Wednesday, 24 July 2013
Ubuntu (Falls Off) Edge...
...or: How Not To Build a Mobile Phone.
If you're in any way into new technology in general, and mobile phones in particular, and especially so if you have a soft spot for FOSS, then you must have already hear of the new Canonical crowdfunding campaign aiming to build a concept mobile phone running both Android (as a phone) and Ubuntu Linux (as a PC when docked) - the Ubuntu Edge.
This post is a rehash of what I already said as a comment on these two Google+ posts with maybe a few additions and/or changes.
First, let's see what's on offer here. You can (should?) also look at Indiegogo project page and browse general tech web sites, but I'll boil it down for you into its essence.
Ubuntu Edge, as described by Canonical, will be a mobile phone with a 4.5" screen with 1280x720 HD resolution, 4GB RAM, 128GB storage (must be flash), 8MP back-, and 2MP front-facing cameras, "fastest" quad-core CPU, and other more-or-less usual and expected gubbins. All of this will be available to generous backers at $600, oh! wait!, $830, oh! wait again! 2 for $1400 - next May (aka May 2014, aka in a year's time - we'll get to that bit of maths later). This thing will dual boot Ubuntu OS for mobiles and Android (versions unspecified) and turn into a full fledged Ubuntu Linux desktop when docked. Sweet!
But wait! Don't you smell something fishy here? OK, maybe not fishy, but distinctly half-baked. Just look - and this is corroborated by interviews given to the tech press - they don't yet even know what CPU they will use (apart from it being quad-core). Now, for your average mobile phone buyer this may not be ringing any alarm bells. Even those who (think they) are in the know usually only care about the CPU when it is listed against the others in the benchmark tests. However, there are two extremely serious issues with this uncertainty.
The first one is potentially a very serious problem for those planning to make a lot of use of the docked desktop experience. Here, the choice between Intel (or AMD) and an ARM CPU can make a world of difference. The former would be much preferable since pretty much every single Linux application is ported to that architecture. Not so with ARM! So, if Canonical eventually decide on ARM CPU you may find your desktop experience potentially severely limited. Would this be an acceptable trade-off for increased batery life (and maybe a tad lower price) is something only you can decide. But you can't because you don't know, and Canonical isn't telling. Now hand over your $600/$830/$700-if-you-buy-two, please.
The second issue with this has to do with the overall success, or at least timeliness, of the whole project. Trust me, because I have been there (and in fact still am): ten months is probably an absolute minimum to complete a new mobile phone project from full design to manufacture, and that is when you already know the key design choice - the CPU. If you leave the choice of CPU until later (apparently "the best available before going to manufacture") you will not complete your project on time. This is because "going to manufacture" is usually only weeks away from "stacking the shelves". If Canonical manages to integrate an unknown part in such a short time frame then they should be in the business of consulting for the big guys, not building concept phones (or Linux distributions, for that matter).
The worst thing here: if Canonical actually knew which CPU they're going with they'd have absolutely no reason not to name it in the proposed spec. Which means they don't, which in turn means they have less then ten months (and less and less, by the day) to integrate an unknown part. A tall task if ever there was one. And don't be fooled into thinking that maybe they already know the type but are dithering and waiting for a faster model. Even the change in speed ripples out to potentially different choices and design of other parts of the phone (in simple terms think at least RAM speeds). So, here we already have seeds of failure and they seem to be being watered quite prodigiously.
And now to the other big part of the almost assured failure of this project: the pricing structure of the backers' "perks".
At the project launch you had the following options: $30 for a warm fuzzy feeling and a mention on the web site, $600 for the phone itself, $10,000 for one of the first 50 numbered phones (eh? who on Earth wants the first 50 off the production line? I thought they were usually ritually destroyed to ward off bad luck), $80,000 you get a cool 100 phones and some corporate perks, too.
After a blazing 12 hours or so, the perks structure suddenly changed: $20 for that same warm fuzzy feeling, $830 for a phone, $1400 for two phones (hip-hip?), and the other two from above.
When questions started pouring, someone from the Canonical camp (who shall remain nameless, but does earn the not-the-sharpest-tool title) admits that the production cost of an Ubuntu Edge phone is in excess of $800 and that the suckers, er, $30 backers were used to subsidise the $600 backers before Canonical decided it's time for people to actually pay for the phone they will (not) get. And they tried to sweeten the new deal by offering two-for-$1400 perk. Now, if this wasn't a stupid thing to do in the first place (change the rules mid game), it sure must make those suckers, er, original $30 backers feel like, mmm, suckers, maybe? Bad Canonical, bad.
The Where (is the money)?
Now, you know what happened next? Of course you knew! The funding ground to a screeching halt. That's what happened, and for very good reasons. Everyone with a $600 burning a hole in their pocket is happy to place a free bet and maybe get a cheap well-specced phone in ten months' time. On the other hand, there are not many people who will stake $830 (or even $700 if they found a buddy) to maybe get a phone not necessarily better than something else that may be around in a year's time. Don't get fooled by Canonical's hype machine (only just bragging about $100,000 an hour funding rate after 36 hours). What happened in reality was: zero to $3 million in 12+ hours when all the $600 perks have been taken, then around $750 thousand! in the following 24+ hours (as of this minute). What I see here, rather than a steady $100k/h funding speed which would comfortably reach the $32 million goal, is $250k/h in the first 12 hours, followed by a much more pedestrian $32k/h for the next 24 hours. To me this is a tenfold decrease in funding speed. At this clip, in the remaining 29 days we'll have raised another $22 million, making the campaign a failure, raising just over $25 million out of $32 required. And that is if even this puny momentum is retained for the remainder of the month.
The Wherefore (art thou common sense)?
By all accounts this is such an ill thought out campaign that it is quite painful to watch it (anyone thinking of slow motion car crashes?). This is, of course, unless Canonical never planned for it to succeed in the first place, but use it as a free marketing and awareness raising exercise, maybe even a sounding board for both themselves and the industry at large. This latter because they also allegedly "briefed" all the major players in the mobile phone arena on both the project and the planned campaign. That they still went ahead with it after talking to the big guys means either they (Canonical) are really stupid and didn't carefully read the minutes of those meetings or the big guys happily pulled wool over their (Canonical's) eyes so they get some free marketing/market research. Whichever it was it seems astronomically unlikely the big guys thought that the project so ill thought out could succeed in the time frame allowed. Or, more likely, none of the big guys actually want a FOSS dual boot/desktop when docked phone to succeed - especially not one they didn't make - so they just smiled and patted Shuttleworth on the back and sent him home to fail in front of the gaping audience that are the mobile phone buyers and enthusiasts.
Now, don't get me wrong. Personally, I'd love for such a phone to be made. Heck, I'd probably be the first in line to buy one. After all, I'm not running a Linux-only shop (but not Ubuntu Unity or Gnome 3, mind!) and Android only mobile shoplet for nothing. And of course I wish Canonical all the very best in building Ubuntu Edge. Once they do, and manage to go for more than a puny 40,000 build run so the price becomes more in line with other similar devices, I'd also probably be the first one to beat the path to their door (provided I can get rid of Unity on the desktop - it is still just dire). But, as this Indiegogo project was conceived, both technically and in funding, I just cannot see it happening. First, I don't think it'll reach its $32 million goal - not by a long stretch. And then, I also just can't imagine how they can meet May 2014 shipping target from where they are now. And every day they miss it is a day which makes your $600/$830/$700-if-you-buy-two phone worth less and less so even thinking about the possibility of delays (and these are likely to be months rather than weeks or days) makes you less and less likely to become a big supporter, and those $20 donations just won't do the trick.
So, in conclusion: if you really have those hundreds you don't know what to do with (which is, let's be frank, unlikely) by all means sign up for a nice phone-to-be. Just don't be too disappointed when in a month's time you realise it just ain't gonna happen, at least not with this campaign. So you may want to either put your hundreds in a bank and wait to see what'll really be on offer come May 2014, or spend them on something else which will give you immediate gratification. And immediate gratification is not to be sneered at, IMO.
Good luck, Canonical and Ubuntu Edge!
There's also one other cheeky bit of Canonical's marketing: the hype they're beating up is all about how much better the specs of Ubuntu Edge are than anything else available. Now, stupid humans we evolved to be, we tend latch on the key phrase "much better", rather than also taking into account that almost unsaid "available" to which we need to add a totally unuttered yet crucial "now". Worst thing, the statement, even when fully parsed, unsaid bits and all, is in fact true. Ubuntu Edge is in fact (or rather will be, or better still, could be) much better than everything else in the market - now. But, and this is a very big but, will it still be so much better than everything else in the market in May 2014? Anyone has a crystal ball handy? No, thought not. Even worse if you also managed to parse May 2014 into "months later than May 2014" because of uncertain hardware spec (and I haven't even mentioned unknown OS versions!). Who's to say other big (or even some newcomer small) guys won't come up with something at least comparable and probably at a similar or even lower price point (big guys can order much bigger production runs - 4 million phones are considerably cheaper to make than 40 thousand). So, why not put this Post Scriptum into the post proper above? To which I respond: why add insult to injury?