Review: PlayStation Move
It's surprising to think that the Wii has been on sale for nearly four years, and we're only now seeing true competition from Nintendo's rivals. Of course, considering the half-hearted motion controls of the Sixaxis, it's probably wise for Sony to have been cooking PlayStation Move slowly since its introduction 15 months ago.
The PlayStation's motion controller sticks much closer to the Wii's mould than Microsoft's Kinect, with E3's Move tagline being "This changes everything". Hyperbole aside, Sony is aiming to take the technological high road, as it does with all things PS3, and beat the Wii on precision and movement accuracy.
The technology used is actually pretty similar to Nintendo's, but on steroids. While the Wii's sensor bar sits by the TV and gives out infrared lights that a camera in the remote pick up for use of the pointer.
Move flips this around, with a camera next to the TV picking up the glowing ball on the end of the controller. The PS3 can then measure distance from the controller using the size of the ball as a reference.
The Move controller also has all the movement sensing bells and whistles of the Wii remote and MotionPlus attachment, demonstrating that Sony's thinking was obviously that they weren't attempting to reinvent the wheel so much as make it a little rounder.
Move controllers are available alone for £35, or with the PlayStation Eye camera for £50. The Navigation controller (the equivalent to the Wii's nunchuck) is £25, though we haven't included it in our testing it here.
The most noticeable thing about Move's primary controller is the glowing ball on top, unsurprisingly. When the controller is off, the orb is white and softly translucent, and looks remarkably like a ping-pong ball (it's about the same size as one, too).
When the controller's in full use, the ball lights up in a range of colours, so it can be tracked by the PlayStation Eye camera. While the ball might seem to be a weak point should the controller ever make contact with your wall/lamp/friend's skull, it's actually squishy, and just pops back into shape after an impact.
The rest of the controller has a more organic look than the Wii remote. It's round, and becomes slightly thinner in the middle, presumably to be more ergonomic.
On the front, you have the four familiar PlayStation face buttons, though the fact that they're arranged in a square, rather than a diamond, makes it a little hard to remember which one is where for a while.
Between those is the Move button, a thumb-sized new addition, clearly meant as Move's version of the Wii remote's big A button.
Beneath those is the PS button, which serves the same function is it does on the DualShock 3 or SixAxis controller of bringing up the XMB. It sits in a concave, which avoids accidental presses neatly.
The underside of the controller is mostly clear, but features a trigger, known as the T button. This is analogue – like R2/L2 on the DualShock 3 or the left and right triggers on the Xbox 360 controller – in contrast to the crisp, clicking B button on the Wii.
At the base of the move is a micro-USB port for charging, a slot for the provided wrist straps, and two mystery connectors that could be used for accessories in the future.
At first, the Move controller feels a little more comfortable than the Wii remote. It's not that Nintendo's controller was uncomfortable, but the roundness of Move does some way toward helping it sit neatly in the hand.
That said, we found that we started to feel the effects of prolonged use faster with Move, and we think it's to do with the shape. The tapered middle means you're often gripping harder than you would have to with the Wii's controller, especially when playing something with hard swings, like Table Tennis on Sports Champions. Discomfort in the wrist crept in earlier than it did during an equivalent session of Wii Sports Resort.
Let's be clear, though: We're not saying Move is painful, uncomfortable, bad for you, or anything like that (assuming you have no joint problems to begin with). After all, we were playing for quite long periods during out review time, though not unusually long for a committed gamer.
We're not even saying it's definitely less comfortable than the Wii remote – as we said, it's actually a bit nicer just to hold – but we do think that the shape isn't ideal for long periods of the more wrist-bending games.
There are a few other things about the Move's design that seem a little odd to us. Why add a new button with the Move logo (which is, let's remember, just a squiggly line, and so doesn't jump out at you on-screen), when Sony could have just used X or Circle?
Of course, the most contentious design decision will always be that orb. We don't deny that accuracy that it brings (more on that later), but it really does look silly. We're not going to make the laboured sex-toy joke, because all of your friends who see it will. Seriously, it's not just a meme – it's the first thing that people who've never even heard of Move say.
The light is also quite distracting. If you're trying to do something else in the room while someone's playing a game with two controllers, your eye is constantly drawn. The Wii remote was designed to be inconspicuous – the shape fits in with your TV remotes, and it almost disappears into your hand when you hold it.
That said, motion-controlled gaming always has and always will make you look a bit weird. It's not like the wild flailing was dignified before the glowing ball was added, so maybe we should just be happy with the extra accuracy and possibilities it offers.
In the games Sony provided us to test with, we were able to get a feel for much of what's possible with Move, but we also became aware that these tests are somewhat limited by the way the games are programmed.
This has always been one of the Wii's biggest caveats, and the worry is always that games will use motion sensing in a way that really should have just been achieved with buttons.
With that in mind, we can safely say that we came away hugely impressed with what Move can do, but it still needs a careful, measured implementation.
The controller's ball certainly seems to offer an accuracy of pointer movement that goes beyond what the Wii remote is capable of. The idea that it could be as accurate as a mouse is probably a bit ambitious, but it's good enough that we could see it giving real-time strategy games, and a few other genres that do better on PCs, a new lease of life on consoles.
When you navigate the XMB with the Move controller, you point it at the screen, hold trigger, and then point it up, down, left or right to move in that direction in the menu. Interestingly, the ball stays dark during this time, so it seems to be using only the motion sensing capabilities, but it's so fluid and accurate that you wouldn't know the difference.
When the ball is being used as a pointer, we found that it often only operated in a very narrow field compared to the Wii, which isn't necessarily a bad thing, but it makes navigating some menus a little more fiddly than we're used to. On the other hand, it works perfectly at much further distances than the Wii remote is capable of.
There are several games where the orb needs to be visible to the camera, even when you'd think the control would be all motion-sensing based. With the Wii, you could walk into another room and bowl if you wanted, but that's rarely the case here. In fact, leaving the camera's field of view with the active controller seems to stop all functions from working – including the buttons.
While one of the criticisms of Kinect that going around is the space needed to use it, this is actually also an occasional problem for Move. //Sports Champions// demands that you stand eight feet (2.5 meters) away from your TV, and then it still expects you to be able to step backwards, and swing your arms all around.
For some people, this won't be a problem, but in our case it meant shifting a sofa back a couple of feet every time we wanted to play (not to mention getting rid of the coffee table). Yes, we always needed to make some space when playing on the Wii, but nowhere near as much as this.
Of course, we're just talking about living rooms here. These space restrictions will just about rule out Sports Champions for bedrooms.
Why is this an issue for Move when it isn't for the Wii (some of the most energetic Wii games be played sitting on your sofa or standing up without too much of a penalty)? It's the restriction of the ball and camera system. You need to be able to swing your arm well out occasionally, and still be on camera.
This restriction follows through into some of the multiplayer games. We only tried with two people, which was fine once we'd made enough space, but if you were playing four-player Volleyball, where everyone has to be on camera at once, it would be absolute carnage.
Annoyingly, Sport Champions wasn't the only game that caused us a distance problem. Our sofa was too close to the TV for us to be able to play that game, but when we popped in Start the Party!, we were too far back!
Start the Party! uses the camera and Move controller for augmented reality (AR) WarioWare-esque party games, so each active player needs to be quite large in the screen. This distance inconsistency is another sacrifice for the accuracy on offer, but it's simply a problem that the Wii doesn't have.
The AR in Start the Party! is very impressive though (and it actually makes an occasional appearance in Sports Champions too). Giant foam hands stick like glue to the end of your controller, wobbling convincingly with the momentum of your swings. Tennis rackets twist in your hand, meaning that you have to careful to hit with the strings, and not the rim.
Win one round in particular and the controller becomes a pencil, enabling you to deface your opponent's image – all with incredible accuracy, though the fact that you're seeing yourself from the other way makes it a little confusing when you're rotating things, or moving them to and from the camera.
Naturally, there's a horror story too. Kung Fu Rider – a kind of Tony Hawk's meets Pain, with a Crazy Taxi heart – is a classic early-Wii case of unnecessary waggle. Thrust the controller up to jump, but to accelerate you have to shake it up and down, resulting in numerous accidental jumps. And yet, when you actually want to jump, it's frequently unresponsive.
It's a perfect example that the Move technology can only be as good as the software harnessing it.
With Sports Champions sitting happily alongside Wii Sports Resort on our shelf, the obvious test for Move was to put it up against Nintendo's offering. There are several like-for-like games here, so how do they compare?
Table Tennis
A tricky one. The Wii version offered a huge amount of control over the spin on your ball, but you still had control over your Mii's movement, and the accuracy of your swing didn't matter as long as the timing was right.
Move couldn't be more different. By stepping left, right, forward and backward, your character will do the same, enabling you to get in close for smashes, or get back for a powerful top spin return.
You also need to think about the height of the ball, because it's quite possible to just swing at air underneath it.
Serving on Sports Champions is a nightmare, though. Not a single one of the people we got to play it could get the hand of it.
Of course, adding elaborate physics to a sports game just means it can go wrong. Attempts to put slice on the ball can result in it pinging off at ridiculous angles for no discernible reason, probably in part due to the precision the game demands from you and Move. Precision that is there, but is hard to master.
Wii Sports Resort is the arcade version, wanting you to put crazy spin on the ball, but this Sports Champions is all simulation. If you want to put tonnes of side on, you'd better get some damn practice in.
Winner: Sports Champions
Disc Golf
There's no proper golf on Sports Champtions, which is a bit of a shame (though perhaps not unexpected, with Tiger Woods 11 already out), we do have a good ol' Frisbee to toss about.
There's barely anything between these two, in terms of the control system. Sports Champions seems to be a tad more forgiving in that it's slightly easier to through the disc straight in front of you, but both games have totally convincing curves and wind effects in flight.
Winner: Draw
Bowling/Bocce
Okay, so this isn't exactly like for like in terms of the games, but the control scheme is the same for the pair, so it's a good comparison.
In Sports Champions, how much momentum you get on the Bocce balls from a throw can occasionally be a bit inconsistent. Throw the pallino hard and low on the S-shaped course and occasionally only travels about 10 metres, while other times it rockets round the course, though you're sure you threw it pretty much the same.
It's also hard to really get the hang of left and right spin on the Move game, especially compared to Wii Sports Resort's bowling. We've always found that the Wii bowling game betrays exactly the same slight left spin that we have in real life, and that adding a different spin is a just a matter of subtle wrist action.
Winner: Wii Sports Resort
Archery
To keep this fair, we compared the Wii remote-and-Nunchuck Archery game to using two motion controllers at once on Sports Champions.
The Wii version was always one of the most impressive MotionPlus demos, with every twitch and sag of the your arm translated to the screen. At first, Move really disappointed us. Control was laggy and accuracy was very tough.
However, it was then pointed out to us that, while we had made some effort to adopt a correct archery pose, we weren't doing it properly. So we turned fully 90 degrees from the TV, outstretched our arm all the way and tried again.
Suddenly, movement was perfect. Going from target to target is smooth (though you get more of an aiming aid from Sports Champions than from Resort), and using the second controller to bring arrows into the bow yourself gives you a great Robin Hood feeling.
The only thing we missed from the Wii version is a way to readjust where the centre of your aiming is (for example, you can aim slightly below the TV as your centre, so you're arm doesn't get in the way). This would be even more welcome on Move, due to the distraction of the glowing orb.
Yes, it's less realistic, but real archers don't have lights on their bows. However, this doesn't take away from the accuracy of the controls.
Winner: Draw
Swordplay/Gladiator Duel
The addition of shields in Gladiator Duel makes Sports Champions offering a little more elaborate than Resort's, but it's still swords.
Alas, the Swordplay game on the Wii was always a bit of a disappointment, because the actual hits tended to be restricted to vertical vs horizontal swipes and blocks. Despite the appearance of attacks at different angles, it pretty much boils down to those gestures, wasting the accuracy of MotionPlus.
Gladiator Duel makes good on these promises, especially with two controllers (for the sword and shield respectively). Attacks do more damage if you hit harder, but there's still the classic situation where a casual swing suddenly deals a huge amount of damage and you're not sure why.
To be honest, the swing strength detection is kind of inconsequential because everyone always swings hard anyway. The trick here is in careful use of your shield and timing and angle of attacks. In this, it's hugely impressive, and Move's accuracy enables truly tactical bouts.
Winner: Sports Champions
The biggest takeaway from our time with Move is its incredible accuracy. Augmented reality instruments move perfectly with the controller, Frisbees fly with the gentle curve you give them and you can select things with superb precision.
As we said, the controller is comfortable, but not for really long sessions of hard-swinging games. The glowing ball is undoubtedly ridiculous, but is the price you pay for accuracy. You'll get used to having it there, even if anyone who sees it for the first time will raise an eyebrow.
Move has advantages and disadvantages over the Wii remote. The orb enables a higher level of accuracy than the Wii can manage even with MotionPlus, in some cases, but also seems to occasionally restrict Move.
Being able to operate only within strict confines of the camera is fine for one or two people, but when there are more of you, it's handy not to have to worry about such things. We suspect future game programmers can avoid an over-reliance on the camera (and the plethora of motion sensors should be able to compensate for this).
Similarly, while the accurate detection of depth that Move has is fantastic for some games, the inconsistency of sitting four feet away for Start the Party! and standing eight feet away for Sports Champions is irritating.
In fact, our single biggest concern is the space required to really go at the games. It won't be an issue for games like MAG and SOCOM, but it could be the difference between whether this or a Wii is more appropriate for your space.
Somewhere, in the gap between the Wii remote with MotionPlus and Move, is an ideal motion controller. But what we have is mightily impressive, even with its flaws. It's not a revolution of motion control, but a refocus from being unassuming and family-friendly to being all about precision and adding options.
However, it's not cheap. The starter pack containing one Move controller and the PlayStation Eye camera is £49.99. We think Sony would have been wise to include a game with that – Wii Play made buying a second controller far more palatable for millions of Wii owners.
At £35 each, the controllers alone aren't that expensive, but they're all /extra/ cost on top of what you've got already.
As a piece of technology, we heartily recommend Move to PlayStation 3 owners. The motion gaming bandwagon is growing and growing, and isn't going to disappear any time soon. Move won't be for everyone, if only because of its steep price as an optional extra, but those that do invest will find an excellent piece of gaming technology.
Related Links
 |  |
Windows Home Server built into Fine Furniture
Check out this great Windows Home Server built into Fine Furniture build over at Once Around the Block. The build is broken down into build post 1, 2 and 3. The finished system looks great but I think my favorite feature is the power on light which is an edge lit piece of plastic that has been inserted into the edge of a drawer. The power on button has also been hidden inside the drawer.
“The server build was completed by fitting the Samsung hard drive behind the power supply. I made some mounting brackets out of right angled metal brackets I found at the hardware store (they are actually “concealed perlin cleats”). They allow enough height for me to eventually add a second hard drive underneath the first, although at this stage that space is filled with the surplus power supply cables.”
Windows Home Server built into Fine Furniture
Check out this great Windows Home Server built into Fine Furniture build over at Once Around the Block. The build is broken down into build post 1, 2 and 3. The finished system looks great but I think my favorite feature is the power on light which is an edge lit piece of plastic that has been inserted into the edge of a drawer. The power on button has also been hidden inside the drawer.
“The server build was completed by fitting the Samsung hard drive behind the power supply. I made some mounting brackets out of right angled metal brackets I found at the hardware store (they are actually “concealed perlin cleats”). They allow enough height for me to eventually add a second hard drive underneath the first, although at this stage that space is filled with the surplus power supply cables.”
In Depth: How to make your website mobile compatible
The days of building websites targeted solely at desktop or laptop environments are over. Users can and will access your website from a variety of internet-enabled devices.
Accessing the web from mobile devices is far from new. However, the popularity of smartphones and cheaper data packages from network providers have driven a sharp rise in mobile web usage, which is not just reserved for the latest and greatest smartphones.
As website owners, mobile devices also offer some fantastic features typically not available on the desktop – functionality such as clicking a hyperlink in your website to call your phone number, or adding your contact details to an address book. And with more advanced devices, a mobile website can provide more targeted, location-aware content for your visitors.
MOBILE FONTS: Computerlove's Advanced Profile on the iPhone demonstrates how the @font-face technique can be used to bring brand fonts onto a mobile site
With the growing importance of mobile devices and the diversity of access it brings, it's vital as website owners, designers and developers that we think smarter and broader about how we enable visitors to engage with our sites.
You need to have a clear policy and strategy in place for making your website accessible to as many internet-enabled devices as possible. Over the following pages, I'll look at some key lessons the team at Code Computerlove has learned while developing mobile websites. I'll cover general hopefully use and apply to your own websites.
Which technology for mobile sites?
The first question you may be asking is: What technology is involved? Well, a technology stack, served by your existing web server, and utilise your existing web development skills. You use XHTML and CSS. There's no need to invest in learning a new formatting language or resurrect an old one such as WML (Wireless Mark-up Language).
At Code Computerlove, our base-level entry devices have to support WAP 2.0, which in turn supports XHTML-MP (Mobile Profile) and WCSS (Wireless Cascading Style Sheets). We also target high-end devices that fully support XHTML and CSS through to HTML5 and CSS3.
XHTML-MP is simply a subset of XHTML for mobile use and WCSS is a subset of CSS. Whilst XHTML-MP has been adopted and superseded by the W3C with XHTML-Basic 1.1, it's still the most widely supported XHTML variant. If you know XHTML, you already know XHTML-MP and XHTML-Basic. The same applies for WCSS.
Styles can be embedded or linked into your XHTML just like normal desktop web pages. What about more interactive technologies such as JavaScript or Flash? Support for these is inconsistent.
However, at the lowest level, like the desktop version of your website, the mobile website should still work without them.
Device diversity
To reiterate, mobile websites are built using existing web technologies. The main challenge is that given the sheer diversity of mobile devices, support for these technologies is inevitably inconsistent.
Certain devices may not support particular features, or features common to devices may be implemented in different ways. So how can you be sure your mobile website will look and act the same on all mobile devices if they all implement the technology in slightly different ways? How can you test your mobile website effectively? How can you assure your clients that their mobile website will work flawlessly on all devices?
BASIC PROFILE: The Basic profile of the site on OpenWave shows the lowest common denominator
The simple answer to these questions is … you can't! It's a tough call to make, but an important one. It's logistically impossible to set up and test all of the XHTML-supporting mobile devices that are available, as well as dealing with quirks that each device offers.
Ironically, appreciating this early on will actually help your build, testing and quality assurance processes. At Code Computerlove, we came up with a two-pronged approach to help us deal with this diversity.
Firstly we agreed to support at the lowest level, only those devices that support XHTML and CSS. Secondly, we used the concept of 'device profiles'. A device profile is simply a way of logically grouping devices.
The criteria for each group can be based on device capability, manufacturer and operating system, or by any other criteria. We created three core profiles. The 'Basic' profile covers low resolution, text-only devices. The 'Intermediate' profile includes devices that support images and have a screen resolution greater than or equal to 240 pixels (such as the BlackBerry, Nokia N95 and Nokia Xpress).
Finally, the 'Advanced' profile includes the latest smartphones based on WebKit browsers, such as the iPhone and Android phones. These profiles help us both to generalise and specialise device capabilities. They help us form a design, build and testing strategy, as well as enabling us to demonstrate to our customers how a mobile website experience will differ between diverse mobile devices.
These profiles are not set in stone. Neither are they static. Profiles may vary between projects and customers. They reflect which agreed device features are to be targeted, as well as the specific handsets we need to support.
During the life cycle of a website, new profiles may be added and old profiles removed. Additionally, they require regular reviews to guarantee their relevance and to ensure that new devices entering the market are properly assigned to the right profile.
Device profiles play an important part in our build process. We assume that all devices adhere to the basic 'text-only' profile. This is our default, fallback profile that will work on all devices. We then progressively enhance a site for subsequent profiles.
INTERMEDIATE PROFILE: The intermediate device profile for the First Group TransPennine Express site shows how it looks on more advanced mobile devices
For example, the Intermediate profile introduces a wider colour palette and the use of images. The Advanced profile extends this further by using advanced CSS3 techniques or more interactive scripting elements.
Device detection
So how do you determine whether a web page request is coming from a mobile device? And secondly, how do you determine what profile a device belongs to?
We do this by simply inspecting a web browser's user agent details and matching them up in a device capability file, which contains extended details about the device. Each web browser has user agent information about itself, such as vendor, version or operating system.
Irrespective of whether a browser is running on the desktop or mobile device, whenever it requests a page from a server, it also sends this information as part of that request. A server can then use this information to determine whether a request has come from a mobile device or not and what capabilities that device has.
To help with this, there are a number of publicly available 'device capability' files that you can use with your server-side technology. These files contain extended information about the web browser and device accessing your site, such as screen resolution, colour depth, image support, touch screen support, manufacturer and operating system.
These database files come with a number of pre-built helper methods, enabling you to integrate them easily into various server-side web technologies such as PHP, ASP.NET and Java.
At Code Computerlove, we use the paid-for subscription service DeviceAtlas. For a minimal yearly cost, this service provides us with regular database updates as well as an online database for browsing phone capabilities.
There are of course alternatives, and a popular open source project called WURFL (Wireless Universal Resource File) is available from wurfl.sourceforge.net. Alternatively, you could write your own user agent detection routine!
Testing
How do you test your mobile website and device profiles? There's no substitute for using real physical devices. Not only will you see precisely how the site will look, but you'll also experience any hardware challenges that the device will throw up, such as screen size or quirky input mechanisms.
Physical testing on every single device is, of course, logistically impossible. But you should try to obtain a few devices that fit each of your device profile characteristics. And if you're building a site on behalf of a client, it's worthwhile trying to get hold of the common handsets they use. There are also a number of software-based testing tools you can use in your daily build and testing process.
TEST AS MUCH AS POSSIBLE: The more devices you can test on the better, but testing on all devices is next to impossible
For testing your mobile site within your desktop browser, use the Firefox User Agent extension. This extension changes the browser details that are sent to your server, spoofing the server into thinking you're accessing the site from a mobile device.
It's important to note that you'll still be viewing the website via the Firefox rendering engine, not how it will be actually rendered on mobile devices. What it will show you, though, is an approximation of how your site will look across various device profiles.
For greater accuracy, there are a number of device simulators available for you to install on your development box for BlackBerry, Android, and iPhone to name a few. Furthermore, there are hosted options offered by Opera, Nokia and Device Anywhere. The latter is a paid-for service hosting thousands of genuine mobile device simulators.
IMPROVING ALL THE TIME: The Intermediate profile on a BlackBerry shows improved formatting and imagery
Perhaps one of the most useful tools is the now-defunct OpenWave simulator. It's fantastic for experiencing your mobile website against a small resolution, text-only device profile. Due to its limited nature, what you'll get is an immediate indication of whether your navigation and content work for mobile.
Multiple websites
Throughout this article, I've been discussing mobile websites as distinct entities from desktop websites. But do you need to build separate websites? Do you need separate URLs for your mobile website?
There's no right or wrong answer to these questions. In an ideal world, your website should be accessible to all devices – desktop or other – and be capable of rendering itself where possible, using the same content and navigation structure.
This singular approach works brilliantly if you know from the outset that your website is to target multiple devices as this can be factored in accordingly when planning structure, layout and content. However, if you have to retro-fit mobile onto an existing desktop site, it may be easier to run the two as separate sites.
Navigation and content that traditionally works well on the desktop may not when displayed on a mobile device. Desktop navigation may seem unnecessarily complex on a mobile device, and content may need rewriting, either to be shorter and more immediate or to be split across multiple pages.
As the amount of server-side branching logic increases to alter navigation and content for desktop and mobile, it's then a case of either re-evaluating your design, structure and content or, if the two sites serve slightly different objectives, running the two as separate instances.
At Code Computerlove, we've separated out our desktop and mobile website. We've found this easier when internally managing site assets such as style sheets, scripts and images and server-side logic.
We are, however, in a convenient position in that our content management system enables us to share content between multiple websites. So, while both are distinct sites, there's crossover and shared content between both.
Do you need a separate URL to distinguish your mobile website from your desktop website? Again, there's no right or wrong answer to this: it's purely down to your own requirements. You don't need a '.mobi' domain, an 'm.' subdomain or a 'mobile' folder as part of your main website's URL.
At Code Computerlove we've used the 'm.' sub-domain approach for mobile – for example, m.codecomputerlove.com – and www for desktop, as this suits our requirements.
If a mobile device browses to the desktop URL, it's redirected to the mobile site. The mobile site contains a hyperlink that enables the visitor to navigate back to the desktop site if they wish to. Conversely, if a desktop browser visits the mobile site, we don't force a redirect back to the desktop version.
Structure and content
Due to restrictions on screen real estate and the various input mechanisms used by mobile devices, relying on a traditional multi-column desktop-based layout doesn't work. This is true even for smartphones that accurately render desktop websites. The novelty soon wears off when scrolling through and zooming around desktop-targeted websites.
At Code Computerlove, we've developed a standard template structure for our mobile websites, initially based on a template that can be found at MobiForge.
This template follows a single-column, fluid-width design. What this means is that the user need only ever scroll vertically and that the page always fits the available width of the mobile device's web browser. The template is purposely 'light'. It encourages simple navigation with tightly written, relevant and focused page content.
Additionally, the template promotes the use of clean and simple XHTML markup that mobile devices with limited processing capabilities, memory and variable network speeds are able to render quickly.
MOBILE TEMPLATE: This diagram will give you an idea on how to structure a mobile website
The template is composed of a header containing the company name or logo; top and bottom breadcrumb trails displayed on all pages except the homepage, enabling the visitor to navigate back through the site with ease; a page content area; sub-navigation links; and a footer containing copyright information plus, more importantly, a hyperlink to the desktop version of the site.
Final tips and tricks
To wrap things up, here are a few final tips and tricks to help with your mobile website development:
Mobile-specific META tags
A number of XHTML META tags that are specific to mobile websites can be used in addition to the common tags such as author and description. A review of these can be found here.
One important one is the viewport META tag, which can be used to set the initial scale of the width of the site to fit the screen. This is especially important for iPhone browsers. It forces your mobile website to fully fit the screen at the correct resolution and prevents the user 'zooming' into a page, for example:
CSS and presentation tips
Given the variable nature of screen resolutions, it's best as a rule to stick with relative units such as percentages and ems. Relative units will help when it comes to making a quality and scalable mobile design. With perhaps the exception of more advanced smartphones, fonts and sizes are in general poorly supported.
It's best to assume that most devices will only use their default font faces and sizes for XHTML elements. CSS background images tend to be well supported. However, your design must work well in the event that they aren't.
We did find a glitch with BlackBerry browsers. To ensure maximum support, make the URLs to your background images absolute, not relative, to the style sheet.
Image resizing and scaling
It's preferable not to push full-size desktop-targeted images to mobile users. Where possible, ensure your images are optimised accordingly. There may be instances where you need to scale images dynamically on the server to match the screen resolution you're serving to.
If you're using DeviceAtlas for browser detection, you can obtain the actual screen width of the web browser and rescale the image accordingly, using such tools as ImageMagick or any other server-side image manipulation library.
Google Analytics
Avoid using the default JavaScript-generated code. Many mobile devices don't support JavaScript and so using the default code will distort your website statistics. Google now offers tracking code options that are targeted specifically for mobile devices.
This code does not rely on JavaScript. You can find the option under the Advanced tab in the section where you normally generate your tracking code from.
| |
In Depth: OS X Finder tweaks to speed up your workflow
Think about your most-used Mac apps. You might choose Safari, Pages, iPhoto or iTunes. However, an app that might escape your furious list-making is Finder, largely because the Mac's file-browser is so ubiquitous that people often forget it's an application.
Finder is designed to be sleek and efficient, providing you with various ways of accessing and viewing documents on your Mac. But like many Apple apps, it's also multi-layered, having by default a simple setup, but offering extra options 'under the hood'.
We'll look at some of the best options, show how to access and activate them, and detail how they can speed up your workflow.
If you find the current Finder a little too cluttered and pine for the simpler days of Mac OS 9, windows can be changed to resemble those from Apple's older system. Click the pill-shaped button at the top-right of a Finder window and the toolbar and sidebar vanish.
This also reverts Finder to a more 'spatial' model; when accessed from within Icon view, clicking folders opens them in a new window, and you cannot open the same folder in multiple windows – instead, Finder reveals the window that's already displaying the folder you're trying to access.
01. The default setup
If you've not made any changes to Finder's setup since you've had your Mac, it will resemble what's shown in the grab above.
Finder window toolbars will show the Spotlight search field and eight buttons: Back/Forward, the four views (Icon, List, Column, Cover Flow), Quick Look, and Actions. On the desktop, icons for files and folders will be scattered around, positioned wherever you left them, and only some volumes will be visible.
02. Tidy your desktop
There are two steps to a neater desktop. First, go to Finder > Preferences. In General, uncheck volumes you don't want to show on the desktop (bearing in mind mounted volumes can be accessed from Finder window sidebars). Then go to View > View Options and set Arrange by to Name to force items to be listed alphabetically.
03. Show and hide item info
Leave the View Options window open. Keep an eye on the items on your desktop and check Show item info. You'll see that for certain document types, additional information is displayed. This option can in fact be set for any folder (via the View Options window) when you're using Icon view.
04. Tweak icon mode
On the desktop, icons are restricted to a maximum size of 128x128 pixels (adjusted by using the slider in the View Options window). However, open another Finder window, change it to Icon view (via the toolbar button) and drag the slider to the right. At its maximum setting, icons are 512x512 – handy for PDF and image previews.
05. Control column sizes
Column view also has some handy resize tricks, this time relating to column widths. You can drag each column's widget, but a double-click stretches a column to fit the longest item's name. Option-click a widget to expand all columns in this way. Shift+Option-clicking expands all columns to match the widest. Option-drag to resize columns manually.
06. Customise Finder windows
Finder windows can be customised. With the sidebar, system-level items are toggled using the Sidebar section of Finder's preferences, but you can drag any file, folder or app there as a shortcut. (Drag a link out of the sidebar to remove it.) You can also drag items to the toolbar; To remove an item Ctrl-click and select Remove Item.
07. View path information
View > Customize Toolbar offers access to more buttons. Path is useful for rapidly moving up a folder hierarchy. Use View > Show Path Bar for an always-onscreen equivalent at the bottom of each Finder window. Path-bar folders can also be dragged and dropped, and items can be dropped into them.
| |
Review: Norrkross Movie
While iMovie gains significant features in every new release, there's still a big leap from it to Final Cut Express, with regard to both accessibility and capability. Norrkross Movie is aiming to fill that space with some high-end options packed into an intuitive layout.
The important steps up from iMovie are the addition of persistent tracks in the timeline for layered editing and compositing, a much wider range of file import options, a larger choice of effects, and several options for keyframing and nodal compositing of video.
Options aplenty
The tracks all have an icon that shows how they interact with the others around them. It's set to overlay by default, but you can add colour-adjusted transparency effects in a matter of seconds. The difference between most of them is subtle, but the sheer choice is impeccable.
A proper timeline means keyframing of effects can be introduced, and it works well here. It can either be added using just the canvas and timeline, or precisely controlled using the Filters Node View on the left.
Our issue with the tracks is that there's no separation of the audio from the video. iMovie enables you to overlap the audio from one clip to another in the precision editor, but in Norrkross you'd be forced to use an overlay edit to hide the first clip from view, and reduce the volume of the second clip to 0. It's clumsy in practice, and we'd prefer the chance to address editing the audio directly.
The inclusion of nodal compositing is very interesting indeed. Its only appearance in an Apple product is in Shake, part of Final Cut Studio, so we're talking about a fairly professional feature. It effectively enables you to add multiple filters to a clip and then see and control the order they're applied.
For example, if you wanted to remove a green screen background on a clip but also apply motion blur, you can use the nodes to ensure the Make Color Transparent clip is applied before Motion Blur. If it were the other way around, you'd be trying to remove green that blurs into your subject. It's a subtle tool, but one that offers a lot of headroom once you gain confidence with it.
A huge advantage of Norrkross is its ability to handle just about any video format your Mac can play – we even managed to import WebM video without any problems. However, it can't capture footage, only import files. Neither can it export to many formats, but it can do various video sizes in H.264.
The real problem with Norrkross is a lack of polish. There are typos in menus, and the Hide Toolbar option doesn't change to Show Toolbar once you've hidden it. You can't position the playhead and then drag to it a clip to trim, because the action moves the playhead.
There are also far too few keyboard shortcuts, and what is there is often convoluted.
Flies in the ointment
CPU usage can occasionally spike for no discernible reason, and the Media browser can only be viewed as a windowed overlay, but it would be best as a tab in the main program.
If you're looking at upgrading from iLife '09, Norrkross offers strong features at an appealing price. If you've already got the latest iMovie, it's a tougher choice.
Related Links
| |
Review: Softhing Entourage Time Machine 1.0
Entourage is the email app that comes with Microsoft Office for Mac. Some people prefer it to Apple's Mail. Others (like us) use it because IT departments insist on it, typically to work with a Microsoft Exchange server, although Mail also now supports Exchange.
Entourage's problem is that it stores mailboxes in a single file called 'Database'. Most people hang on to emails as a record, and it's not ideal to have all your eggs in one large basket.
Any file that's constantly read and written is liable to corruption. And if you include it in your backups using Mac OS X's Time Machine, that's a big chunk of data to update.
Entourage Time Machine from Softhing aims to help. With one click, it copies all your messages to a set of XML files. You can then browse and search these in a viewer.
It's simple and it works, but not perfectly. Currently only POP accounts are handled. Archiving around 50,000 messages, we found the initial backup glitchy.
Incremental backups are quick, but they can't be automated. The viewer only shows plain text, so HTML emails are garbled. The developers claim searching emails is faster than with Entourage.
But entering search terms often gave us an error, and when it worked it was slow – as was folder switching. The app's not quite there yet, but it has some interesting potential.
Related Links
| |
Guide: How to solve a Rubik’s Cube
We've all seen the algorithms for solving Rubik's Cube by hand using a step-by-step approach: get all the corners done, then get the side cubelets done. Or: do a complete side, then do the next one, then the next.
There are several approaches and, by dint of some fairly rigorous practice, experts can solve a cube in under a minute. But how jumbled can a cube get? Or, to put it another way: what's the minimum number of moves necessary? Enter God's Algorithm.
Way back when, I studied mathematics at Kings College, London. Every year, in the summer term, the Mathematical Society organised a weekend away in Windsor Great Park, where we'd invite guest speakers to present topics we wouldn't normally encounter in our regular maths courses.
Gleaning the cube
In 1979, we had Professor David Singmaster as our guest. His topic was a brand-new toy called Rubik's Cube – not yet officially available in England – and the use of combinatorial mathematics to solve it.
The cube had been invented by Erno Rubik in Hungary some five years previously and at that point Ideal Toys were just on the verge of licensing the cube from Rubik for worldwide distribution. Singmaster had a set of cubes with him that we could buy and, needless to say, after his talk he sold the lot.
Within a couple of months, I'd got the art of solving a cube sufficiently practised that I could regularly solve one within a couple of minutes. As we were maths students, we understood the mathematics behind the cube.
The initial solution that Singmaster discovered used combinatorial mathematics to solve it. In essence, he'd devised a set of combined moves (let's call them Moves, each containing about seven to 12 individual face rotations), that would move around three corners or three sides.
All of the Moves were of the form aba' – that is, a set of rotations a, followed by a single rotation b, followed by the reverse set of rotations that formed a.
Instead of hopelessly randomising the cubelets, the Moves were designed to only swap the positions of three cubelets around. By identifying three cubelets that were out of position, you could solve the cube by repeatedly applying these Moves.
I practised two Moves by heart – one to swap three corners, one to swap three side cubelets – until I could do them in my sleep. With my tuned cube, that meant I could solve a random position in about two minutes. That's not a brilliant time to be sure, but acceptable.
Two questions left open at that time were: how randomised could you make the cube, and what would be the optimal number of moves that an omnipotent solver – in other words a solver who could perfectly analyse the cube – would take in order to render the cube to its default state?
Obviously our combinatorial solution would require many moves – possibly 100 – but what about if you could visualise the solution perfectly? 10? 20? 42? This optimal cube analysis became known as God's Algorithm, not because there is such an algorithm necessarily, but because it gives us something to aim for in our ever-better algorithms for solving the cube.
Back in 1982, Singmaster hypothesised that God's Algorithm might only need a number of moves "in the low twenties", but he was unable to refine that hypothesis much further.
The magic cube
Before we can even begin to solve the cube, we need some notation so that we don't drown in descriptive phrases. Even today, we still use the same notation devised by Singmaster back in 1982 in his book Notes on Rubik's Magic Cube.
The Cube consists of three types of cubelets, assembled together with what looks to be utter magic in a 3 x 3 x 3 cube.
There are 12 edge cubelets, each with two faces of different colours.
Similarly, there are eight corner cubelets, each with three visible faces, with each face a different colour. Finally there are six centre cubelets each showing one face.
The centre squares form the sprung matrix that holds it all in place.
These centre pieces define the colour of their sides in the solved state. Hold the cube in front of you, such that there is one side directly facing you. The six sides of the cube are called Front, Back, Left, Right, Up and Down.
We use Up and Down instead of Top and Bottom because we're about to use the initial letters to signify the rotations of their respective face, and to use both Bottom and Back in this case would clash confusingly.
The letters F, B, L, R, U and D denote a clockwise quarter-turn of the respective face. By clockwise we refer to the direction you rotate the face if you were looking directly at it. A half turn of a face is denoted by either repeating the letter (for example, FF or UU) or by squaring the letter (such as B2 or R2).
A quarter-turn anti-clockwise is denoted by using a prime mark or apostrophe (such as D' or L'). Of course, a quarter-turn anticlockwise could be denoted by repeating a letter three times, but this is rarely seen.
Cross wits
As an example, here's how to get the simple crosses pattern from a default cube: L2R2U2D2F 2B2 (or LLRRUU DDFFBB). To return to the solved cube, just reverse the moves.
For the cleverer looking centre dots pattern, try L'R•U'D•B'F•L'R (here I've separated the moves in pairs to make it easier to see what's going on). Again, to return to the pristine cube, just reverse the moves.
Singmaster's original solution was in three main stages: First, choose a colour (I always go for white as it's the most visible) and then restore that particular face. In general, you do this by first restoring the edge cubelets and then the corner cubelets.
CROSS PATTERN: Putting the Cross pattern on the Rubik's Cube can set you up for some quickfire puzzle-solving
Second, restore the middle layer. This of course means making sure the four edge cubelets are properly positioned and in the correct orientation.
Third, restore the final face. Singmaster did this part in four main phases: flipping the edge cubelets so that they all showed the final colour, forming a cross with the centre cubelet (of course, they could be in the wrong position); restore the edge cubelets to their proper position; place the corner cubelets in their proper position (although they may be oriented incorrectly); twist the corners until they are in the correct orientation.
Singmaster's algorithm was guaranteed to solve the cube, but the number of moves was not optimal in any sense of the word. It could take over 100 moves to solve the cube using his algorithm.
Once Singmaster had published his algorithm (a solution that required you to learn six basic Moves and then apply them over and over), the race was on to reduce the number of moves drastically in order to solve the cube more quickly.
Quite soon after Singmaster published his initial book, Jessica Fridrich devised a four-pass algorithm known as CFOP (Cross, First two layers, Orient last layer, Permute last layer) that proved to be extremely fast for the new sport of speedcubing – that is, solving the cube very fast in competitions.
Unfortunately, the algorithm requires the knowledge and use of some 120 Moves, but offset against that a practiced speedcuber can analyse and solve a randomised cube in about 55 rotations.
Picking up speed
Philip Marshall then described an algorithm that only required learning two Moves (plus the art of knowing how to recognise when to apply them), but that would solve the cube in somewhere around 65 moves.
It's a five-step process: Cross, centre section edges, top edges, five corner pieces, end game. Next up was Lars Petrus' method, which he devised at roughly the same time as everyone else in the early '80s.
He decided to avoid the traditional layered approach used by everyone else and to restore the cube from one corner, building it out via a solved 2 x 2 x 2 cube, to a 2 x 2 x 3 rectangular block (otherwise known as a cuboid) to the completed cube.
Although the first few passes use several types of Moves, the final stages of the Petrus System only use three. Overall the cube can be solved in 45 moves, provided that time is available to study the cube in advance.
In speed contests, the number of moves increases somewhat to something in the region of 60 moves because there's less time to study the cube in order to devise the most efficient solution. Apart from some tweaks of these methods over the years, that's where human-solving now stands.
The fastest speedcubers use some variant of these methods. But what about computer solutions? Can they get closer to God's Algorithm through lengthy analyses of the randomised cube?
The first approaches were made by professor Morwen B Thistlethwaite at the same time as Singmaster was explaining his method, and were published in Scientific American in 1981 by Douglas Hofstadter. In essence, Thistlethwaite divided up the solving process into subproblems.
Rather than concentrating on solving portions of the cube and endeavouring to not jumble up those parts as you tried to solve the remainder of the cube, he concentrated on the kinds of moves you were allowed to make. To do this, he made use of group theory and searching by computer.
He started off with what's known as the cube group. This is a mathematical group whose operations are all the usual moves we've discussed here: F, B, L, R, U, D and the moves obtainable from them (F 2, F', B2, B' and so on).
The number of possible positions in this cube group is immense: 4.3 x 1,019. He then posited another smaller group, one that only allowed the following moves: L, R, F, B, U2 and D2 . Next he worked out a set of tables of the Moves that would take the cube from the larger group to the smaller group.
Once in this smaller group, he devised yet another smaller group that only allowed L, R, F 2, B2, U2 and D2, and then worked out how to transform the cube into a member of this group. From there he went to the next more restrictive group that only allowed L2, R2, F2, B2, U2 and D2. From this particular group it was a small search that led to the final and smallest group of all: the identity group (the solved cube).
It is important to note that Thistlethwaite's algorithm requires many searches at each step down the group chain and is only feasible for computers to do, not humans. Using this algorithm, it is possible to solve the cube in a maximum of 52 moves.
Nearing God's algorithm
The final improvement was made by Herbert Kociemba in 1992. He built his algorithm based on Thistlethwaite's by removing most of the interim groups. Kociemba's algorithm just used three groups: the cube group, the U, D, F2, B2, L2 and R2 group, and the identity group.
He called it a two-phase algorithm, because you transform the cube into a member of the smaller group, and then transform that into the only member of the identity group.
The important thing about the U, D, F2, B2, L2 and R2 group is that the orientations of the corners and edges cannot be changed using those particular operations.
Furthermore, the edges in the middle slice between the Up and Down faces stay within that slice. The first phase uses a modified A* search algorithm known as iterative deepening A* (or IDA) in order to find the moves that will constrain the corners and edges (and the middle slice) of the cube to fit into the second group.
The second phase then searches for the moves to solve the cube using only the restricted moves allowed. In fact the algorithm is a little cleverer than it may at first appear: it solves the cube multiple times in order to find the shortest solution path available.
First it uses the shortest path provided by the first search and transforms the resulting cube to the solved state. Then it uses the less successful paths from the original search and tries to transform those to the solved state.
After completing this process, it chooses the shortest path it finds as the solution. In general, it finds a path that is 20 moves or shorter to solve the cube. Note however, that the shortest path it finds is not necessarily guaranteed to be the most optimal solution.
So, Kociemba's Algorithm, although very effective, can only ever approximate God's Algorithm. We're still waiting for that one.
| |
Name the Thing Contest – 147
The prize this week is a loupe magnifier, this will let you get a close look at whatever you are working on. This contest will run for one week (Sept 4 - 10, 2010) . Ending time is based on central standard time. To enter, identify the item pictured above and give an example of what can be done with it.
Please do not give the answer in the comments.
Send an email to contest @ hackedgadgets.com with “Name the Thing Contest” as the subject, and the message body consisting of:
- The name of the item in the above picture
- An example of what the item pictured above can be used for
The winner will be chosen at random from all of the correct entries.
Below is a picture of the prize.
IFA 2010: Hands on: Toshiba BDX3100KB Blu-ray 3D review
Toshiba's triptych of 3D goodies is complete with the BDX3100KB 3D Blu-ray player, which joins the WL68 3D range of TVs and Satellite A665 laptop in delivering content in the third dimension.
The 3D Blu-ray market may be a small one – currently there is just one disc on sale in the UK – but this is set to explode, with a mountain of 3D content set for cinemas and the newly crowned king of 3D, James Cameron, set to release Avatar in 3D form.
If you aren't lucky enough to own a PS3, which will get 3D Blu-ray capabilities through a firmware update, then your next port of call will be buying a standalone 3D Bu-ray player.
Toshiba's BDX3100KB is a fine-looking device which features more angles that a Lady Gaga haircut.
The from fascia of the machine is almost triangular, which makes it stand out from other players on the market.
Supplied on the front is an easy to access USB, which means you don't have to fiddle around the back when playing content from a USB drive.
And use this you will as Toshiba has kindly made the TV compatible with the DivX and MKV format. So any content crimped from the web should play like a dream on the machine.
Other than the USB slot, the front has a minimal feel. There are a couple of manual control buttons – power, eject, play and stop – and an LCD screen which offers up timecode information.
Playback of discs comes with the standardFull HD 1920x1080p 3D output and this is at 24fps, so your flicks should retain that cinematic feel.
When it comes to audio, you can play either Dolby TrueHD and DTS-HD Master Audio and if you want to you can hook up to 7.1 surround sound to the player.
If you have any home HD movies that you want to play on your TV, then you can do it through the BDX3100KB as it is able to crunch through AVCHD footage. It does seem that whatever codec you throw at it, the Blu-ray player will play it.
It's worth noting that the player is BD Live compatible, so any Profile 2.0 content you want to watch, you can do so. Not that we have found anything decent out there, but the option is open to you.
We have to admit it is not the slimmest 3D Blu-ray player out there but with dimensions of 430mm x 45mm x 225mm it is not the chunkiest either.
1080p upscaling rounds off what is a rather smart device from Toshiba and one which will handle itself against the rest of the 3D Blu-ray players on the market.
The Toshiba BDX3100KB Blu-ray 3D and its its 2D companion (the BDX1100KB) have a UK release date from October 2010 with pricing to be confirmed.
| |
