"[Apple's] problem is that they thought they did not have a problem."
Crucial insight, there. From the perspective of an observer who has no special affection for Apple, it's surprising that a company of that competence would succumb to such a basic process error.
Apple actually makes this kind of mistake often: they positively excel at some things (retail relationships, hardware supply chains, physical design, software integration) but really suck at other things (social experiences, data harvesting and search, data synchronization and security, dealing with developers): even "within" the Apple community this leads to a number of well-known failures, from Ping (a similar situation to Maps, with Apple suddenly being forced to compete with Facebook as a deal fell through for functionality they considered critical) to MobileMe (now replaced by iCloud, which has its own set of glaring issues which, if you haven't noticed them now, you might the first time you need to restore an 8GB iCloud backup "all-or-nothing" on the same day tens of millions of other users are attempting to restore their devices with a new version of the firmware).
There's plenty worth criticizing about Apple, but you didn't even mention radar://
1) "Social Experiences" - Ping failed where... who succeeded?
2) "Data Harvesting and Search" - These are two terms that are either exactly the same, or extremely different. Neither points at a clear criticism. You realize Apple pioneered desktop search, right?
3) "Data synchronization" - Where's the problem, exactly? Is there any company other than Dropbox that can claim the crown for this? Do you know how hard this is? Who is third place after Apple's second?
4) "and Security" - Again. Huh?
5) I'm terribly sorry you had a software failure and only had a single solitary backup of 8GB of your vital that was difficult to retrieve when Apple was clearly spiking 8 petabytes of traffic in one day. In this clearly entirely hypothetical scenario you've invented.
I hate defending the only company that finally brought UNIX to the desktop just because I happen to wear designer glasses.
5) I'm terribly sorry you had a software failure and only had a single solitary backup of 8GB of your vital that was difficult to retrieve when Apple was clearly spiking 8 petabytes of traffic in one day.
Let me give another example where iCloud fails: iTunes in the Cloud and iTunes Match. I'm regularly unable to download or stream certain songs for days. I can play the rest of the album fine, but one or two songs, you cannot play nor download.
Sometimes, I cannot access my matched music at all. iTunes will do an iTunes Match update, which fails for some reason, and then suddenly all songs are greyed out. Usually, a match update will only succeed after a few hours of trying. During that time I can only play music that is locally available.
Match is a big mess and I will probably not renew it after a year.
> "1) "Social Experiences" - Ping failed where... who succeeded?"
Facebook + Spotify. I actually see people sharing the music they listen to. Not to mention, it is indeed very possible to fail without someone else succeeding - i.e., launching a product nobody ever wanted (say, Ping).
> "You realize Apple pioneered desktop search, right?"
And now the competition is much better than they are. Your point?
> "Where's the problem, exactly?"
iCloud. Have you written an iCloud-enabled app recently? I was in the iCloud session at this year's WWDC, and damn that was a tense session - developers approached this session with "hopefully this shit finally works", because iCloud up until that point was broken, and still largely is.
If you look behind Apple's hyperbolic claims and actually look at the API and work with it, you will realize iCloud is pretty much loosely held together by string. It breaks often, the API is obtuse and more or less undocumented, and there are very few ways to look under the hood when it does break, even while in dev, much less production.
Apple really, really sucks at data synchronization. Here's an interesting test: take an iCloud-enabled app, go into your settings and delete the app's cloud data. You would expect this to delete your data and let you start fresh right? But nope, none of your devices associated with this Apple account can ever use iCloud with this app again.
1) Many other companies have succeeded at getting social to work: Ping did not; do you use Ping? I explicitly go into the story with regards to Ping. I know of no one who considers Ping a success.
2) Apple is fairly good at doing things on the desktop, however their server work leaves much to be desired: the way they attempt to rank search results in their various online store fronts is a commonly cited example of this. I have explicitly been told by people at Apple "we are not server people" when reporting issues. (Also, "harvesting" is the collection of data, and "search" is the process of finding data you need in the massive pile you have collected: these are orthogonal concepts.)
3) There are so many issues with this that there is a small industry surrounding tools and teaching people hacks to make this work correctly; I'm kind of assuming you have sufficient context in the Apple community to know much about the specific issues, which is why I felt it was perfectly acceptable to just list "known failures" rather than having to say "this is how why this one failed".
4) You realize who you are talking to, right? Do I really need to defend this? The sheer fact that I have been, for years, operating what is probably the largest private-sector man-in-the-middle attack ever (many tens of millions of devices worldwide, literally billions of stored signature blobs), on Apple's products, an attack that would be trivially solvable by techniques literally taught in intro-level encryption classes, should be sufficient (but of course when you do what I do you have tons of stories to tell).
5) I haven't had this happen to me, as I don't use iCloud: this has happened to numerous other people, however. Here is one example user (of many), from May, with the non-saurik names changed:
--- Day changed Tue May 29 2012
01:04:34 <xxx> this iCloud shit is getting ridiculous
01:04:59 <saurik> xxx: ?
01:05:16 <xxx> saurik: I upgraded my phone, but my iCloud backup was 7GB, i've spent the last two days straight trying to restore it
01:05:38 <xxx> (if the download is interrupted for one second, it fails, and there is no download resume or anything similar or way to download to iTunes or anything)
01:05:50 <xxx> and on my slow internet, 7GB takes about 13 hours to download
01:06:10 <saurik> ouch
01:06:20 <lmrpq> I backup everything minus the Camera Roll, so my backups stay small. If I update, I use iTunes for a one off backup/restore.
01:06:56 <xxx> yeah; i should have disabled the camera roll
01:07:33 <xxx> saurik: I don't think it fails if the screen doesn't turn off, so right now I'm sitting with it for the ~9 remaining hours on my 7th attempt tapping the screen.
01:07:35 <lmrpq> However, the backup I made before upgrading to 5.1.1 turned out to be corrupt (wtf?) so I had to restore from iCloud minus Camera Roll.
01:07:46 <xxx> lmrpq: how do you restore minus camera roll??
01:07:50 <xxx> oh, iTunes backup
01:08:02 <xxx> i want to restore from iCloud without camera roll, for maybe ~500mb
01:08:08 <lmrpq> No, my iCloud backups don't have camera roll, so I had to do that
01:08:12 <xxx> hmm
01:08:15 <lmrpq> I was not happy.
01:08:19 <xxx> maybe i could finish setup
01:08:20 <xxx> jailbreak it
01:08:26 <xxx> remove the "setup finished" flag
01:08:32 <xxx> hook into the setup script
01:08:41 <xxx> skip parts of the download
01:08:45 <xxx> :<
01:09:25 <xxx> hmm maybe i could also go somewhere with > 10mbps internet
01:09:37 <lmrpq> Apple has a lot of money. I wish they put some of it into making stellar cloud services.
01:11:18 <saurik> xxx: what kind of device?
01:13:03 <saurik> ah, I'm going to resume 4S? (as you said iphone)
01:13:47 <xxx> yaeh
01:13:55 <saurik> I guess its time for the drinking bird
--- Day changed Wed May 30 2012
00:07:33 <xxx> iCloud restore failed.
00:07:37 <xxx> :(
00:07:42 <xxx> time to start again!
The best part of the entire report (which you will note was still plaguing this user a day later, as near the end the timestamps reset for May 30th):
00:42:52 <xxx> all i really want is my chrono trigger save game :(
1) Which companies succeeded at "getting social to work?" Facebook?
2) Okay. I've tried to buy large numbers of servers from Apple in the past. They really aren't server people, it's true. I was trying to buy these servers because I was working for a company whose primary business is searching through petabytes of data. Apple was and still is kicking the shit out of everyone with Spotlight.
3) I'm not saying there have never been bugs in any of Apple's data synchronization software, but Dropbox is the only company doing a better job.
4) I guess I don't realize who I'm talking to. Apple has a great reputation when it comes to malicious code. Until I finish posting this and get to researching your history, I'm going to assume you're behind jailbreakme, or similar. Thanks! Great job! Really, it's hard for me not to come off as hypocritical here, but that kind of work is incredibly valuable.
5) Eh, I have difficulty garnering sympathy for anyone complaining on IRC about ~160KB download speeds for downloading their personal backup from Apple while Apple is also distributing iOS6 binaries to at least 15 million devices or so (based on napkin math derived from here: http://allthingsd.com/20120920/usage-of-apples-ios-6-hits-st...)
You're not wrong about Apple's hubris in general. These are just bad examples.
1) A more relevant example in this case might be Spotify in specific, but sure: Facebook. Do you honestly disagree? If so, I am pretty certain you are in the minority. In a list of 7 reasons a reporter from the Huffington Post "left iTunes for Spotify", #7 was "Spotify gets social" (with specific contrast to Ping).
4) I currently host jailbreakme.com, but comex developed that specific jailbreak. I helped with Corona (5.1 exploit) and run Cydia, Substrate, etc.; I guess thinking about it more, I produced the protection fix for one of comex's JailbreakMe PDF exploits (making me somewhat relevantly related to that project).
Apple does not, in fact, have a good reputation when it comes to malicious code: they are simply sufficiently small players that people don't target them. A lot of people believe otherwise, but as far as I can tell this is because their knowledge of Apple products comes only from Apple's marketing efforts.
In fact, in 2010, reknowned security researcher Charlie Miller (who was winning Pwn2Own every year until he decided to stop attending to protest a rules change) was fuzzing PDF renderers, and found many more exploitable PDF files against Apple's Preview (30-60 failures) than in Adobe's Acrobate (only 3-10).
^ This, combined with first-hand experience with the zero-day PDF exploits from comex (where the second exploit was to the same mechanism as the first, as Apple apparently failed to fix it the first time around) are the reason I install Adobe Acrobat and deactivate Preview on my Mac: at least Adobe manages to fix the bugs that are found.
5) iOS 6 was not released in May. In fact, no iOS version was released concurrent to that reported issue, AFAIK. How is this relevant to the example I dragged up and posted? I specifically went out of my way to find an example that would not fall to simple "but the bandwidth is too much for Apple/Akamai to handle!" arguments, and you didn't even pay attention.
1) I think Ping was a shitty idea, and that no one anywhere has gotten "social" right.
4) I said hello on IRC a little bit ago. I'm not going to argue the point on security when I'm clearly outclassed.
5) Whoops. I'm just being careless there. Sorry about that.
I'm really only arguing any of this because of the remark "Apple actually makes this kind of mistake often." This maps thing is a rare, but colossal fuckup from Apple. First other thing that came to mind was when they cut the original iPhone sticker price by $200 only a couple weeks after it came out.
I realize my question will be off topic but I'm sure everyone in this thread would appreciate the answer. Apart from deactivating Preview, what other measures do you take to protect your Mac?
saurik, thank you so much for responding in this thread. Your answers are very insightful. In regards to 4, I've always assumed that to an extent, Apple tolerates jailbreakers and doesn't put as much effort as they could into fixing the exploits (unless they're critical, like the PDF one.) What do you think?
Apple has a schedule by which they fix bugs; if an exploit is not "dangerous" (such as a remote web browser vulnerability), they do not seem to alter their schedule to fix it, even if it helps people jailbreak: they treat it like any other bug.
As an example, 5.1.1 was subject to Rocky Racoon for many months throughout the beta releases of 6.0. This exploit requires physical access to an unlocked (as in, not at the lock screen: PIN code already entered) device; once the device is unlocked, you can just use it, getting access to e-mails, the address book... whatever you'd like: it isn't really a serious security hazard to also be able to jailbreak it.
If nothing else, I consider it a "common courtesy" to do minimal research into people (comment history, Google search) with whom I am communicating: it lets me get a feel for their level of language fluency, a sense of context for interpreting their values, and a better respect that the other person is actually a "person".
I personally believe that not doing this (either caused by not finding it important, or by an endemic challenge to the action by the communication medium), and not anonymity (as many people in the "use real names" debate claim) is what causes commenting community chaotic collapse (of the form of YouTube, 4chan, and random blogs).
I don't know what you've used for search and replace, but in some cases you have obscured the "saurik names" and left the "non-saurik names" as clear text. :)
? "xxx" and "lmrpq" (my failed attempt at "lmnop") are not the users' actual names. I know the names of the two users involved, and neither of them are present in the text. (I have also been sitting here re-reading the block over and over again, as I tend to do after I post any comment on HN for about ten minutes, looking for typos or other mistakes, and have not found any with respect to the name changes in that giant block of transcript.)
Maybe the format of a message is confusing? Each line is "timestamp <from> message", with the convention that people use a colon to do a targeted address; were I talking to you, the message might be "12:00:00 <saurik> blub: this is my message". HN then collapsed it to a single massive line, and I figured that the contents was sufficiently unimportant that that was better not to expand into an equally-massive vertical wall of text.
Is this true though? It's easy for everyone for all the armchair CEOs to say that Jobs never would have released such an epic fail, but I'm not so sure.
If there's one thing that Apple hates as much as a half-baked UX, it's dependency on another tech company. It seems pretty obvious to me as a casual developer and observer of Google maps over the years that global mapping is really really hard. How could Apple not have known this? I don't think even they have this much hubris.
Now we don't know why negotiations with Google for the maps API failed. Maybe Apple was overconfident, maybe Google was playing hard ball for Android's sake, maybe a bit of both. But regardless of the reason, the contract was not renewed and they had to release something. The iPhone without maps is not the iPhone.
The fact that they used an algorithmic approach to QA says more about the timelines than Apple's beliefs about the overall best approach. There was simply not time to get people on the ground to do a proper QA job. Apple's hands were tied.
Are Bing Maps this horrible? Couldn't they license those if they didn't have a choice? Or what about MapQuest? What about the maps used by GPS units? Couldn't they have made a deal with Garmin or Tomtom or someone like that? It may have taken more work than integration with Google, but it'd almost definitely take less work than starting from scratch. I definitely do not believe that their hands were tied and they were forced into creating their own shoddy, incomplete system.
On top of all that, I think Google has more business intelligence than to actually let the contract for iOS maps go. Maybe both sides were just bluffing and never thought the one party would cut the other loose, or something like that, but you'd think with a deal this big at some point someone would come back with their tail between their legs and reopen negotiations.
Of course, it's all speculation at this point anyway. It'd be nice if someone with knowledge could come in here and elucidate, though the likelihood of that is really small.
Last I heard, Bing was contracting out to Nokia for their maps, as part of the whole Nokia-Microsoft strategic surrend^H^H^H^H^H^H^Hpartnership.
(And I know that for all its faults, Nokia is smart enough to apply human QA to their mapping system, because in my last job I worked on the Web site that those human contractors used.)
> What about the maps used by GPS units? Couldn't they have made a deal with Garmin or Tomtom or someone like that?
They have made a deal with Tomtom (teleatlas).
Concerning Garmin, they don't do maps but use navteq (a subsidiary of nokia). Apple should probably have bought nokia, they would have acquired the best mobile maps technology.
I'm not sure what all that deal entails, so it's hard to say whether it could have been better executed upon, other than to say it's clearly not ready yet. I find it hard to believe, however, that all of the incorrect placement of destinations and, in some cases, even the inclusion of destinations that have not existed for decades, also occur in TomTom's car navigation systems.
TomTom is pretty good at directing you to a street address but is just terrible in terms of finding local destinations. I did a ~2000 mile road trip a few months ago and several times that I was looking for something specific I couldn't find it or found the wrong thing using TomTom. It turned out the best workflow was to look something up in Google Maps, get the address from Google, type that address into TomTom, and let TomTom direct me to that address.
One can always use http://routes.tomtom.com/ as a source for comparison. It would also be good if certain journalists and bloggers did this before shooting their mouths off about data quality, but that wouldn't make a compelling headline.
From the article "2. I suspect that the data and routing functionality that they have from TomTom, while not the best, is simply not the source of their problems. ..."
The author of the article is suggesting the problem is the integration of lots of other data in with the TomTom data is the problem. Not the TomTom data
Apple is smarter than this. They would not have released such a half-baked replacement unless they were forced to. I suspect Google refused to renew knowing Apple were working on a replacement but were not yet ready.
This is Apple's "Vista" moment and will inevitably draw the "wouldn't have happened under Jobs" comments. How Apple recovers will form business school study material for years to come.
I'm not so sure they were forced to. I thought that too, initially, but another commenter somewhere else pointed out how hasty and nervy Google's "maps" announcement was, just before WWDC. In that context, it doesn't sound like something Google was wanting to happen.
What we do know is that Google raised their general prices for map access since the original deal with Apple was made. If we assume Google also wanted to increase the rate for Apple's access, it seems logical that they would be upset that they lost a multimillion dollar deal.
Jobs didn't exactly lack "cajones" but there were other approaches than just "no maps at all"
- try to make the deal with Google happen; Apple has so much money it's hard to understand why they couldn't buy their way into Google Maps for another year or two
- make a deal with someone else (Microsoft / Bing Maps, other map provider) while still working on their own, in-house solution
It's great that Apple decided to build its own map system; the problem is they released it when it was far from ready; Apple didn't use to present half-baked products to the world.
I bought it. It was an MVP. The M stands for Minimal. The V stands for Viable. I still have it and note that Maps in iOS 3 over EDGE is more responsive and slicker looking than Maps in iOS 5.
> Oh, so when it suits your argument it's an MVP, when it doesn't, it's half baked. Got it.
Nice straw man and shoving words into another's mouth.
When the product suits one's purpose, it's viable. When it doesn't, it's half baked. When the original iPhone was released, the bar for viable was much lower. Other smartphones were a mess, UX-wise. Just fixing the UX with a smaller feature set won over a mainstream audience.
Now, the bar for viable is much higher, and it carries the expectations of a Maps application that works at least as well as Google Maps for iOS.
Customer expectations change over time. Pretending that they don't for the sake of argument is either foolish or dishonest.
the decision to switch to there own maps was made while jobs was around. its also not the first time apple replaces a product with an inferior option and insists its better, final cut pro x comes to mind.
nothing really changed with .1, they added some placebo fixes, like xml import/export (except that its a new xml format that only represents trivial edits and thus is useless).
>Crucial insight, there. From the perspective of an observer who has no special affection for Apple, it's surprising that a company of that competence would succumb to such a basic process error.
What insight?
Apple makes most of their stuff themselves -- or gets them from multiple sources.
Maps were an exception, as a key technology that was in the hands of a competitor.
But not because Apple could help it at the time: they made a foray into a new territory, mobile phones, and Google had all the available technology and was a non competitor at the time (the first Android phone was out a year later, Oct 2008).
It's a few years that they have already started the process of moving to something of their own. They bought some mapping companies, etc. Something forced their hands to make the switch before that was over.
In any case, you can't just jump over 8+ years of refinements that Google maps have (not to mention all the "driving endless miles for Google Street View and road info, etc).
Crucial insight, there. From the perspective of an observer who has no special affection for Apple, it's surprising that a company of that competence would succumb to such a basic process error.