In retrospect, what I really enjoyed about my early days of using Twitter was that I could construct a narrative about my life. Everyone naturally creates such a narrative, but writing it out gives you the opportunity to shape it. You can emphasise and omit, and shape your perception of your own life.

So for the next month, I want to run an experiment: I will tweet to construct a positive narrative of my life.

This is not intended as some kind of bullshit oprahesque magical thinking exercise: I do not believe that positive thoughts are magic. But I do believe that your level of happiness depends on your perceptions. There are plenty of people who are unhappy because they are relentlessly negative: Their narrative is an endless string of disappointments and aggravations picked out from their experiences, instantly forgetting all joy. I want to make myself do the opposite.

This is not about lying. If I get hit by a car, I won't tweet "Hit by a car! Best feeling ever! #spinalinjuries". I want to highlight things that I enjoy and things that make me feel good about myself:

"Having delicious coffee. Now awake!"

"Just managed to shave off 25% running time from text recognition."

"Off to the theatre tonight to see X."

It's astounding just how hard it is for me to even contemplate publicly going on about my achievements and qualities. As a shy and nerdy kid I learned to be unobtrusive to avoid bullies, and most of a decade spent in the UK, the land of understatement, hasn't exactly changed this attitude. So I feel that I have to write this preamble and frame this (probably actually quite healthy and normal) thing as a time-limited experiment. I would put a giant blinking disclaimer in front of every single tweet if I had the space.

(And no, the feed still isn't public.)

Here's a video of me getting through level 3 of Dawn: 8 minutes of being chased by a vampire - it is possible!

(This one is recorded against a desktop version with audio I'm working on. I quite like the breathing effects, but dislike what I have for the other sound so far, which is why I went for the minimalist approach here.)

The finished version of Dawn is now available. Compressed, it weighs in at just under 4 kB, so it's ready for submission.

A new beta version of Dawn is available. This version has some bugfixes, some more items of furniture, and somewhat easier gameplay.

No, not the Federal Republic of Germany. Nazi Germany. We already won, you say? Well, they do still seem to be occupying pockets of France.

The Geneva Conventions make a clear distinctions between combatants - people who identify themselves with some form of uniform - and non-combatants, who are not in uniform and hence not part of the fight.

So if you find yourself in a restaurant with some Nazi soldiers, all you have to do is identify yourself as part of a military force, and the war can be on. Ideally, carry a cap or helmet with you, though traditionally speaking, a brightly-coloured handkerchief tied around your shoulder together with a verbal declaration of your combatant status should do as well.

While it is a bit unclear whether you are allowed to shoot the enemy soldiers at this point, as they are likely unarmed, you are definitely allowed to take them prisoners of war. You can then keep them in a prison camp (perhaps a large pit in your back yard?) until you receive confirmation of a formal surrender from a high Nazi official: say Hitler, or maybe Goebbels or Goering. Until the time where you have Hitler on your phone surrendering, you may keep them in that pit.

I envision a glorious if rather lengthy war.

The beta version of Dawn is now available to try out at this finely-crafted link.

If you have any feedback/suggestions, please email me.

This version features a much bigger house. Since a bigger house translates into more data used, I ended up removing the crucifix. It didn't work very well, and I needed the space to have some randomisation of where in the house you starts. (Otherwise, as Dave rightly pointed out, you quickly develop a particular pattern of movements that lets you win, which is boring. The game also has levels now, doubling the time you need to stay alive with each level, and decreasing your chances of finding things.

According to my compression chain, I have either 54 or 4 bytes spare at this point, depending on whether using JoGa breaks the code or not. I can't tell because you need special weird sauce to actually host hyper-compressed java applets. If I figure out the sauce, and it turns out I have 54 left, I might add a bed as an item of furniture.

As always, you can see the code for it on Github.

Your possessive, manipulative, jealous boyfriend has snapped and is coming after you. Unfortunately, he's a vampire.

You must survive until dawn.

I've finished with the alpha version of Dawn, which weighs in at 4025 bytes after a lot of compression. You'll have to wait a little bit for the playable version, because the controls are currently kind of awkward, and because I want to re-do the map to be more interesting.

I've managed to get a reasonable compression script working, so now I have a better idea of how much space I have left. Right now, the compressed game weighs in at 2609 bytes, so I have another 1487 bytes to play with. Plenty of space.

You can play the newest version here

Here's a new version of Dawn, this time with some actual gameplay. You're being chased: try not to get caught. Try not to get too exhausted.

You can now try out Dawn in your browser. Needless to say, the game's not done yet, but I'm making good progress. My main worry is that there is no way that I'm going to be able to stuff in all the things I've planned: I'm currently at 9.5 kB uncompressed jar file, and while that probably squeezes down to about 2.5 kB compressed, I'm really not very far yet! Unfortunately, Riven's Compile n'Shrink, which is what I used to compress the jar last year, is currently down. And my own attempts were never nearly as good. (Most of the instructions for jar hypercompression are vague and Windows-specific.)

I'll just have to see.

For now, I'll direct you to my absolute favourite Java4k game of recent years: Burning Man. You seem to be on fire; you probably should fix that.

I entered last year with Moo4k (warning: Java applet), which didn't do as well as I hoped - mostly because it was graphically very drab and the reviewers saw it as a more plodding version of Galcon instead of a demake of Master of Orion. That said, it was a lot of fun to make.

This time I'm going for a realtime game with the working title of "Dawn". I thought it'd be fun to document the development process here, so that's what I'll do along with some terse notes - I mostly want to make the game, not just blog about it!

You can also follow the game's progress on GitHub.

We've all seen this: an indie game is grandly announced, makes a lot of progress initially, and gets to a playable stage. You play it, you like it, you eagerly anticipate updates. But then the time between updates starts getting longer, and the updates get more terse. Eventually, there are long stretches of no activity at all. Then, there is a post insisting that the project "isn't dead". That's usually the last update you ever get.

I really don't want SE:SS to be another instance of this. So why, exactly, have you seen no updates for the past year and a half?

Two years ago, in October 2009, I quit my day job at Linguamatics to finish SE:SS. I had savings to last me to the end of the project, and some contract work lined up. Two days later, I had RSI so bad I basically couldn't type anymore. I proceeded to limp through to finishing dev 10, months behind schedule, working as fast as I could, maybe an hour a day. By that point my money was running out, so I switched to spending that hour doing contract work.

Since then, I have been slowly and unsteadily improving. Today I can work for maybe four hours a day on average, if I space them out through the day. Of all the things that I tried to improve the RSI (or whatever it is I have, exactly), the one thing that worked the best is exercise, so I try to get in enough of that as well.

So that's one reason for the lack of updates. The other has more to do with my own foolishness. You see, I have been working on other projects. Things like my Java4k entry, Which Fish?, Kobold Pit, Project Fruitbat, and, um, an entire OCR engine. So why didn't I spend this time on SE:SS?

When I first started putting up a feedback form for the game, I thought this would be a great way of receiving bug reports and knowing what people thought of the game. What I didn't realise was that I'm frankly pretty thin-skinned, and really could not cope with some of the feedback I got. Now, for the Internet, even the most negative feedback was rather mild. It wasn't "I will track you down and gut you like a fish," more "This game sucks. You should change it around completely. Make it 3D. And massive multiplayer. And about ponies." The vast majority of the feedback was much more polite than that, and a lot of it was very constructive.

So the stupid thing I did was that I got really defensive over the exact way I'd written the game - even bad design decisions I'd made on the spur of the moment at half past midnight. People told me again and again that the game's responses were too random. I hid behind a succession of justifications for why it had to be like that instead of realising that for a game to actually work, there has to be a clearly visible link between the player's action and the game's reaction, and there simply wasn't in many instances. Send out a well-chosen crew to explore ruins? Bang! One of them dies. Send out a single trembling archeologist to hunt prayer monkeys? She captures them with ease. This is frustrating, but I'd decided that "in real life, this is just how it works", evidently forgetting that I was making a game.

I felt that I didn't want to do a game based on the ideas of whoever sent me the most strident emails (even though I'd asked for them, even though a lot of the feedback was very constructive). I had also made he mistake of rather heavily promoting the game as nearly done, with a clear release schedule it was now impossible to meet. So I ended up pretty much not working on SE:SS at all. And when I worked on other things instead, I tended to talk about them almost furtively, since they were using up time I "ought" to be spending on the game.

The game on my hard drive is somewhere between dev 10 and dev 11 right now: dev 10 with some combat mode enhancements (like a tractor beam) and one or two more encounters. There is obviously a temptation to just ditch the project and do something completely different. Concentrate on my other projects. Start making a new game based on one of the myriad ideas I've had since I started working on SE:SS, um, four years ago.

But actually, I don't want to. Not out of stubbornness or out of a sense of owing the completion of the game to others, but because I actually really like the game. I take dev 10.5 (so to speak) out of its dusty toolshed every once in a while and just play it. And I really like playing it. I have fun. I like the graphics, the sound, the setting, the characters.

After a year of perspective, I've realised that I should make the best game I can, not the game I'm least worried about people disliking. I'm the developer, so if I don't think ponies are a good theme for the game, it won't be about ponies. By doing that, I can also let myself see where there are real problems, like the aforementioned frustrating randomness.

This has been a pretty long and maybe slightly over-sharing post. The short version is that I haven't worked on SE:SS because of RSI, and because of not coping with others' expectations.

So when will you see the next dev release? Not quite yet, I'm afraid. I have been tying up loose ends (like shutting down my unnecessary company), but I'm not there yet, and dev 10.5 needs more work before it's worth putting out as an update. I also absolutely do not want to make promises about a schedule. And I'll be working on other things too, and will actually write about them, and link to them properly.

I want to remain kind of stealthy for now, and only give a firm date on release when I have on my computer a beta version of the game, ready to be released save some typos and minor bug-fixes.

It will be done when it's done. And it will be.

whichfish.org is a website I've just launched that helps you easily tell which fish are sustainable to eat. Why? Well...

Source: Information Is Beautiful / The Guardian

This picture terrifies me. In 1900, we can see huge tracts of ocean where there are massive numbers of fish. In 2000, the map is empty apart from a small leftover smudge clinging to the American coast.

There are no longer plenty more fish in the sea.

Where there once were swarms of fish so abundant we could catch them with simple nets, we are using ever more sophisticated equipment to locate and catch what little is left of them. Once innumerable fish like cod have been reduced to a few isolated stocks and majestic creatures like bluefin tuna are nearly extinct.

At the same time, a lot of the fish we pull out of the sea is thrown straight back in or ground into fishmeal, simply because there is no demand for them. They don't taste bad, they're just not the fish we're used to eating.

The thing to do, of course, is to eat more sustainable kinds of fish. We know this. But that knowledge alone doesn't help much when you're standing in the fish aisle in the supermarket and need to buy dinner now.

There are some very good sources on what fish to eat, like the Monterey Bay Aquarium and the Marine Conservation Society. Both have done a huge amount of work collecting and assessing data on sustainability.

I have combined their data, and data from other sources, and turned it into an intentionally simple web page that just tells you which fish are OK to eat. You can view it on your mobile and download it for offline viewing as well. It can help you make a quick decision on what's for dinner.

Give it a spin. It should display well on most mobile phones, and if you have an iPhone or similar, you can also download the site for offline use, so you don't need a signal to be able to use it.