Here it is folks, My Nine Lives sprint
to prototype postmortem analysis.
I'd like to divide this assessment into
three major sections, time management, art, code. I would have
included a fourth 'design' category but this is a game that I had
previously designed and so during this week I didn't do any design
work. Check further down for my TL;DR section.
Time Management
When I first thought of doing this I
had planned to give myself 48 hours and keep the schedule of a game
jammer. Two of my three game jam experiences went over extremely well
and I liked putting a game together before I could get bored with it.
However, when I first sat down to write this dev blog I decided that
I should instead give myself a week. I was, after all trying to
achieve a few things I had never even attempted before and I would be
doing both the asset creation and game implementation. Plus, I would
have to have time to cook meals, shower, work out, and generally live
life. I couldn't abandon all of my other responsibilities in order to
indulge myself.
I started Tuesday night. I was
confident at the outset that I could tackle the pixel art goal I had
given myself. Designing a character and sketching it are my
specialties and so that took all of 5 minutes to iterate through a
few designs, settle on one I liked, sketch a walk cycle, and scan it
in. I looked up a pixel art tutorial and got to work. I spent twice
the amount of time I wanted to spend on pixeling and animating my
avatar. I laid my first pixels the morning of the 15th and
finished the night of the 16th. Two full days spent on a
character animation the didn't even make it into the final prototype.
From a purely, 'did I reach my goal?'
stand point making the pixel cat and animating it's walk cycle was a
complete waste of time. However, from a, 'did I learn pixel art?'
stand point it was a huge success. I can now definitively say that I
understand more about pixel art and how to implement it than I ever
have before. More of this will be covered in my art section later.
The 16th was a Thursday and
the last day I worked on creating the art for Nine Lives. I spent
most of Friday backing up files, renaming files, flipping my sprites
so my avatar can walk left, and reading about Game Maker. I even
completed a few tutorials but they were all drag and drop tutorials
and left me feeling like I didn't really know what was going on. I
hardly worked on anything again until Sunday when I sat down and
really dug into the code.
I hadn't coded for a few weeks and
certainly not in an environment like Game Maker so by the time Sunday
was over I felt as if I really didn't have any chance of completing
this game. I posted about how it was a learning experience and that I
may have bit off more than I can chew. I was mad that I had to post
that and so the next day, with two days left, I went for it. On
Monday I was able to get my cat walking correctly and have it's other
8 lives follow it. Then, I got it jumping, kinda. I had a made a few
mistakes in my movement code and I was getting stuck in the floor.
Tuesday rolled around and I was able to work out the bugs. For all
intents and purposes I emerged victorious.
Overall, I believe my time could have
been managed much better. For one, when building a prototype, get the
game play done first. The art wasn't as important to the game at this
stage as the programming and design were. I'm happy that my prototype
works and that my avatar is a pretty cool cat but I would rather have
it working with no avatar than an avatar with a broken prototype.
Also, learning new things takes time. I know that sounds obvious but
what I actually mean is that in this project I not only tried to
learn two new things but I tried to execute them as well. I went in
cold and could have ended up coming out cold. Luckily for me that
wasn't the case. However, if I had learned the tools before hand and
then tried to prototype this game I think I could have done it much
faster.
Artwork
I've always had
an aversion to doing pixel art. It's time consuming and required that
I not only be good at drawing, color theory, and be familiar with how
certain digital image files work but it specifically required me to
be good at working with pixels. This may seem incredibly obvious but
pixels are an incredibly restrictive media. I believe that under
tight restrictions true creativity can grow but damn is it hard.
I've practiced
and worked most with traditional media. I have sketchbook after
sketchbook filled and resting on the shelves around my desk. Last
year I took a color theory class and expanded my work out of
graphite, charcoal, and ink into colored media such as colored
pencils, water soluble crayons, oil pastels, water colors, and
digital media. Even after years and years of practicing and honing my
art skills I was still feeling timid to try a new media. Going from
an intermediate level to the very beginning level at something is
daunting.
Wednesday morning
I put down my first pixels. I made the canvas 256x256. It's a
generous size when doing pixel art but the more room you have to work
the easier it is to draw something and I thought I'd apply the same
aspect to my cat avatar. I scanned my sketch in and sized it so that
I could put it on a lower layer and use it as a guide. If you check
my earlier posts I link to the tutorial I followed. By the end of the
day I had created my avatar with some arms and legs.
The entirety of
the next day was making the eight frames needed for a walking
animation. By the end of the second day it was no longer difficult to
get the shapes I wanted. I understood how to create curves a lot
better and it was becoming more natural. However, when I went to put
my avatar into the game I noticed something horrible. My avatar
didn't take up the entire canvas. It's was actually 121x208. I didn't
have time to fix it as I was a day behind my original deadline to
start coding.
I think the next
time I do pixel art it'll be much easier. Since I plan to take this
game idea further out for a couple of weeks I'm sure I'll have plenty
of practice. I'll have to remake my avatar in an acceptable and
probably smaller size. Pixel art is tedious but it's extremely
rewarding.
Programming
This isn't the
first game that I've coded. I recently completed all of the
coursework necessary to graduate with an associate of science in game
development. The degree program is mostly a computer science degree
with a bunch of multimedia sprinkled in. I have experience in Java,
C++, C# in XNA Studio, Action Script 3.0, and Unity's Javascript.
I've never attempted to learn a language on my own. I've dug deeper
into a language than my coursework would have lead me but that's it.
I know GML isn't supposed to be a difficult language and for the most
part it wasn't.
The first hurdle
I had to overcome was how variables and objects were made. Variables
can be assigned value without any conventional declaration. There's
no variable keyword needed. You can simply type, “cats = 9,” and
a variable called cats will be created that stores the integer 9. At
first I wanted my cats to have certain member variables and in any
other language I would include that in the cat object. However, in
Game Maker I had to include that in the creation event so that when
it was instantiated all the variable were created at certain values.
The second hurdle
was movement. From C# I'm used to the top left corner being the
origin. In Game Maker I'm not entirely sure. When checking if a spot
6 pixels to the right is free you don't have to add the sprite’s
width to the x coordinate and then add 6 you just add 6 to x. I don't
know why but that's how it's done. Same with the y coordinate. That
tripped me up more than I'd like to admit. It caused me so many
problems with getting stuck in the floor, thinking my place_free(x,y)
function was returning the opposite values it was supposed to, and
just generally causing mayhem. Thankfully, I spotted the problem
fairly quickly.
The third hurdle
was a data structure hurdle. A class I'm sad to say I didn't do very
well in. My cats were governed by an invisible cat manager object. I
wanted the manager to dictate how they interacted with each other
while the cats controlled themselves individually. This ended up
giving me a lot of head aches. At the start of the room the cat
manager would instantiate a number of cats equal to the number of
lives the player had and designate one as the active cat. I put all
of the cats into an array and then gave them a member variable called
index that reflected there spot in line. The active cat would be
index 0 and the last cat would be index 9. I did this because
originally I wanted each ghost cat to follow the one in front of it.
They could ask the manager for the cat in front of them by there
index. However, that created a huge mess.
The index didn't
correspond to there index in the array. So when each cat was asking
for the index in front of them they were actually asking for that
spot in the array which wouldn't necessarily be the next cat in line.
So then I tried doing it a different way. When I switched cats I
wouldn't assign a new value to a member variable but would rather
temporarily store the cat elsewhere and then shift all the cats
forward eventually ending up with the first cat put in the last
position. As I was implementing this I realized what I was actually
creating was a queue. I quickly googled to see if there was a built
in queue and luckily for me Game Maker had one.
My code instantly
became easier to read and all around more elegant. I simply de-queued
the head and the en-queued it. The biggest problem now was the code
to get the cats to follow the cat in front of them. Queues don't have
indices. I couldn't access each cat in the queue unless it was at the
head or the tail. Fortunately for me it came to me at the gym. I just
had to give each cat a member variable of index again. Then, when I
de-queue it I add the number of lives to it and then en-queue it.
Basically, if I de-queue the first cat (index 0) and then en-queue it
and add 9 (the number of starting lives) it's index becomes 10. Now,
instead of having indices of 0-8 I have 1-9. This way each cat could
ask itself how far it's index is from the head and follow from a
certain distance. Using this formula the relative distance would
never change. Even if I lost a life and ended up with 2-9 (8 lives).
When I switch again I'll have to add 8 (the number of current lives)
to 2 and end up with 3-10. Their indices would climb and climb and
climb but I could always check and make sure they never went too high
and if they did I could simultaneously set them back down.
My last piece of
coding was the inventory. This was exceptionally easy thanks to Game
Maker. I just had to specify what would happen if the active cat
collided with my powerup object. One of the events that is collision
and then you choose the object. It's fantastic. When the cat collided
I executed a code. Basically, destroy the object, set the powerup
variable to “powerup” or literally any other variable. Done. Now
that cat has a powerup in it's inventory. When I press certain keys I
just need to test what that powerup variable reads and boom, I've got
special powers.
I'd have to say
that Game Maker, although limiting, does make things super easy. I
learned the language and prototyped a game in a week. It has it's
quirks but I think it's worth working in. I'd say it's a great tool
for anyone that wants to make games rapidly.
TL;DR
- I should've spent less time on art
- Postmortems take a long time to write
- Learning new skills as you go is time consuming
- Pixel art is tedious, limiting, and rewarding
- GML variables are weird
- Data structures are important
- Where's the origin of my object?
- I'm continuing development for a few more weeks to see where it takes me
