Android Bitmap Idiosyncrasies

Whilst improving my Android app Chromakey Photo Edit, I've come across a few idiosyncrasies with Android Bitmaps which I thought I'd share for reference:-

• Bitmap.createScaledBitmap() always returns a bitmap of type RGB_565, regardless of the source bitmap.  See http://code.google.com/p/android/issues/detail?id=13038
• ARGB_4444 is now depreciated, so you'll always need to use ARGB_8888 if you want to store transparency information.
• When using BitmapFactory.decodeFile(), always try to pass a BitmapFactory.Options().inPreferredConfig to ensure you get a bitmap in the format you need.  By default, bitmaps loaded this way are always immutable (editable), so you'll need to do a Bitmap.copy() if you need to make it mutable.
• Bitmap.recycle() is not neccesarily required, but can help improve the garbage collection.

Fixing GRUB

Grub seems to be a constant source of problems.  After updating my Ubuntu I got an error along the lines of "grub may need sorting out for various reasons".  I hate it when it does that, because if it does go wrong, you're usually stuck without a computer and are thus unable to browse the internet to find out how to solve the problem, a'la Catch-22.  Needless to say, when I rebooted, I just got a completely blank screen, not even a cursor flashing in the corner.


Anyway, for my own reference, here's what I did (after booting from a LiveCD - what did we do before LiveCD's?)

Mount the disk:

sudo mount /dev/sdX /mnt

Re-install Grub:

sudo grub-install --root-directory=/mnt /dev/sdX

Obviously, change the sdX bit to point to your HD.

If you're a real masochist, you could try reading this.

How to be a productive programmer

Here's my humble checklist of what I think are the most important requirements for being the kind of programmer that can create something ultra-quickly.

* A stock of library code, built up over years so that simple and common tasks are already covered by a pre-written function or two.
* Be either the only programmer on the project or have your area clearly defined, and not be dependant on someone elses code that is still being produced.
* Have it clear in your head exactly what you are creating and how it is generally going to work.
* You have already "solved" any major technical hurdles that you know you are going to encounter.
* A quiet-ish place to work.
* A desire to actually create this program.
* Access to the internet to solve those niggly problems or to copy-and-paste code from websites.
* As a bonus, have a similar project that you previously created that can be easily subverted to the new task.

If you've got the above covered then you should be able to churn stuff out quickly, and then you can spend the rest of your time writing your own more interesting projects.

Megaproject #2: Worldcrafter

I call it a "megaproject" since I'm pretty convinced I can see this one through to the end without getting bored or sidetracked onto something else.

It's called Worldcrafter. In a nutshell, it's a 2D Minecraft-style game for Android. Yes! Another one! This is what it looks like so far, after about 5 days since I started (most of those days being spent trying to decide what to call it):-



Does the world need another Minecraft game? I'll let the world decide. What I like about writing this kind of game (and what I like about Stellar Forces) is that it's almost "modular". You (i.e. me) can play a bit, then think "ooh, wouldn't it be good if it did this". And then add it. However, the basic game is quite easy to get up and running, and that's the important bit when starting a new project: have something playable ASA-goddam-P.

Domain Squatters Can Die!

If you've ever tried finding a domain name for your new project, no doubt you'll have come across the problem of cybersquatters, the vermin that polute the internet with ad-farms by buying up any and all domains that might be remotely useful for anyone. The rest of us have to try and think up obscure domain names that still try and relate to whatever subject our website is about, and hope people will still be able to find our site on the net.

Of course, you can usually buy a domain back from them if you want, using one of the their very helpful "make us an offer" forms. Gee, thanks. You're doing mankind a service.

Software Patents Metaphor

I'm a chef, and I've just patented a new recipe: chips with *2* sprinkles of salt over them. So if I see anyone sprinkling between 1 and 3 sprinkles of salt over their chips, expect a letter from my lawyer.

-- Update --

I've just had an amazing idea: chips with 3! Yes 3! sprinkles of salt over them! I don't know how I do it. I might start to sell this new recipe over the internet, but of course it will take you more than 1-click to buy it.

Tower Invaders: Hints'n'Tips

It was my deliberate intention to make Tower Invaders quite difficult from the start. I didn't want players to have to wade through loads of screens of baddies before they got to the challenging bit.

Anyway, here are some quick hints'n' tips for those who might be having some problems defeating the invaders. I like to think of the game as more of a puzzle or even strategy game. There is a lot of strategy involved in what at first seems to most simple of choices:-


* Position of towers is important!
Even though the invaders move from left to right (and back, and so on) covering the whole width, the position of the towers is still very important. Statistically, the invaders will be mostly above the middle of the screen, so that is the best area to place your pulse lasers. However, this also means that this is where the most bullets are going to fall; if you have the full version, shields are a good idea, especially on towers you've spent a lot of creds on upgrading.


* Sell towers that are about to be destroyed!
You will get the same amount of money back for a brand new tower as for one that's about to crumble. You do get more for upgraded towers, though obviously not as much as you paid for them. Wait until a tower is about to go belly-up before selling it.


* Keep Zappers to the side!
Since Zappers will always hit an invader, keep them out of the way of bullets. As mentioned earlier, the most bullets will fall in the middle of the screen, so you might as well build Zappers to the far sides, where they will have much less chance of being hit.


* Upgrade towers to limit your exposure!
It costs the same to buy two towers as it does to buy one and upgrade it. Even though an upgraded tower will shoot about 10% less than two basic towers, those two basic towers cover twice as much area, so they have twice as much chance of being hit by a bullet. Also, you'll need 2 shields to protect them. Of course, they will be able to incur twice as much damage, so things aren't always so clear cut.