A little unfair. A lot more could be going wrong (it could be bricking every phone, for example). And to be honest, while the article seems to think the first update should be the least likely to have a problem, for me the first update is where I will cut them some slack - this is a massive distributed update process that has never been tried before in the wild. Any of us real developers know that it's incredibly hard to actually test a process like this completely....
Am an embedded systems developer, and have dealt with similar updates before.
I think the author is fair in expecting proper testing since there weren't too many variations to handle. Firmware upgrades aren't too complex, especially when you have h/w spec lock-downs. You just need to build in the right safe guards, and build a culture of thorough testing. This looks like a typical bureaucratically executed project with time+cost deadlines to meet, and narrow-minded bosses to please.
Sorry but I can't agree. First they are late to market. They then have an underwhelming product introduction despite spending who knows how much money. Then they flub things the first time they touch the user's phone. From an embedded engineer's point of view it may be acceptable but from a product marketing viewpoint it is ridiculously bad.
What this article points out is that MS has the toughest phone update process in the industry. And for all the whining people do with Android, they may have surprisingly got it right. Let me explain.
Apple's process is pretty easy. They own the phone, firmware, software, etc... It's easy for them to test and roll out. Everyone knows this.
Now Android went with a completely different approach. We give you the OS, now you own everything else. The OEMs have to ensure that any update works on their phone. This is great for Android for two reasons: (1) The OEMs COMPLETELY own testing each new build of Android. Google doesn't have to think about Gingerbread, except for systemic issues. And (2) if there are issues on a given phone, it's not Google's fault -- it's the OEMs.
MS is in a rather horrible position of delivering the OS directly to the phones thus having to test them as they are responsible for failure.
One big problem that comes up here is imagine you have one phone where the update just doesn't seem to work. Are you going to hold up all of the other phones just to get that one phone working?
If MS pulls this off, kudos to them. But I don't envy them. With that said, they pulled it off in the PC market, but that market built itself around MS-DOS, rather than MS-DOS coming out the gate with support for multiple manufacturers.
The only thing I would quibble with is: "This is great for Android". I think you mean "this is great for Google" or even "this makes it much easier for Google".
Assuming "great for Android" means "makes Android a better and more popular platform" -- I think that it will simply lead (indeed, has led) to horrendous fragmentation.
Fair point. I guess arbitrariness is like intuitiveness - it describes an individual's experience more than any objective characteristic of the software.
Contrary to popular belief, sometimes a delay in releasing updates really is a feature.
The article is hilarious on this point. "This just in! OS update kills phones, curdles milk, terrifies children! Also, users might be angry that they're not getting it right away!"
By the way, I shouldn't need to say this, but the plural of a terrifying anecdote is not data. Every upgrade ever pushed breaks something for someone, and the noise from one failure sounds louder than ten thousand successes. We'll see if this dire emergency still looks as bad three days from now when more reports are in.
> I shouldn't need to say this, but the plural of a terrifying anecdote is not data
I hear this repeated a lot, and I think people parrot it without really understanding what they say. I'm not saying that this particular article is presenting properly sampled data, but in reality, a lot of social sciences collect their data by sampling a population and pulling together a plurality of "anecdotes".
That's basically the definition of data. An anecdote _is_ data. Whether or not it's statistically significant depends on the plurality, and sampling procedures, but to call it not "data" is quite unwarranted.
If you believe in western medical science, then you had better believe in anecdotes as data, because to this very day, medical journals report single case studies as data and evidence for dissemination across the medical community.
Just wow. Fire the entire mobile division and get out of the market; you're cursed and/or hopelessly incapable of executing properly on anything in the space.
Actually, my solution was to fire the whole division AND get out of the market.
And this isn't a technical problem; it's a process and culture problem. MS almost certainly has the technical and financial resources to have prevented this. The fact that they didn't points to an utterly dysfunctional company. Getting out of a market that they're rather obviously too hidebound or screwed up to compete in, is definitely a solution.
Yeah, like they got out of the game console market when they were getting red rings of death. Or that time Windows ME was complete and utter crap, and they completely got out of the operating system market. Or that time they did something else wrong, and instead of firing everybody and getting out of the market, they learned from their mistake, came back, and fared very well or perhaps dominated it.
Microsoft, for all their power and ability, are not remotely above mistakes. They also do a pretty decent job of learning from them. I don't know of a home that doesn't have an Xbox 360 in it, or a home that doesn't have some installation of Windows in it.
Sure, it takes them awhile to hit their stride, but the worst thing anybody can do is discount them as a strong competitor, in any market they're in.
The fact is; Microsoft continues to make the mistakes because it never hurts them to make them. By now everyone expects this of MS [1] and everyone is used to not doing anything about it but complaining.
[1] As supporting data, I point to the other HN front page thread that points out the exception that proves the rule: Windows 7 has SP1 and no one cares.
Not necessarily disputing your point but those figures are for 2008.
Current sales figures are roughly 50m worldwide sales for the 360, 48m worldwide for the PS3 (having been available for a year less), 85m sales for the Wii. For reference worldwide sales for the PS2 are currently 150m (and still selling).
All figures for Dec 2010 / Jan 2011. Worth noting that if those figures don't tally with what you see around you, the Xbox360 is way bigger than the PS3 in the USA (18m to 10m I think), the PS3 either equals or beats it in all other markets (notably Japan and Europe).
So yes it's "one of the top selling consoles" and 50m units isn't exactly chopped liver, but it's pretty much third equal in the market. Also worth noting that the whole Xbox division has lost massive amounts over all time (though it's now profitable, it's never recouped the original investment).
But I would agree with the point that while they take time to hit their stride, they do get there and I think Kinnect is a good example of the fact that they're now executing pretty well in this space and indeed innovating which isn't usually a word you'd associate with Microsoft.
Nah, I'm serious - lot of folks have no interest in game consoles, and for those who do, my understanding was that the Wii outsold them both put together.
Your solution is rash and doesn't do much more than illustrate your ignorance. This article is one of the first reports, and doesn't even contain numbers in terms of how many users are affected.
Sit back and wait for the real story to unfold, then make an analysis and contribute something. Simply bursting out with "Fire them all!" and the first report of a problem is childish at best.
Woah, that's a whole lot of bashing. It sounds like a serious fuckup, of course, but we don't know a single detail yet. It could be carriers screwing with firmware, for example.
Count your blessings if it only bricked a handful of phones and forced users to fiddle with their devices to get them working again. At least it didn't install Windows Phone 7 Genuine Advantage.
Something tells me that the author doesn't really know what he's talking about.
> And because every phone is running the same software, well, it should all just work, shouldn't it?
As we all know, an operating system is simply software. It doesn't interact with hardware and, if it did, the hardware is all completely identical, right? There's absolutely no difference between any of the WP7 phones on the market.
> The company was non-specific about the purpose of the update; it's not the copy-and-paste update that will be shipped next month, but rather an update to somehow improve the update process.
It's a test update, one whose sole purpose is to help Microsoft work with the carriers to figure out how to roll out updates. It's not like remotely updating hardware is a trivial problem. It's better to have a test update fail than a functional update fail.
An operating system is simply software. It doesn't
interact with hardware and, if it did, the hardware
is all completely identical, right?
No offence, but that's the stupidest remark I've ever read.
And btw, hardware is almost never completely identical. Take 2 pieces of hardware built after the same spec by the same factory, and under certain conditions you may find them exhibiting different behavior. Hardware is not as predictable as software. It's the job of the operating system / drivers to make hardware predictable.
It's still winter, there's a blizzard outside and my sarcasm detector is malfunctioning. I wish for spring to come, but it will take another 20 days according to the latest forecasts :(
> It's a test update, one whose sole purpose is to help Microsoft work with the carriers to figure out how to roll out updates.
If it's a test update, why is anyone who doesn't work for a carrier and/or Microsoft aware of the problem? If real live customers are having issues, that's not testing, that's screwed up deployment. It doesn't matter how minor the update is.
