More Apple/Android Junk?

I have so many things I’d like to write about – topics having nothing whatsoever to do with Apple or Android, like IoC, ARM assembly, rational NoSQL, and of course pragmatic software architecture – but various mobile issues keep mentally demanding that I hop up on a soapbox about them for a bit. I have to get this out of the way.

Mobile computing is absolutely the most important realm for this industry over the coming decade, so pay attention to what is happening because it really, really matters.

CIBC – Leader and Innovator, or Me-Too Wannabe?

The CIBC Platinum Card

CIBC, a large Canadian bank, just launched a nationwide advertising campaign to promote their newly released iPhone banking application. You can see the video on YouTube.

It's notable that CIBC isn't targeting iPhone-only venues with these ads, which they could easily and cost-effectively do, but instead they are promoting this during primetime Olympics coverage. They're putting it front and center on their website.

If you caught the commercial you might have mistook it for an Apple ad, given that the strongest takeaway is a subtle "there's an app for that" message, followed by the implicit declaration that iPhone customers – a small minority of CIBC customers – are the elite, their walled garden needing its own special flowers.

Maybe Apple subsidized the ads and the product itself. It’s a little surprising for a large and profitable bank to look for ad subsidies, but it’s the only conclusion I can draw.

It’s either that or CIBC has an Apple fanboy wreaking havoc from the executive level.

In any case, big deal: CIBC makes an app for the iPhone (first bank in Canada to do so, they proudly boast). Just serving customers, right?

Consider that CIBC has never offered a rich-client Windows application for banking, which is a statement that is true for every Canadian bank as far as I know.

They will let you download data to Quicken or Money or what have you, on whatever platform you’d like, but if you want to bank electronically as an end-user the cross-platform web browser is and has always been your electronic banking tool, even when it limited them to a very simplistic interface.

They knew not to fall into the ActiveX quagmire like, say, South Korea. The banks have always supported just about any modern client equally.

Think about that for a bit: They have never directly supported the rich interface of the overwhelmingly dominant client platform for PCs.

And for very good reasons.

Yet now they have special, premiere support for one far-from-dominant Smartphone.

Given that history of device and platform treatment, it’s natural to presume that they have some fantastic and compelling reason for making this change: Maybe they’re using amazing 3D graphics of money flow or something on the device. Maybe a breathtaking augmented reality experience that allows you to visualize your debt load increasing when the camera is pointed at that new must-have device you really want at the electronics superstore, a virtual banker sternly shaking his head no.

There must be something that they just couldn't do without "going native", right?

Nope.

Their iPhone app features shockingly basic functionality. The single place where it could use something even remotely client-rich – to get the user's location to find the nearest branch – they screw it up and force you to type your location in.

This Is What Web Apps Are Made For

This Is An HTML5 App on the iPhone

This application was made for HTML 5, which humorously would easily allow them to use the Geolocation API to get the user's current location for richer and more intuitive mapping.

And let’s give credit where it is definitely due: The iPhone features excellent web app support, arguably best of class, likely because that was originally the only way to create applications for the device.

Jobs’ original vision was that the phone would offer a native Apple experience enhanced by a rich and robust web application ecosystem. That was the phone that they originally delivered.

That web richness allows you to make apps that look and act just like an iphone application with some simple targeted styles and scripting, offering rich and robust functionality and features.  It also allows you to avoid going hat-in-hand through Apple's app review process for every update, as is demanded when you publish via the App Store.

So imagine a world where CIBC decided that they didn’t need to kneel in worship before Apple, trying to suck some Apple-idolizing droppings from the dirty ground, and they’d release this as an HTML 5 app.

It would feature the same look and feel, could easily support all of the same functionality (without breaking a sweat). It would almost certainly be far more maintainable, and could function like a minimized version of the web app they already have for PCs, without necessarily demanding new public-facing web service APIs.

Win all around, right? Well that’s just the start.

If CIBC did it the correct way – as an HTML 5 app – it would also work on Android devices (including crazy features like local databases and geolocation and all of the snazzy dynamics), such as the hot new Milestone coming to Telus, and the anticipated Acer Liquid and Sony X10 coming to Rogers, and more importantly – this is the land of RIM – it would work on the newly revamped Blackberry webkit browser coming shortly, which is worthwhile given that Blackberry remains atop Apple in the “smartphone” category, especially here in Canada.

Remember that Apple far from dominates the Smartphone category, and competition is only getting fiercer. Canada has never been as iPhone-crazed as the US, and a number of compelling non-iPhone smartphones are just hitting our shores.

So if they went the HTML 5 route, they could offer a rich experience on all capable devices, easily stylable and feature-scaling to optimize the platform experience. Anything would be better than the WAPishly rudimentary "everyone else" dumpbin interface they currently support for every other mobile device.

Didn’t They Listen To Steve Jobs?

Juxtapose Steve Jobs telling us that the iPad doesn't need Flash because HTML 5 makes it irrelevant – a premature statement, but the time will come when his words will seem prophetic – with organizations like CIBC porting absolutely rudimentary web functionality into native apps, wasting time and resources and cash, primarily benefiting Apple, while undermining marketplace choice.

Very backwards move, CIBC. It doesn't make you look hip and on-the-ball, but instead makes you look like Apple-salesmen hoping for your little bit of me-tooism hipster credibility.

Given how boastful CIBC is about being the first bank to feature an iPhone interface, it would be delightful to see another Canadian bank, such as my old workplace RBC, take the high road and come out with rich and robust mobile web apps that don’t favour one walled garden without cause.

They could show it running just as richly on a Blackberry, gaining benefit from the glow of Canadian patriotism.

A Place for Web Apps and a Place for Native Apps

While Jobs is quick to declare the end of Flash in pitching the iPad, the reality is that there are serious gaps in what HTML 5 web apps are capable of.

Graphical games, for instance, aren’t a web app reality without either adding Flash to the equation or going native (e.g. OpenGL on Android or the iPhone). Some day down the road it will be possible, but that isn’t reasonably the case right now, aside from some tech demos that make a high-end desktop grind to a halt.

Apps that exploit special features and functions of the hardware generally can’t be web apps either, at least until the feature is so common and prolific that it gets baked into the shared standards, as geolocation has. I’m sure at some future point we’ll have “camera” and “webcam” and “DSP” APIs to access from JavaScript, but for now those are native app domains.

Mr. Jobs’ statement could more honestly be worded as “Flash isn’t necessary when you have HTML....and apps from the Apple App Store”.

Platform specific apps are needed for a lot of solutions. That goes without saying.

Still, porting absolutely rudimentary functionality to native apps is a backwards repeat of mistakes made in the past, walling the garden off for no logical reason.

CIBC is hardly alone in making this foolish, foolish move, but given that they seem to be so proud of this mistake they deserve particular criticism.

So You Own An iPhone and You Don’t See The Problem

My Youngest Doesn't Have an iPhone...yet

Even if you own an iPhone, and you happily imagine a world where your children’s children will have iPhones, you should still view moves like CIBCs with intense cynicism.

Not only are they limiting your choice unnecessarily if you ever decide to consider alternatives (as everyone should always be doing), even if you’ve declared fealty to Apple forever and ever the movements of organizations like CIBC are diminishing Apple’s need to be competitive.

Recall what happened to Internet Explorer after so much of the web (outside of Canadian banks, notably) decided that they would treat IE users as first class citizens, everyone else ignorable chumps. Once that lock-in was established Microsoft had little incentive to work on their browser and they took their users for granted. We’re still trying to pull ourselves out of that mistake.

Apple isn’t Microsoft, but by the end of 2010 they will likely exceed Microsoft’s market capitalization, which is absolutely shocking. Corporate sludge is inevitable at some point. If something happens to Steve Jobs, for instance, and they recruit Ballmer to run the place, you might decide to consider alternatives only to find that you're tied to the platform in a thousand seemingly minor but cumulative ways.

Competition is good. Building up the walled-garden of the iPhone undermines competition, and encourages a foolish Windowsification of the mobile world.

The NAS Gets a New PSU

In March of last year I wrote about replacing the home NAS with a custom-built Linux box.  

Almost a year in and the device has served the purpose well, providing a solid foundation for a connected home. I’ve been very satisfied with the change.

The only downsides of the unit are the higher power consumption (averaging around 38W), and the groan of the two fans inside: the power supply and chipset fans. The audible part isn’t really an issue given that it’s stashed away, but considering that a probable failure point on most new electronics is the fans, it becomes a reliability concern.

I junked a laptop because of an impossible to repair broken fan. I’ve lost several video cards for the same reason.

I can even hear the irritating whirring of my blu-ray player’s fan (do not buy the Samsung BDP1600. The thing is complete junk even without factoring in the noisy fan trying to upstage the even noisier optical unit. Speaking of junk, the Sony alpha-200 is another garbage product that made me regret ever turning my back on Canon).

As promised in the original entry, I got around to replacing the power supply with a PicoPSU 90W unit, which was basically a plug and play swap.

In my original entry I estimated a 4-8W power reduction, which turned out to be an underestimation. With this PSU the power consumption dropped a whole 10W, going down to a constant 28W (only slightly spiking under load), making me feel a little less enviro-guilt. There’s still the noisy chipset fan, but that’ll be another project.

The case was built around the expectation of a power supply fan exhausting heat, so some extra natural ventilation was required. With that the sensor readings now hover at low operating levels.

Economically this is a change that will not pay off. From NCIX the new PSU cost me $73.49 all in. Given a savings of 0.01kWh per hour, and a fully loaded electric cost around $0.16/kWh, it would take 5 years for the 10W to pay for the change.

It would be nice if all power supplies were mandated to be efficient (they aren’t for most devices because they know it plays zero part in your purchasing criteria. It’s unfortunately one of those areas where legislation is really the only effective solution), because right now inefficiency is the standard. Of course environmental choices don't always yield the expected results.

The Dream is Over...Wake Up With New Phone

In July of last year I wrote about choosing a new smartphone to replace the MotoQ that I had been using. While the MotoQ served a good tour of duty, it was seriously showing its age and was falling behind in the empowering mobile revolution.

While I’d been using variants of Windows CE since before the turn of the century, Windows Mobile was obviously lost in the wilderness. Not only was each equipped device essentially abandoned right after being released, the clearest sign that Microsoft lost the plot could be seen in PocketIE, where the preloaded bookmarks to various Microsoft Mobile pages led to 404 errors.

The team moved onto something new and shiny and had no concern at all for the existing base. Microsoft has a very short attention span to products that don't earn them Windows Office type revenue numbers, so it wasn't a surprise.

For various reasons I did not want an iPhone (we don’t need another restrictive and innovation crushing Microsoft scenario playing out, and I want to develop for the device without embracing the whole cult), despite it being the easy choice. I opined in the first entry that Android seemed to have a very bright future ahead, which is a prediction that seems quite obvious now given that it is the platform of so many incredible devices recently released or on the horizon.

The future is so bright for Android that the robots have to wear shades.

The options in Canada were (and remain) limited, so I went with an HTC Dream (G1) given that it had a keyboard and otherwise had largely the same specs as the newer HTC Magic, aside from what seemed like a minor difference in memory capacity.

 I have to confess to being disappointed with the device.

Functionally it is amazing, and even with Android 1.5 the platform is simply brilliant. When everything operates correctly I am over the moon with the device.

The problem is that everything didn’t operate correctly. For whatever reason the device seems to be horrendously overloaded, so even with virtually no apps installed and nothing beyond the base system running, most actions are plagued by obnoxious pauses, even on a fresh start-up.

I hate pauses.

I stopped using brilliant apps like Weatherbug because they seemed to make the situation worse.

Alas, my long term plan was always that I would buy one of the newer, faster phones when they came to market, while using the starter device for development purposes until that time. If an unlocked Nexus One or Droid/Milestone worked on Rogers’ wireless band, I’d grab one of those when it was a possibility.

Nonetheless, I was pleasantly surprised recently to find that Rogers was offering all HTC Dream owners a free HTC Magic for $0, with the caveat that your term length pushes out. Given that Dream owners can only possibly be 6 or 7 months into their term, that isn’t that tough of a demand. I am on a very reasonable family plan that allows me 5GB / month (which I seldom use more than 1% of), so I feel fairly future-proofed with that foundation and for me it was all win.

So the next day a Magic arrived in the mail and moments later I was up and running with it. With the SIM card removed my existing Dream still works on wifi, where it can browse the web and play media and respond to emails and take pictures, and I can of course put another card in it and continue using it online. I’ll likely install Cyanogen on it now.

Quite pleased about that.

The most shocking thing, though, is that this Magic is much more responsive. It has the same processor as the Dream, so that doesn’t explain the difference. If I had to guess, I’d point to RAM, which on this device comes in at 288MB, compared to the 192MB in the Dream. For comparison both the Droid and the iPhone 3GS feature 256MB of RAM.

The extra headroom over the base OS seems to make all the difference in the world. On the Magic I can see that the free memory is usually less than 90MB, even on a fresh start-up, which notably would put it over the limits of the Dream.

HTC and Rogers claim that they’ll release Android 2.1 for this device in the near future, which makes me especially pleased.

Great move, Rogers. The new HTC Sense update and free month of data is icing on the cupcake.

Firefox 3.6 Released – Web Worker Performance Remains the Same

Back in June I wrote about Web Workers, a fantastic new method to move processing out of the UI thread. To support the entry I posted a variation of the SunSpider benchmark I named Moonbat.

Safari kicked Firefox around in this benchmark. I just tried it with the just released 3.6, and it doesn’t look like much has changed: FF 3.6 does 10 iterations with 4 threads in ~11 seconds, Chrome does it in 2.6 seconds, while Safari leads the pack at 2.3 seconds.

Alas, web worker performance isn’t a critical factor in choosing a browser (my favourite browser remains Firefox), but it would be nice to see it moving in the right direction.

Celebrating My First Home High Speed Overage

Got the cable bill — a bill that pushes into the $250 range per month these days — to find a surprising $11.25 "internet overage fee". Apparently I used 67.5GB last month, while my limit is 60GB. The Steam sales, several purchased HD movies and a couple of on-demand games for the kids on the 360, added to the normal internet usage apparently really added up to a very atypically throughput-intensive month.

I'm not going to cry many tears about it, even though I do think $1.50 a GB is a bit absurd (in an average month I doubt I use 10GB, so now I almost feel obligated to max it out), given that I think by usage pricing would lead to a far better, more open, more honest system for everyone.