So Android has this “accessibility” feature that lets you hang up on a call by tapping the power button.

Except what happens if you’re on a call, the screen turns off (say, you’re using the speakerphone, it’s on a table). How do you turn the screen back on?

You’re right! The answer is: you can’t! Instead it helpfully always hangs up on your caller.

Should only hang up on people with that button if the screen is already on. Not if it’s off.

Of course, this isn’t if during normal use the phone doesn’t decide to turn the screen back on when it’s held against your face, start pressing the hangup button then get the touchscreen “jammed” because of persperation.

My normal phone arrangement consists of an iPhone 4 and a Nokia N900. iPhone gets my main SIM card for calls and the majority of my web browsing, N900 gets a mobile broadband SIM card for SSH, and other mobile Linux usage that doesn’t require a full laptop. Unfortunately the N900 is starting to show age, and I would use it as a primary phone, if not for the lack of UTMS-850 support that I require, and living in a blackspot for all carriers but Telstra.

Recently, a friend challenged me to put my Android trash-talking where my mouth was, and loaned me a Nexus S for a week, Google’s previous flagship Gingerbread phone. It’s okay — he has to use a HTC Mozart (a Windows Phone 7 device). In both cases we’re using devices that are released in late 2010, so they’re now showing their age.

I should point out from the start that he loaded the latest Gingerbread CyanogenMod ROM onto his phone, having failed to get an Android 4 CyanogenMod ROM to work (it popped up lots of crash messages). He got the Mozart with some custom ROM made by some Russian that I found on XDA developers, that gave it Windows Phone 7.5 (instead of 7.0) as well as all the Nokia applications, effectively turning it into a poor man’s Nokia Lumia phone. In both cases the phone is running a custom ROM with various enhancements not available in the stock ROM, and are rooted and unlocked — it’s a fair fight.

My friend attempted to load an Android 4 ROM on the device which didn’t work properly at all, with all the applications crashing. After loading the previous stable Gingerbread ROM instead and attempting four times to get my microSIM aligned in the slot, I was ready to go.

Immediately I’m told “download ConnectBot and Hacker’s Keyboard”. Okay. Done that. Installed it. The nice part of this is I don’t think it required any special sideloading trickery to get these things on there, and now I have a Dvorak soft keyboard.

I also notice there’s a big difference between this and my co-worker’s Samsung Galaxy S, which has an entirely broken build of some custom Android ROM on it. Talking my friends who use Android, CyanogenMod seems the way to go, and every one of them has loaded this onto their device. Manufacturer’s custom ROMs don’t last long.

Playing with it, there’s a nice complex “permissions” system for Market applications. I like this. But the problem with this system is it’s “take it or leave it”, there’s no way to revoke individual permissions to applications. I’m told later that there is infact a way to do this with CyanogenMod’s hacks, however it means temporarily granting applications the full requested access, and there’s no guarantee that the application won’t crash.

If I was to “fix” this, I’d force every Market application to either have an option to degrade gracefully in the absence of functionality, or give it access to a subset of or false information. For example, an application that wants access to your contact list may be granted access to an empty contact list just for that application, or only certain groups of your contacts. It could also push writes to the contact list to another list entirely, or allow you to prompt on those. In the end the applications that leak your entire contact list to third parties would be exposed, shamed, and removed.

The default homescreen, while it looks cute, I’m told uses a lot of battery life. It also lags the interface a lot. In fact, even with the fancy homescreen, the phone is still very unresponsive, and UI transitions are choppy. This is the price for running your UI in a virtual machine.

The calendar widget doesn’t appear on the home screen by default — almost like the calendar is an afterthought. Several UI problems present themselves, all of them are fixed in Android 4:

• The homescreen applet only shows your next appointment, but still lets you occupy large amounts of screen space to hold the applet.
• When you want to edit an appointment, the only way to do this is by a context menu. There’s no edit button that’s just shown, despite loads of space to put one.
• When you actually edit that appointment, the button order is: Done, Cancel, Delete. It is very easy to confuse “Done” and “Delete” (who reads labels), and there should be a very obvious cue that you’re about to do a potentially disasterous operation. As an example, on the iPhone, delete buttons are marked in red.

I’m having a lot of reception problems with the phone at home on data, it seems to be using Telstra’s GSM-900 GPRS network (despite being connected to WiFi). The 3TELSTRA UTMS-2100 network has patchy coverage at best at my house, and the lack of UTMS-850 on this version of the phone is really starting to show a problem. I imagine this is only going to get worse once I try to teather with it. There is another version of this phone that supports UTMS-850, however this is at the exclusion of GSM-900, meaning it won’t work on other Australian mobile networks.

But in the end, these rules don’t matter at all — because of the way that auDA policy works it is trivial to squat domains even when it violates auDA policy.

How it works is very simple — it is to do with the way the auDRP (Dispute Resolution Policy) is enacted.

In the policy, it requires the services of an independent arbitration panel. The complainant must pay a fee of 2000$ for a 1-member panel, or 4500$ for a 3-member panel. The complainant may ask for the domain license to be revoked, or transferred to them in the case that they are eligible. The worst that can happen for a defendant is that the domain is taken away from them.

As a result, a nefarious party may register large numbers of domain names hosting adverts on them (so as to satisfy that the use of the domain is “relevant”). They could use common types of businesses, or locations of businesses, things that people are likely to type into their web browser and add .com.au on the end of.

The extremely cunning will go a step further, scraping business names from sources like the Yellow and White Pages directories, or the Australian Business Register, and directly target businesses. There’s already people out there doing this with Adwords listings trying to target traffic from competitors.

They could then offer listings on the site or to buy the domain outright for several times the registration cost (say, 1000$ to have the domain transferred to the business), and it would be still cheaper than dealing with the auDRP, even though it is in clear violation of policies.

Basically, the auDRP is a huge joke when it comes to domain squatting of this nature. The only people who can afford to deal with it are big business (so registrations of micosoft.com.au or gogle.com.au will never be a problem), but small to medium businesses, often those that have a poor or non-existent internet presence get the short straw.

What’s more is some of these tactics are actively being exploited today. There exists a company who operates over 2,500 “business directory” sites using generic names, that are “Copyright 2011 example.com.au” who sell listings on these sites at 75$/yr — more than what it costs to register a .com.au domain for two years! This is being operated by Ollority Pty Ltd, who in turn seem to be owned by Online Marketing Group Pty Ltd, who are in turn owned by Fairfax (one of Australia’s largest media companies).

This kind of wholesale pollution of our country’s ccTLD needs to be stopped. Unfortunately there’s a lot of money in this industry, and I don’t think that those companies are going to go away quietly.

Google managed to establish themselves as a household name of the first place you go on the internet when looking for anything. They beat other search engines and directories by offering very good quality results in a lightweight, no-nonsense format that was accessible. They beat business directories like the Yellow Pages, who deliver printed directories every year to every person with a telephone line, who have been a household name for decades.

I have nothing against people starting up a business directory. You just shouldn’t resort to pollution of our country’s domain namespace in the process.

Basically the idea behind a 28-hour day is that instead of following a 24 hour circadian rhythm (16 hours awake, 8 hours sleep), you follow one that is 28 hours long (20 hours long, 8 hours asleep). This is inspired by the xkcd comic (which is always a bad idea to take things on it seriously)…

I made some adjustments to the plan, namely that my day is divided up into 18:40 of waking time, and 9:20 of sleep. This means that I ended up with the same amount of sleep as you would on a typical 16/8 24-hour cycle. I also shifted the days around slightly so that I would wake on Tuesdays at 00:20 (you can calculate the rest from there). The advantage is that it gave me some time on Monday mornings during business hours, at the cost of some loss of time on Friday afternoons in business hours.

The bonus for me, theoretically, is that I am always exhausted when going to bed. Which is great when you have sleeping problems.

The biggest problem you’d encounter with such a schedule is the rest of the world. I get around this by having a bit of flexibility in the hours of my employment (but not everyone has this). If you’ve got flextime, you can basically go nuts. Then if you’re living with anyone else, their schedules are a problem.

You also lose a day. Some people go crazy when they’ve realised a day has “gone missing”.

My eating habits were also a problem. I noticed I was a lot hungrier on this cycle. Typically I can get away with three meals a day with ease (sometimes two, by having a bigger dinner), but on this, I found I needed to eat four to five meals a day. This is because of the bigger energy demands of staying awake for longer.

My bowels caused trouble as well — they went absolutely haywire. I think if I stuck with the schedule for longer and setup artificial lighting in my room, I could trick them into working properly again. This could be diet related as well, YMMV.

I wrote a clock to help me monitor how my sleeping is supposed to be, and calculate it into something “normalised”. I also published the sources to this on github. I’ve got an updated version with a HTML5 canvas diagram in the works, I just haven’t finished it properly.

There’s a number of improvements in 2.8.4′s development branch already over 2.8.3:

• Local sign-in and events management system.
• Port forwarding system.
• Migrating several systems over so they work with Django 1.2, and removing some older legacy code.
• All the local authentication hacks have been eliminated.
• Fixed XSS issue in captive landing page.

The code is released under the GNU Affero General Public License v3, and is available from my GitHub repository.

At some point in the future this’ll be pulled into a 2.8.4 release, after some further janitorial work is done. It’s been sitting around for a little while, may as well get it out there now. Eventually this will be migrated to become a part of the LanConnect project, with optional integration with it’s databases.

There is a VGA adapter available for them, however I notice that the buttons on the controller and menus do not operate properly on Linux — after the first button press, the menu stops responding. This makes it impossible to select 3D mode or adjust any settings.

The controller works fine on Mac OSX and Windows in my tests, which left me scratching my head. It also works fine when you plug the VGA of a Linux box into it, and USB into another computer. I’ve also tried it with several other configurations, where I had a powered hub not connected to upstream USB, and plugging it into an iPhone USB wall charger, and still wouldn’t work.

Turns out, the controller presents in addition to a USB interface (1BAE:014B) for an internal sound card, a HID interface for the (optional) head tracking module. Linux doesn’t know how to handle it, so it presents it as a “hidraw” interface (typically /dev/hidraw0). You just need to read from the device, and suddenly it starts working! You can’t just ignore the device entirely (with a usbhid quirk or unloading the module), you must read from it otherwise events are blocked. This is probably due to a bug in the VGA adapter firmware, but was never found because Linux isn’t a supported platform.

This is simple to deal with. As root, run: dd if=/dev/hidraw0 of=/dev/null.

All the events will be read from the device, and simply discarded. If you watch it with hexdump, there’s some events that come up whenever you press a button on the controller. In the end, they don’t matter. I haven’t looked much in to it yet, but there’s probably a way that you could parse them, and possibly even reconfigure the device from software. This is needed regardless of whether you have the optional head-tracking sensors installed — I don’t have them.

I think that Mac OSX and Windows both do this in-kernel when it doesn’t know how to handle events, it’s probably just a quirk in how Linux handles these devices.

One important parameter if you’re adding more sources is border-alpha. This needs to be set to 0, so that the black border that is added around a video source is transparent so you can push it on top of another video frame.

This is handy if you have two webcams, you can use this to setup side-by-side stereo display to 3D glasses. Just make sure to point your cameras at the right angle, otherwise your convergence will be all messed up.

You need the gstreamer-tools package in order to use the script.

#!/bin/sh
gst-launch \
	v4l2src device=/dev/video1 ! "image/jpeg,width=320,height=240" ! jpegdec ! "video/x-raw-yuv" ! videobox right=-320 border-alpha=0 ! videomixer name=mix ! \
	ffmpegcolorspace ! xvimagesink \
	v4l2src device=/dev/video2 !  "image/jpeg,width=320,height=240" ! jpegdec ! "video/x-raw-yuv" ! videobox left=-320 ! mix.

I’ve written a QR Encoder patch for Quartz Composer based on libqrencode. You give it a string, and then it’ll encode it into an image.

There’s another patch in my github which provides a client to GPSdX. It’s a bit buggy still, and only works with GPSd running on localhost with the older encoding method. Unfortunately libgps is ridiculously annoying to work with, so I took the easy (incompatible) way out.

I’ve also patches for Openbravo POS. They make the “customer” price for an item be the buy price, allowing you to implement a simple discount for registered customers that isn’t based on a percentage. It also includes Australian currency images and some minor UI modification to waste less screen space.

I’m currently working on cleaning portal2 up for release. Which is fun. Most of the legacy authentication hooks have been removed, and it’s getting closer to running standalone in a nice way. As part of this I’m getting rid of lots of ugly and legacy code that was for Django 0.96.

Unfortunately the problems are quite deep seeded. I’m still actively participating in the LAN scene (I recently travelled to Sydney to help out with LANTACO, a LAN run by the Macquarie University Computer Society and that was a lot of fun), however I’m not involved the ongoing operation of any public events.

On the roadmap is the completion of the Jabberwocky project, which will provide a complete integrated stack of services for LAN parties. I’ll also be looking at releasing the software I’ve developed for StreetGeek as open source in the coming months.

Update 2010-11-08: I’ve read a lot of stuff about “bad blood”, and a lot of people are frustrated that I’m not involved with this event anymore.

I would ask that you do not engage in attacks against StreetGeek or it’s admins in my name. It doesn’t do anyone any favours, and just rubs further salt in the wounds.

If you don’t like where StreetGeek is going, either work towards fixing it or don’t attend anymore. Just don’t do anything stupid.

Thanks.

Just thought I should post this, because I do not wish to be attributed as having written such horrible software, nor impose it on anyone else.