

My poor partner had to put up with me being almost entirely unable to form coherent, completed sentences on Friday night. But we were pretty confident that the system would be robust enough at that point for Sunday. 

And you know what? It actually was. We had 15 participants in the end, working in 7 family groups. They came from all over the place, Africa, India, South America, and even the UK. After intros and so on we played for around 3, 3.5 hours, and got through 3 game years. It got a lot faster once we'd got through the first season, but even that first season the difficulties were mostly around decision-making and not our interface. 

At the end of the last cycle our project sponsor John lead the reflective discussion, and we honestly could not have paid for better results. They mentioned almost all of the learning points you could want, decisions around schooling, better understanding of why that didn't happen much, a feel for intensification and the importance of saving seed over providing the best possible diet all came up. What didn't come up was the interface. It disappeared, in exactly the way it ought to have. 

In short, it was more successful than I have ever dared to dream it could be. 

We possibly made it a little tough - more people died than we meant. There are things we could improve, things we might need to include for a fully featured game. The poor game manager at the moment needs to constantly manipulate the database directly, rather than work with the game interface. There's still one irritating occasional bug that needs a browser refresh to clear it. But over all, it worked, it was fairly solid, and it produced the kind of learning and reflection that the table-top games have always caused. 

The only thing I can compare my feelings at the end of the day to is finishing a marathon: exhausted, proud, pleased and a little bit startled to have got there. In reality I'm probably running a 50k (or 50 mile? Hopefully not) ultra, but at least the 26.2 has broken the back of it! 

(This would also be a great time to thank the lab group, who actually came back and played the game again for us and highlighted issues we would never have found without them. Thanks everyone!)


They played it!

The lab group. They played our game! For real!

Well, kind of. Actually, we only got through two seasons (well, and the intro). In about an hour and a bit. Clearly there are some issues there, as I seriously expected it would take a maximum of about 15 minutes per season. But we had 8 people logged on simultaneously, and it did not fall down in a big heap. 

I'm not claiming there weren't errors. Mostly they were fixed by either trying to do the thing again (I really need to check my hibernate stuff, there were unclosed transactions all over the shop and I really thought I'd caught all of them) or restarting the flash player. That's clearly not acceptable, but for the first time we've ever run it as a group I'll take that. 

There was a lot of feedback around the (lack of) feedback - which I was aware of, but not quite aware how pervasive it was. I know this game intimately, it's so hard for me to work out what a new player won't know. There was also a lot of clicking on things that aren't buttons, and mis-reading things that are (exit and home didn't actually mean what people thought they meant). I still don't have drop-downs working 100% well.

So much useful stuff to see and take on board, so much coding to do! But that is a really big step. I think we've been at this for 2.5 years now, from the very first paper sketches of the screens, to our PowerPoint prototype with the Flintstones as our family members, to a static flash prototype and now finally to people logging in and playing. Quite a long journey, and a lot of code. 

What I'm most pleased about was the buzz in the group. They wanted to get past the issues, they wanted to keep playing. I have had this nagging doubt in the back of my mind that even if we made this work we would actually be creating a game that noone wants to play. So if nothing else, I am hugely reassured that this is not the case. 

I guess the next big step will be throwing that in front of people we don't know, who don't know how much effort has gone in or have to talk to us next day. I think it would be nice to iterate and show the group again, and see if we can get that playing time down. I still think that 15 mintues per season is do-able. Just need to improve that interface!