Oh, come on. Android is a sufficiently ill-defined platform that applications often end up broken on subsets of devices. It's not difficult to find cases of breakage going between version updates. And that's only on the app side - internal interfaces regularly change, making it impossible to update devices if the manufacturer won't forward-port the binary-only bits of their stack. If it were as perfectly compatible and stable as you imply then you wouldn't need to own redundant phones because updates keep breaking things for you.
Android's not massively worse than other mobile platforms, but we never had an opportunity to figure out whether or not Meego would have provided any level of stability. The N900 shipped with something that was almost, but not quite, entirely unlike Meego, and it shipped a year after the G1. Android had spent that year moving on from first-cut developer phones to devices that people actually wanted to use. Nokia released something large and heavy with a resistive touchscreen. It wasn't so much turning up to a gunfight with a knife - it was turning up to a gunfight while in the final stages of a fatal stroke.
There's plenty of justifiable reasons to explain the utter failure of Meego as a platform, but blaming platform stability when a total of zero devices ever shipped with Meego is just trying to fit your pet argument into an inappropriate hole.
Posted Jul 9, 2012 3:47 UTC (Mon) by jcm (subscriber, #18262)
[Link]
Android isn't perfect, and yes, I do own redundant phones and this is a sad requirement. But my operating rule is that I try to run stock firmware if possible, and I don't just randomly update it. I don't consider it a failure of Android if there aren't updates for older phones. If Google (or whatever phone vendor) want to push an update out to existing users then cool, great, it ought to work, but if they don't, it's not a failure if some user/developer downloads a build of Android (or makes one) and it doesn't work with an older phone. That's like expecting your car manufacturer to field upgrade last year's model through a recall just because there happens to be a newer in-dash software system available that you don't need.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 11:48 UTC (Mon) by mjg59 (subscriber, #23239)
[Link]
I think you've missed my point. If updates break things, how is the platform stable? Given that it's irresponsible to run a network-connected device without applying security fixes I'd really hope that most people do run the updates.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 18:27 UTC (Mon) by markhb (guest, #1003)
[Link]
I think "most people" run their vendor-supplied Android version and apply whatever updates trickle down from the mfgr. and carrier. Under that model, I've never had an update (as opposed to a version upgrade) break anything, and I don't carry or own a backup phone.
So far as the version upgrade goes, the Eclair upgrade for the original CLIQ broke some things, but that was largely because the Moto team did yeoman work forcing 10 lbs. of manure into a 5 lb. bag.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 23:42 UTC (Mon) by landley (guest, #6789)
[Link]
Android has more smartphone preinstalls than anything but iPhone. Non-Android Linux has fewer smartphone preinstalls than blackberry or symbian. The number of consumers who reformat their phone and install a different OS is too small to measure, so in the short to medium term compatability is entirely a vendor issue.
This may change as the platform matures, but as yet we've seen no sign of it. Right now we're in smartphone version of the "ROM Basic" era of PCs, where all the apps are java blobs running in Dalvik and nobody really cares about native code. The PC outgrew ROM basic as it commoditized and people started pushing the limits of the hardware, but to get there the platform had to open itself up to third party vendors who didn't get distribution through IBM.
As long as installing apps goes through Google's app store, it's Dalvik all the way down. And as long as "people who don't have smartphones" is a bigger market than "people who have inferior smartphones", saying that X is an incremental technical improvement over Y will get lost in the noise. And by the time that stops, we'll have a winner.
Feel free to complain about how Google's version is less important than aftermarket Cyanogenmod while iPhone passes 50% and locks in the network effects. Did you know iPhone sales have surpassed Microsoft's entire gross revenue? Sure, Apple's being dickish, which will obviously hand the market over to Linux the same way Microsoft's repugnant behavior handed the desktop to DR-DOS, OS/2, BeOS, and Linux. Obviously they're doomed. (And of course _technical_ inferiority prevented Windows 3.1/95/98 from ever amounting to anything, and network insecurity immediately gutted XP's market share.)
But somehow, I'm not finding these arguments persuasive.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 10, 2012 5:53 UTC (Tue) by Cyberax (✭ supporter ✭, #52523)
[Link]
> As long as installing apps goes through Google's app store, it's Dalvik all the way down.
It's possible to develop Android apps purely in native code without even touching Java. A lot of games do just this, for example.
Java is necessary only if you're using Android's widget toolkit and functionality not exposed in NDK. And even then you can use thin Java wrappers.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 18:55 UTC (Mon) by daniels (subscriber, #16193)
[Link]
You're also missing his point that getting apps to run across the breadth of deployed devices is really, really difficult. It's almost as bad as trying to target desktop Linux.
Android is strong proof that you can actually develop a current smartphone platform while learning nothing from Symbian/Series 60.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 18:56 UTC (Mon) by obi (guest, #5784)
[Link]
Say what you want about the closedness of the Apple's walled garden, but at least they get this right.
I still have an app that's been pulled from the Apple store years ago (trademark violation), and so this hasn't received an update since. Yet it still works after multiple major OS updates (of iOS 3/4/5), and hasn't stopped working.
People shouldn't be afraid to update their phones. And there's no reason perfectly decent Andriod phones should be abandoned and relegated to "legacy" after a year or even a few months. The 3GS will still get a major iOS update in september, three years since its release. This should be the norm.
I don't know if this is in spite of Apple's closed nature, or because of it.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 6:05 UTC (Mon) by drag (subscriber, #31333)
[Link]
> Oh, come on. Android is a sufficiently ill-defined platform that applications often end up broken on subsets of devices. It's not difficult to find cases of breakage going between version updates.
Yeah and this may cause severe problems for Android and possibly on a surprisingly short timescale. Some groups are convinced that Android is going to fall from being the dominate phone platform this year or the next due to these problems.
Microsoft learns from it's mistakes and has enough money that they will keep trying over and over and over again if they think it's important enough. They believe that the mobile platform is critical to the future of their company. They will examine Android and copy what works and abandon what doesn't. Unless Android can keep up then they will lose market share massively.
> Android's not massively worse than other mobile platforms, but we never had an opportunity to figure out whether or not Meego would have provided any level of stability.
I agree.
Meego was a dead on arrival due to the judgement mistakes of people that ran the projects. I don't know if it's a problem with Nokia or other groups (ie: who in particular bears the responsibility), but I do know that what happened is due to mismanagement. When projects and corporations fail in endeavours like this it is 100% a management problem and not technology.
I can safely say this because Meego technology never made it to the market. Consequently, not a technology failure. THAT sort of thing is the failure of the leaders. It can only be their failure. They had years, lots of money, lots of talent, and lots of attention and they didn't do a good job. We can't say that the technology provided X and Android did Y because their is no real information to go off.
The only thing we can do is try to examine why it never made it to market and learn from their leadership mistakes. Their wrong decisions and wrong directions. Why they missed the boat.
(well, in all actuality leaders bear the responsibility and source of most failures. (there are exceptions, of course) Technology problems stem from bad decisions, but in this case it's easy to see, at least from my perspective, the technology never had a chance)
Nokia had Linux tablets on the market in 2006. That was 2 years before Android. That was a year before iPhone. It was 4 years before the iPad. I don't know exactly why it took them from 2006 to 2011 to add telephony capabilities to what they shipped on the Nokia 770, but I have my theories.
Then on top of that even after shipping a development mule type product (N9) to their target developers they decided to go through another rewrite, which seems a bad move.
It's not a fun subject. I loved the concept of the platform.
(This is a bit different from the situation with Linux and netbooks. They were able to get it out and for a while it was very popular (Linux systems topped Amazon best sellers, for example) and had a lot of attention. Unfortunately the platform(s) was soundly rejected by the consumer public in favor of more expensive systems running Windows.)
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 6:09 UTC (Mon) by drag (subscriber, #31333)
[Link]
> Then on top of that even after shipping a development mule type product (N9)
*
Sorry I think I was confused with it's predecessor, N900. I got the timelines mixed up. 2006-2009 vs 2006-2011 and such things. I apologize.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 7:32 UTC (Mon) by kragil (guest, #34373)
[Link]
Android will do fine, no worries. Stupid software patents are the only thing that can stop it now.
Once cheap $100 phones(without contract) run 4.1 everybody will be very happy with Android. 4.1 is finally a great mobile OS, maybe even the best.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 13:23 UTC (Mon) by pboddie (subscriber, #50784)
[Link]
(The referenced article is really good, but it's not possible to say so on the blog, so I'll say so here.)
Nokia had Linux tablets on the market in 2006. That was 2 years before Android. That was a year before iPhone. It was 4 years before the iPad. I don't know exactly why it took them from 2006 to 2011 to add telephony capabilities to what they shipped on the Nokia 770, but I have my theories.
"We can't give the consumer everything at the same time - this is not a phone!" is probably a very large one. As for the general reasons about Nokia snatching defeat from the jaws of victory, here are a few:
Always having the benefit of another iteration to get things "right": unlike the Apple and Android people, developing and then immediately delivering wasn't seen as feasible or even desirable. The five step plan is very telling, I think, and hints at the attitude mentioned above.
Turf wars: I wonder how many e-mails were exchanged about getting access to hardware people for hardware details and specifications, and how many feathers were ruffled because some department head's pet project was going to be upstaged by the N-series tablets.
The obsession over control: people who would have been happy to develop for these platforms were constantly rebuffed when it turned out that various parts were proprietary or secret, and Nokia representatives appeared to want to cultivate the impression that it was their "show", on their terms, and that outsiders should feel lucky to be able to participate in a limited way and, of course, popularise and legitimise the platform by developing for it.
The exercise of control: every time Nokia iterated and changed the platform, no-one had anywhere else to go to for hardware (or support for that partially open hardware) in order to continue what they were working on, being forced to either rebase and keep up with the circus or give up and do something else. It's the rodeo style of community management.
I'm sure there are plenty of other reasons, but each one of those is pretty significant.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 13:33 UTC (Mon) by mjg59 (subscriber, #23239)
[Link]
"I don't know exactly why it took them from 2006 to 2011 to add telephony capabilities to what they shipped on the Nokia 770"
I'd always just assumed that it took them that long to overcome the Series 60 powerbase. Including Maemo, Nokia were developing at least three mobile platforms simultaneously.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 10, 2012 18:02 UTC (Tue) by Kluge (guest, #2881)
[Link]
Which is why it seemed sensible to me to move from GTK+ to Qt, since Qt has had more success running on multiple platforms. Providing a Qt-based API for all their platforms seemed like a good way forward. Though I can't say whether Qt offered the right kind of abstractions for a mobile platform.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 10, 2012 19:31 UTC (Tue) by drag (subscriber, #31333)
[Link]
Moving to QT was a bad move because they abandoned something that was working for something that didn't and wasn't going to for years.
The whole cross-platform thing was a complete red herring and could not possibly make up for the loss of productivity. Even if GTK is non-portable, which it isn't, there is vastly more productive ways to solving that problem.
Another problem is that the toolkit is mostly irrelevant. GTK has warts, but so does everything else. No matter what you choose you are going to have to work with it and modify it to fit your system.
Application developers care about things like documentation and developer tools. As long as the APIs are not completely stupid they will tolerate it the same way that a guy driving a pickup tolerates the fact that he is driving a truck with white cab and automatic when he wants to have one with a white cab and manual transmission.
Customers care about applications and UI.
Changing the architecture over and over again served no purpose but to waste resources and time is a core reason why Meamo/Meego/etc failed miserably.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 10, 2012 19:37 UTC (Tue) by drag (subscriber, #31333)
[Link]
> Even if GTK is non-portable, which it isn't,
Eww. bad double negative. 'even if gtk is non-portable, which it is portable,' was the intended meaning.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 11, 2012 16:53 UTC (Wed) by Cyberax (✭ supporter ✭, #52523)
[Link]
Move to QT was OK. The programming environment in Meego is/was pretty cool.
However, the way they did this move is certainly suboptimal. They could have gradually phased in QT instead of throwing away everything (including DEB->RPM move).
No user value
Posted Jul 15, 2012 14:26 UTC (Sun) by man_ls (subscriber, #15091)
[Link]
Mandatory Spolsky reference of the day: Things You Should Never Do, Part I. Solid advice. Developers might be happier, but users did not care one bit about graphical toolkit or package management.
No user value
Posted Jul 15, 2012 14:36 UTC (Sun) by Cyberax (✭ supporter ✭, #52523)
[Link]
Yet without developers there won't be good programs. So you have to keep developers at least somewhat happy.
Microsoft understands this ("Developers, developers, developers, developers!") and provides nice tools for their platform.
Devs without users are no good
Posted Jul 15, 2012 14:43 UTC (Sun) by man_ls (subscriber, #15091)
[Link]
Ideally you should keep both users and developers happy; but if in doubt, always please your userbase first. I submit Sony PlayStation as an (anecdotal, second-hand) example: even if their dev tools have sucked for a long time, it was not until their users were really unhappy about upgrades that users started leaving the platform.
Devs without users are no good
Posted Jul 15, 2012 15:02 UTC (Sun) by Cyberax (✭ supporter ✭, #52523)
[Link]
Actually, PS3 is a good example. It took a couple of years for decent games to appear on PS3. That's why PS3 sales were initially growing very slowly.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 19:30 UTC (Mon) by obi (guest, #5784)
[Link]
> Then on top of that even after shipping a development mule type product (N9) to their target developers they decided to go through another rewrite, which seems a bad move.
Yeah. I was excited about the platform and got me a 770 as soon as they became available. I really wanted to develop for it.
About a year later they dropped all support for the 770 as the N800 came out, and alarm bells started ringing. Another year after that they announced the current platform based on Debian and GTK+ had no future, and that developers where basically wasting their time and should switch to QT, but oh yeah, QT devices weren't available and wouldn't be for several years. Or, “how to kill your developer base as quick as possible”
I dropped the platform completely right then and there as it seemed to be a project that was completely mismanaged with no future. I'm sorry to see that what I feared came to pass; just glad I didn't invest any more effort in it in the last four years.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 9:27 UTC (Mon) by sumanah (guest, #59891)
[Link]
almost, but not quite, entirely unlike Meego
A more subtle Adams reference than one usually finds, and thus restful; thank you.
a total of zero devices ever shipped with Meego
Tiny nitpick: I presume you're saying the N9 doesn't count. Why?
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 10:12 UTC (Mon) by bros (subscriber, #75198)
[Link]
> Tiny nitpick: I presume you're saying the N9 doesn't count. Why?
I'm sorry to step in, but formally N9 does not have MeeGo OS on-board. What it has is called MeeGo 1.2 Harmattan and that is much closer to Maemo 6, rather then MeeGo.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 12:30 UTC (Mon) by p2mate (subscriber, #51563)
[Link]
I would say it is maemo6 :)
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 12:49 UTC (Mon) by mjg59 (subscriber, #23239)
[Link]
The N9 was based on the Meego core framework but I don't think it met the Meego Compliance spec. On the other hand, I may still have been too harsh there - Meego still seems to exist in the IVI market, so it may well end up shipping there if it hasn't already.
Bergius: The Dreams of the MeeGo Diaspora
Posted Jul 9, 2012 14:00 UTC (Mon) by sumanah (guest, #59891)
[Link]
Ah, got it. Thanks to you, bros, & p2mate for clarifying.