Yesterday, I blogged about Motorola’s Motoblur UI, which adds an additional SDK for its specific APIs beyond the standard Android stacks. I reckoned that this might mean more fragmentation, which would push it a step closer to the nightmare that was/is J2ME.
I received two quick reactions to this: one reader commented that this was only bad if you wouldn’t have good tools and compilers. To him (@tederf), I would respond that, while it is certainly true that good tools reduce the friction, raise efficiencies and alleviate overall pain, the smallest common denominator is always just that. In my previous companies, we used to produce up to seven or eight different J2ME builds in order to maximise performance of our games on the huge spread of handsets. Could we have done with one build? Probably. Would the result have been great? Almost certainly not!
Anyway, the more interesting reaction came from the good folks at Motorola themselves. They reckoned (via Twitter; they are @motoblur) that:
with all due respect, I feel you’ve misunderstood motoblur, and android fragmentation concerns are a wee bit overblown.
Now, now. I offered them a guest post here in order to explain this further. I have unfortunately not yet had a response (which I take, applying Twitter attention spans, as lasting silence). But I still wanted to use the opportunity to elaborate a little more on this (and, no, I will not lament Moto’s lost opportunity to feature their wares on this humble site).
To clarify a couple of things outright:
- I would be delighted would I be mistaken (and note that I am not a techie, so this is a distinct possibility!).
- I would be equally delighted would Motorola manage to regain some of its lost ground. The world clearly would be a better place with another strong manufacturer regaining old strengths (although maybe with better UI this time around – which Motoblur certainly seems to offer [see picture on the right of the Motorola CLIQ!).
But let’s go back to the general issue of Android fragmentation threats (the fact that I pointed this out – again – en cas de Moto is of course purely coincidental).
So let’s dive in: with open source software, there is always the intrinsic possibility that fragmentation will occur. Why do people customize it? Because they can! vendors, developers and operators that make up the Open Handset Alliance (which releases Android) can tweak is in whichever way they like (or “need” to) and for any number of reasons: to protect IP, to optimize performance on their network or for certain devices or simply because they feel they need some distinguishing factors, some degree of uniqueness. The result can be, however, as one analyst puts it that
there will be multiple flavors of Android, all of them incompatible with each other. That, in turn, necessitates different versions of each application or updates to accommodate the entire device ecosystem. On the whole, such activity negates the cost efficiencies inherent in the idea of a standard, open operating system, and potentially makes the Android Market a confusing place to shop for widgets.
And that’s what you call fragmentation. Interestingly, there were rumours that Google had made the Open Handset Alliance members sign “non-fragmentation agreements” but it seems that this is either not true or not enforceable.
Others point out that HTC, Samsung, Dell, Verizon, (may I add Motorola?) all have phones on the way that run on different software to the others. Reports of version conflicts, lack of backward compatibility, etc, etc. I mean, hell, there is even an “alternative” Android app store (with 223 apps as of tonight)… Sounds familiar?
Dear Motoblur, if it is different with your SDK, please enlighten us! I am sure I will not be the only one applauding!
Image Credit: http://www.visionmobile.com
Everyone with an iPhone will by now have seen and “played” them: the first generation of multi-player games where most of the computing is done in the cloud, relieving the handset from such labourious tasks. This, according to new report, will soon (well, by 2014) be the norm and will in fact erase downloadable apps (games and other) as we know them today.
The huge advantage of this approach is that it would largely do away with the fragmentation nightmare of the handset market: multiple operating systems, a plethora of screen sizes, all sorts of “interesting” middle-ware solution imposed by either carriers (Sprint anyone?) or OEM to “tweak” an OS to their specs, with the sole result of making things even more cumbersome for developers. It is one reason why the Apple iPhone ecosystem thrives as it does: one OS, one model, one distribution channel, worldwide deployment. If the heavy-lifting is done in the cloud rather than on the handset, one can – presumably – do away with less complex front-ends reducing the work necessary to port the app to various handsets (and, no, please do not choose iMob as the benchmark!). Coding for the cloud would resemble web development rather than today’s mass market application development with armies of post-production and QA engineers filling offices (and overall development costs).
Another, easily overlooked, advantage is battery life: the more (and more complex) processes are computed on the phone hardware, the more the device will consume of the precious and scarce juice; battery life is better with cloud computing!
The obvious downside of mobile cloud computing is a blue sky: no cloud, no app. In posh tech lingo: intermittent network availability, which is to say that the app won’t work if you have no coverage (so no displacing downloadable apps in Oklahoma or on the Scottish isles then…). The aforementioned report suggests HTML 5 and its more sophisticated caching options as the solutions but I am fairly skeptical in that respect: you will not bring about the “unprecedented” sophistication with a little bit of cached data.
Anyway, the more likely scenario is an abundance of network availability by 2014 anyway: by then, large parts of our computing will be done over the air and LTE (and further evolved) networks will surely rule good parts of the world.
So, will it work? The report points out iPhone apps for the likes of Amazon, eBay, etc and, yes, for such apps I am sure it will work: they basically extend a front-end of a web service to a mobile device. The “app” is basically a client that facilitates input and serves as a crutch for the smaller screen where it is necessary to optimise the available data so as not to overburden the little available space with non-core information.
Will it work for more complex apps, in particular games? It will depend. On what? On the type of game of course! How many online racing games do you know? I mean really good ones. None? Ah… Why is that? Because not even the abundance of bandwidth and computing power available today gets to a game experience even close to the 50+ frames per second that the top iPhone racing games can do today (check out Firemint’s Real Racing!). The experience on next-gen consoles is even more astonishing. Need for Speed: Under Cover on XBox 360 or PS3? Woah! All these games require very serious hardware acceleration to convert the high-end graphics into smooth, fast, flowing action. And broadband networks aren’t good enough for that. Period.
For other games, turn-based card games, board games, the Sims, or indeed full-blown MMO, this is another story. These games may require a lot of data but the speed is not as important for good gameplay. So for this group (which is arguably the larger one), cloud computing will probably be the way forward. The opportunity exists in particular (and will be largely sufficient) for all “serious” business applications: these feed for the most part off server-side data sets already and a light mobile client (or browser?) is all they need to provide serious productivity boosts to the corporate world.
So will these run in a (mobile-optimised) browser or via (lighter) downloadable front-end apps? Due to the constraints of the small screen, I believe it will remain very beneficial to the UI and playability to have a specific front-end (and core graphics and processes) on your handset because it will optimise the user experience on a mobile device. To keep this bit as generic and light will be key. Otherwise, it would be porting hell all over again.
But we’ll get there! Just watch out!
Here’s one that nearly slipped through the (well, at least my) net: according to a recent press release, the Eclipse Foundation is set to unveil a unified development platform. It is said that some major players, including Nokia, RIM, Sony Ericsson, IBM and Motorola have joined this initiative already though Android and – predictably – Microsoft and Apple are notable in their absence.
We can depend on the researchers from Juniper after all (or maybe they simply felt bad after reading my post on their last report). Whichever the reason, apparently the mobile content industry could be worth a hefty $167bn (!) if – yes, if – the operators would resolve to allowing a workable commercial environment, namely by limiting themselves to lower revenue shares. Whatever the caveats (which are, as usual, hidden in the expensive main report) this number is topping even the loftiest predictions to date; right on in times of the doom and gloom. The key apparently lies in whether operators would act as dumb pipes (no richness for anyone) or a smart pipe (lots of play money for all players on the value chain). In their own words:
“If MNOs are to benefit financially, they need to move away from their Dumb Pipe roots to the Smart Pipe model, though they will clash with the content providers which already dominate the Smart Pipe. A compromise needs to be found.”
It must be truly bleak: even the best friend of every young telecoms entrepreneur on the fundraising trail whose reports rarely failed to feature as a footnote in an investment memorandum for the next big digital thing now sounds a word of caution. Juniper (whose reports I still cannot afford) issued its latest report on mobile gaming and it actually reduces (for the first time, I’m sure, even if I haven’t checked) its prior predictions on the growth and size of the sector in the next, erm, 20 years…
Funny. Sometimes a theme somewhat haunts you… After I have posted about the demise of Tira Wireless (and added some alternative views on the labyrinth that is platforms and handset fragmentation; also go and revisit my posts on the same topic here and here), today we can read that it will all get worse (or maybe not). I bet they read my recent post on the issue… 
“Hey, we noticed you tried downloading content to your T-610. You may not have realized that this phone is utterly outdated and will give you no joy when playing games. We would like to offer you a discounted upgrade to the brand-spanking new N76/ W880i/ Pearl/ iPhone/ Viewty/… and your life would be so much cooler. We are confident that you would then also have more luck with the girls/boys… Best. Your carrier”
My post on Tira Wireless‘ apparent demise triggered a few e-mails, and it was pointed out that, whilst my observations generally seem to have been accurate, I forgot a few players that actually do deliver porting solutions across platforms (e.g. from J2ME to BREW) rather successfully (and do work with some of the larger publishers, too). There is for instance Innaworks, whose Alchemo solution is pretty powerful.
