Hacker News Reader

Comment by qzervaas | Hacker News Reader

Author here, if you have any questions about the game or blog post, I'll be here for a while.

This is interesting, especially the bit about getting featured right off the bat, because I'm just getting started with mobile game development and had for long time an idea for a pretty similar game. Similar in the sense that it was a hex-tiled spelling game where you need to string together words to get somewhere.

However, my partner and I are looking at a different game right now.

Also, I know there are worse reviewers than this, but man ...

"The vocab in this game needs a serious update, I have no idea which words to go for due to the fact that it does not have some retry basic words built in. Not toention the selection process needs a tad bit rethink. Overall it is a pretty great idea, it just needs to be fully realized. I would give it more stars but I'm a pass or fail kinda guy, and this did not pass... Not yet anyway :)"

Yea, it can be confronting reading reviews.

There's over 250,000 words in our dictionary (that's not to say you can necessarily make all of them in the game).

Most people just query "Zen" not being in there though ;)

Are you using SOWPODS? I had thought to stick with TWL being somewhat of an old school Scrabble player. Also, little technical question if I may: what sort of data structure do you use for your dictionary?

We're using CSW12.

It's stored in a single-table Sqlite database with some techniques to make word lookup really fast. I believe it could be faster still, but this is likely beyond my technical expertise.

Next problem is to scale the game to multiple languages without bloating the download size of the game too much. Haven't decided what's best yet for this: either compressing the dictionaries for distribution (and decompressing on first launch), or an in-game download (this seems like so much effort though and just adds more friction)

I wasn't aware of CSW but from the size I guess it's similar to SOWPODS.

An Sqlite db sounds like the sweet spot of performance and easy of use.

Couldn't you automate the in-game download? I.e.: If the user selects another language download it on the spot with a little popup telling the reason for the delay?

All options are being considered. I like that idea in theory but it just adds to the complexity:

- More technical debt - More points of failure - Need to host the data somewhere (plus pay for it... the App Store is basically a free CDN)

Off the top of my head, it may be possible to do a Free IAP where Apple hosts the content, but I may be totally wrong on that.

Regarding the drawing delay on older devices: Is there no way to start that process concurrently while the countdown is running? It looks like you could perform 50%-75% of that work before the countdown even finishes.

It does start immediately when the countdown happens, it just can't do it quickly enough for the number of tiles there are.

There's three modes in the game: Escape, Survive, Explore. The Explore mode uses a much smaller radius and no countdown, and doesn't suffer from the same issue. Instead of thousands of tiles on the board there's hundreds.

It was interesting playing with different mechanisms for drawing the tiles though. Between pre-generated letters, real-time letter drawing and everything in-between, the solution we have works well on most models of Apple devices.

Some of the issues were also with how SpriteKit did certain things (such as how inefficient cloning a text node object is)

How did you iterate on the game mechanics in the early stages of development? Was the original version of the game significantly different from the final released product?

It's actually remarkably similar now to what it was really early on. There's a lot more refinement and polish now.

Here's a video from about 2 weeks into the project (it was codenamed Fractured at that time):

https://www.youtube.com/watch?v=BeMt9UJIPS0

One key mechanic in the game now that isn't in this video is the auto-panning after the user makes a word

(there's a bunch of other stuff too, like the timing mechanism, size of tiles, size of game grid, etc.., but that's more gameplay rather than mechanics)

What sort of daily install numbers are you seeing?

I don't want to give out that kind of info just yet. We may do a debrief post in the coming weeks once we get a better idea of everything.

Having said that, as noted at the top of the post, at time of writing about 600,000 games of Hexiled had been played in the 3 days since the feature started, which has totally blown us away.

What tools did you use to write this game ?

It's all written using SpriteKit.

We didn't like the standard Game Center pop-in notifications, so we designed our own - these use normal UIViews, and we use the built-in Twitter/Facebook sharing view controllers.

The loading spinners, alerts / dialogs are all in SpriteKit. I'm not sure if this is optimal (I believe other developers lean on UIKit a bit more), but it all fits together nicely this way.

So it does not in fact run on android ?

One of the youtube movies on your linked page is titled "Hexiled Gameplay iOS & Android iPhone & iPad HD" and uploaded by "AndroidGameTech".

Correct, iOS only. I have no idea what their motivation was for posting the video. I simply used that as an example since it taught me a few things we hadn't encountered yet in our 2+ months of testing.

I am wondering what most of those game developers use for writing apps. Your app looks pretty minimalistic, and I do think it could use at least some additional decoration.

But writing my own game apps, I do realize why you're not doing that when implementing games directly in things like SpriteKit.

Are they using flash ? Unity ?

I am a web developer/designer and certainly not a game designer or illustrator so I guess in tackling a word game, we avoided 3D models or other things we couldn't do ourselves. We abandoned our initial game concept because it was going to require more challenging artwork.

The music was something we used a freelancer for, however.

Hopefully if Hexiled continues to prove successful, we're in a position to get help on the visual side!

I think Unity is pretty common. There's tools for converting cocos2d code to Android, so that one too.

Being the first game I've ever made I wanted something I could jump into really quickly just to play around. Got a bit of momentum on it and kept going.

I do wonder about the future of SpriteKit because of this, since having to code it two or more times is a huge effort. Apple seem to be all-in on it though, especially with SceneKit coming soon.

There are a ton of cross-platform frameworks, both proprietary and OS. Here are a few of the bigger ones: Marmalade, Cocos2d-x, LibGDX, Xamarin, and Unity (you can 2d with Unity now, I believe).

In comparison to Unity, Monogame is a good option if you would really like to write a complete game in a high-level cross-platform language (C#/F#) rather than scripting pre-made chunks: http://monogame.net

I "like" that they used the music with AudioJungle watermark

Reply on news.ycombinator.com