Monday, October 01, 2012

Gridiron Solitaire #23: Beginning the Beta

The first group of testers received the build Friday night.

I invited 15 people, which was probably too many. That may sound silly, but this isn't a beta with a highly-developed tracking tool for feedback, or forums.

It's a text file and e-mail. Those are my tools.

I know--that sounds stupid--but it's in keeping with what has always been my number one rule: don't  use anything  I don't completely understand. I am 100% sure that I can manage the process this way, and I won't have to spend time implementing any kind of tool that I'm really not familiar with.

Early on, I realized that there were code samples for literally everything.  Some of them were absolutely incredible in terms of their output--oftentimes, they were exactly what I wanted.  I've never used any of them, though,  because I never understood the underlying code.  Even though what I've written is, I'm sure, very crude, I understand how it works, and I can maintain it when something goes wrong.

So to facilitate my Flintstone-era tools,  the beta is going to be in waves of 15 users, basically. Plus, and this is just as important, I see the beta as progression on an archery target, with each wave getting closer to the center. At least, I hope that's how it works.

What I've immediately learned about a beta test is that compared to regular development, it's awkward. Instead of a nice list of features that I'm implementing, or a list of bugs that I'm familiar with but haven't fixed yet, it's like a meteor shower.  Stuff comes in from everywhere.

No one has had a crashing bug yet--I don't think anyone will--and nothing is broken, but there have been a few surprises. Most notably, text on some screens isn't scaling properly.  This is a good example of what happens when something goes from a developer's desktop "into the wild."

Gridiron Solitaire is a WPF application, and one of the biggest benefits of WPF is that it can scale everything automaticaly (using something called a "viewbox").  That means you can rescale the game window at any time, to any resolution you like, and everything scales perfectly.

On my system (default resolution 1920x1080), I can open any of the game windows, scale them endlessly, and everything still looks perfect. Immediately, though, I received multiple reports from people running at 1400x900 resolution that certain text elements were clipped because the text wasn't scaling.

Ah, crap.

This is the awkward part: hunting down things like this can take an enormous amount of time. You're basically just jumping into a rabbit hole and hoping it's not 10 feet deep. Of course, this is the time you want to jump into a rabbit hole--not after the game ships.

There have also been some excellent usability comments. For instance, almost everyone wanted a more visible highlight when a card is selected. That's incredibly helpful, and it took all of 5 minutes to fix.  It also made me realize that the very nice animation I created for matched cards had disappeared a few months ago.

Oops.

I didn't even realize it was missing, and neither did John Harwood. So I found what had gone wrong, and it's working again. 

It's very fun to hear that people enjoy playing the game, and that they're seeing some of the depth of the game world. It's also very interesting to hear some of the features that people want. There's a fine line between complexity and accessibility, and while I want is to have plenty of football elements in the gameplay, I also still want it to be a game that a casual gamer can enjoy. So people have made some excellent suggestions that aren't going to get put in, because it would make the game too complicated for the casual player.

It's also been very interesting to see what kind of information people want about the team they're playing. Several people have said that they'd like to see team ratings in the pre-game broadcast. This goes to the core of how I want the game to play, based on the concept of giving the player incomplete information.

Not incorrect, but incomplete.

An example: the pregame broadcast kid you a rating comparison by using an announcer, but no numbers. So one team's running game might be "slightly stronger" than their opponents. The description changes to reflect the size of the ratings gap, and there are multiple phrases for each level.

If I just showed the explicit team ratings, strategy would be transparent and extremely mechanical. With the text,  strategy is derived and non-exact.

Plus, a player can go into the league statistics screen, and a team's statistics will roughly correspond to their ratings. Again, general information about ratings can be derived, but it's not exact.
To me, that's how real football works. It lacks 100% precision, and there can be large variations from game to game. That's how I hope GS plays, to reflect the advantages and uncertainty of the real world.

Site Meter