The usual pride before a fall:

Ludum Dare Mini 31: Fear – Team and preparation

That mentions our little ‘barebones’ framework we’ve made for our LD entry, and a tiny bit about our team members.

It seems the audio and video kind of lose sync when youtube transcoded it – the original was about 40 seconds longer and somehow it got squashed.  Most atypical of Googlers.  Hmm.   Anyway, excuses aside, its understandable.  Mostly.

I trimmed off a spoiler describing what we’re going to build; I’ll post that and timelapse and stuff when we start our 48hrs, which will be Friday evening European time.

“Mr War” hasn’t much free-time this weekend so he’s been putting some of his 48hrs in in advance.

Shameless blog promotion: http://williamedwardscoder.tumblr.com/

Back in the January mini #31 we tried top prove that platform games don’t need to be made from 2D block tiles!

(Anyone wanting an OpenGLES 2 and file IO wrapper that compiles for Windows, Linux, Mac and runs natively in the in Chrome browser can feel free to use the scaffolding we made for mini #31 – barebones)

This time around we have an overall game concept that is entirely different, perhaps unique, and we’re super excited imagining how we might bend it to suit whatever the LD 23 theme will be!

We might try to make the game play two different ways – sort of almost two games in one.

When we last did a mini Ludum Dare we tried to show that platform games don’t have to be made out of square tiles; we used meshed.  Naturally, we struggled to finish in time!

We wanted to go break convention in a different way this time: resurrect the classic text adventure!

That was the type of game we wanted to build when we were small.

We did do a slideware scrolling illustrated text adventure in Turbo Pascal 6 and VGA mode 13h IIRC; so now, 20+ years later, we’re re-entering the fray…

Here’s a video of our illustrated text adventure:

By using a map top-down we get spatial reasoning for the going north/east/south/west etc

51 locations, 27 objects, 7 npcs, and a classic mode (code mode to come) it’s going to be awesome

LD #23 Jam: “The Small World of Professor Strange”

https://github.com/williame/ludum_dare_23_tiny_world

After 2 of 3 days (but who can do a full day tomorrow on a work day?)

(sorry no audio)

Oh, a really neat SAFE-FOR-WORK mode too

Its mostly finishing the puzzles and adding text descriptions to everything.

You can play on-line already (images load slowly): http://williame.github.com/ludum_dare_23_tiny_world/

Play it online here

After an engineered launch on Hacker News a few hours ago, Google Analytics tells me that we have had over 9000 players and wasted over 100 hours of their collective time so far!   (Yeah, average time/visitor not so high, but hey)

We are super proud of the game.  I am super proud of the game.  I am rather in awe of Mr War’s artwork and the pace he works at.  And he uses MS Paint!  The top-down map for a text adventure gives a spatial feel that works really well.  Anyone ever played a game like it?

Text adventures really bring back warm memories.  And they are shockingly underrepresented in LD, so we tried to remedy that!  We deliberately didn’t want to fall into the point-n-click improvements trap and we wanted an overall Victorian HG-Wells / steam-punk feel to it all.

Back when we were small we used to make text adventures in Basic and later Turbo Pascal with still images in VGA 13h.  I actually can’t remember how we did it, but I imagine it took far more technical planning then.  Nowadays RAM is so cheap we just have big JSON arrays of everything in dynamic languages…

We did a dry run back in the mini 31 in January.  That showed that artwork was our strength but that we were under-resourced programmer-wise (my area) and we completely lacked musicians.

We didn’t recruit any musicians for this LD either.  Tempted?

We prepared the weeks before by acquiring permission from loved ones and deciding the kind of game we wanted to make.  We settled on a text-adventure, using HTML/JS, and the top-down spatial map idea.  I wanted the programmer UI but that didn’t need Mr War, so would have been a solo entry.  Not cool.  And I’m so glad we did the illustrated one; its by far the defining feature of the whole experience.

I didn’t actually know Javascript, so the week before I set about asking on Stack Overflow how to do things like the scrolling that I thought I was going to need.  I got not leads.  I also asked on gamedev about parsing, but the advice there was to use Inform; not what we had in mind!  I settled on super-simple string matching in the end.

This entry was low code high content.  The basic engine – in Javascript – was workable within hours, and by the end of day 1 we were more content oriented.  We panicked a little bit about the actual narrative, and roped in James at short notice to give us some lines.  By the end of day 2 we had 90% artwork done, 90% coding done, but only 60% of the puzzles complete.  You could play the game but not solve it.

Day 3 was of course a Monday – a work day.  It was also the day that my flu kicked into overdrive and I was feeling decidedly groggy.  We didn’t return to the challenge before the evening (Europe-time), just hours away from the deadline.    We managed to get the puzzles complete and get some of the glest community to do play-testing for us.  We submitted in time and went back to our families.

This went super-smoothly.  We ended up with a playable game that can give several hours of fun (if you enjoy slow-paced text adventure puzzle games without bloodshed).

Technically, it has some weak points:

The command parsing is primitive in the extreme.  Its just a list of legal commands.  (You can see them all by pressing shift-space).  I had planned text auto-completion.  Its not the meaning that you have to guess the commands.

The programmer SFW mode doesn’t have the auto-completion I imagined as central to its look&feel.  I experimented with code-like layout of text, but the shortcuts I tried to massage text into code nesting and adding faint punctuation and camelCasing etc made it completely unreadable.  Chopped.  I think with more finesse such a mode could be successful.

The NPCs don’t have the full role we imagined.  We had wanted much more emergent dialogue, and work in how they all knew each other and what a coincidence it was that they converged to retrieve the treasure.  That got cut on time grounds.  And writing dialogue lines is hard.  The NPCs were to have a role in the solution, and be a hindrance, but that never got fully developed.

And we forgot to add the vicar.

With 50+ locations, 30+ items and 7+ NPCs its quite a full enough game anyway?

The suggestion system for rating 20 random posts for me is now chock-full of Windows games, which I cannot play.  If I could, would I anyway?  All that downloading and installing and trusting and stuff… yuck!

So I’ve taken to going down the top-10s people are posting now; if the game was good for them, well worth a look for me?!

And when I go down everyone’s top-10 -type posts for LD 23, almost all of them are WEB playable games!

What does this mean?  Will we be able to work out how rateable things are by their platform choice after voting finishes?  Is there a trend here?

I’ve been stuck staring at Java applets that just show blank and being asked to download Unity players that don’t exist for my Linux OS, and I’ve even borrowed a Windows machine to play a few that looked particularly promising.  But the general Webness is goodness and the right direction!

Haven’t seen any NaCL entries yet though, which is a shame as I tried it myself for a mini once: Afraid of the Dark

My favourite so far is Tondie and Zupe:

Its just so playable!  A complete little platform game.

Something completely different is Tiny’s World

I really loved the music!  I forgave the gameplay, even.  Go listen to it

Tiny Civilization is super-nostalgic for me and I’m not even going to try the new colour version

My own game is a web game, and I’m rather hoping people will appreciate the pretty graphics and the CTRL-B mode-switching:

The Small World of Professor Strange

(nothing like trying to rustle up a few more ratings before the deadline 😉

Go Web Go!

MiniLD#31 we tried to show you don’t need to make a platform game out of tiles.

LD#23 we tried to show there was live left in text adventures.

I think in LD#24 we’ll surrender to popular pressure and make a game that the LD masses will like better 😉

Current hobby influence: ray tracing

a massive multi-player on-line space-fighter beat’em-up!

no pansy foyers and menus; just straight in, spawning right in the middle of crazy space cage combat.

End of day 1:

watch?v=fS-JMsLi2E0

Write the name down: Cage Flight – because you’re gona wanta play it next week for sure! 😉

Play Cage Flight now!

Wow its just been so much fun   After the frantic coding over the weekend, we haven’t really used today to actually code or tweak much.  We’ve just been playing the game to bits!

At times we’ve had 10+ total strangers all wheeling, weaving and diving!

Played Cage Flight yet?  You ought to

The framerate is set by the server, and people were complaining that I’d set it too low.  I’ve reconfigured the server to run at 12 fps, but its hosted in Sweden and is there’s a lot of latency over the Atlantic to contend with.

So I’ve also published instructions on how to run your own server.  If anyone can run a server at a higher framerate in North America, that’d be awesome!  Please post link

Multi-player games need YOU!

We ought to get together a list of multi-player games and, on say the last few hours of voting or some time we can all agree on – rendezvous in-game.

Because for the poor soles rating games when they are alone in the game-would, it must be hard to appreciate them.

I’ll be sitting in my multi-player game hoping you’ll come and appreciate it!

github Game Off for web-apps

https://github.com/blog/1303-github-game-off

Prizes := iPads;

I figured people want something to do between the October Challenge and the next LD in December

(I’m not affiliated with them in any way, I just saw the announcement n proggit and thought ‘cool, LDers will want to know!’)

A very nice write-up about the never-released Hard Vacuum RTS game http://lunar.lostgarden.com/game_HardVacuum.htm

And, in excellent timing for LD25, the artwork is up for grabs!

Jam entries can use existing artwork etc.

And recently Command&Conquer was ported to HTML5 http://www.adityaravishankar.com/projects/games/command-and-conquer/

Surely all the pushing and hinting you need, right?

I usually develop my webgl entry on a VPS and then publish it on github-pages.

For this purpose, I have a little python web-server that I run on my VPS which serves its files from a git repo.

Rather like github-pages, but without the build-times and aggressive caching headers.

Ideal for quick development.

I have now added some very neat collaboration features:

• you can download a zip-ball of the project easily from a URL
• you can upload files to the git repo!  So, your artists can load new/edited content into the game, and then the game can upload that to the central server!

I hope this is as useful for you as I hope it will be for my jam team!  Feel free to use, abuse, profit!

tiny1web.py

Hmm that’s a working title.

Here’s how the editor/player looks at end of day one.  Now, today, I work on play mode

And here’s a video of me clicking around in the editor:

http://www.youtube.com/watch?v=7O5PpYcFSOc

Wouldn’t it be nice if you could click on a randomise button on the ratings screen and it swaps out all the choices?

And blacklist/whitelist popular tags, so you can say you aren’t interested in those that have “windows” (and “source”), but are interested in those that have “web”… and so the ratings screen will be full of games you actually stand an outside chance of playing…

I expect this has all been said a thousand times before…

We need a mod-ludum-dare-itself mini weekend

(And thanks to everyone who has rated my game, what a great turnout yet again for LD

Opening myself up here, but to be honest I’m deeply disappointed about our LD score history. I expected more of us; we aimed higher, and we hoped to make something that would please and wow the crowd.

And we abjectly failed, again.

We’ve always aimed to challenge norms; in LD23 we’d tried to put graphics into text adventures in an interesting, new way. We’ve tried to enter a multi-player on-line game in LD24, and we’ve tried in LD25 to make a 2D platform game without tiles.

We really are overactive underachievers.

Our highest LD25 score was for music. We were perhaps not clear enough that that was the one thing we had not ourselves made It was a track we’d picked up on the Internet, and we thought we’d explained that.

So why did we fail so miserably? Bad control.  We should have had a lot more and we lacked in plenty of areas, but I think that the playability was most affected by the bad control.

Its fairly non-trivial to do collision with arbitrary lines and so on, and do it efficiently too. Our collision code, which we’d written for the game and mostly as the deadline loomed, was immature and incomplete. We should of course have learned how to code it before the jam, or borrowed a library or something.  But we’re not really borrowing-code-libraries kind-of-people.

And what happens is that you do badly in a core area such as control and people voting just give you a low score across the board. We got 1.92 for innovation which just means we hid the code well I guess.

I’m deeply saddened that we didn’t score higher for graphics, as I myself am in awe of my team-mate’s artistic prowess and thing it deserved more. A very big part of that is my fault, as I also didn’t get around to good GLSL lighting. Meaning the whole making the scene out of meshes thing didn’t work, and it was hard to see where the platforms were because of the uniform shading.

In previous LDs the winners have always wowed me. I’ve always looked down the winners list and been awed and inspired. This time its been a whole lot, well, distanced. I don’t mean to take anything from the winners and everyone else in on the game; I think its in a large part because I didn’t find the theme exciting.

We had had an alternative game that was a FarmRomanVilla thing; Zynga being villains and villain being a medieval French word for farm worker and all. All actually inspired by a recent episode of Time Team I’d watched where they excavated a Roman villa in Britain…  But whilst that was very much the kind of game I like making, I think it would have disappointed just as badly as the lets-make-a-platform-game,-they-always-do-well game that we tried.

Its been a introspective start to the new year. Our strength is art and we do ok-ish at coding too, but general story and game-play, audio (we’re both terribly untalented musically) and overarching fun we fail miserably at.

I know I’m supposed to end on a high note of future optimism and enthusiasm and a promise to try harder and come back next time. But its hard to muster the bravado. Time heals, so maybe we’ll be back for another beating; we’ll see.

yay!  I’m the first on mini#40: http://www.ludumdare.com/compo/minild-40/?action=preview&uid=10313

Its a joint effort – already got lots of help from the LD IRC channel! – so do head over and help code/art/sound/gameplay/concept etc! 😀

I can’t enter LD26 but, rather than complaining and doing nothing about this travesty of priority, I’ve done my own little LD in my evenings.  An hour here and an hour there and, after about 2 days worth, this is how far I got with my city-building game (video):

Now I must press on and make a proper full game of this!  Its a big project, of course…

I ask for your help!  How would you simulate a city, at a very deep technical algorithmic level?

I’ve played around with a “priority flood fill” so that, from each zone of a type, it fills outwards over neighbouring tiles using its score as the priority.  There’s a cost lookup by tile type so residential influence travels further along roads, for example.

This is very pretty and its neat but is it useful?  Will a large building with a high score ‘drown out’ nearby low-scoring buildings?  Will the influence of a high score building cause it to reach lots of target zones – e.g. residential finding industrial – quickly, meaning it sates the needs of far-off target zones at the cost of nearer but not-yet-high-score zones?   Will it mean a concentration of score?  Will it be intuitive?

What ways can you think of of actually computing, each sim step, if a zone should increase or decrease score?

My barebones.js webGL library now has a basic heightmap-based terrain and simple camera system:

It is hopefully simple enough and isolated enough that you can borrow it, adapt it, or learn from it without needing to adopt barebones.js.

It does top-down and 3D views, and makes it easy to work out where the user clicked and to update the map tiles.  Grid is optional.

Each tile can have its own texture etc.

So go make an RTS or something!

Good luck!