Anecdotally, Yosemite (clean install) had an order of magnitude more breakage than Mavericks. YMMV. Perhaps we're so used to the unreasonable expectations of almost everything working that more breakage than usual precipitaes a whinefest "crisis" that gets picked up on TechCrunch as "news." That somehow Apple is "over" because of a few bugs that are fixed in the next patch release. (New releases are buggy, obviously. Best to wait for everyone else to sort them out or help catch them in beta.)
discoveryd is shit. The unofficial no broadcast argument breaks WiFi. You can't disable discoveryd, or it breaks DNS and DHCP.
If there's somwthing you want to block (multicast DNS), get icefloor and/or (Hand Off OR Little Snitch). I use the former and latter.former. OSX Yosemite contains the awesome pf firewall forked from FreeBSD which was forked from OpenBSD. (Apple send some cash to FreeBSD and OpenBSD plz.)
SJ used to be relentless about nitpicking to keep quality up. Probably not happening consistently across all apps and platforms as much. Tim may need to appoint a Quality Czar whom is detail-oriented, accepts no bull and has "wrath of God" authority to make folks take them seriously.
Long-standing, time-wasting bugs I've noticed:
- Mdns broadcast disabling doesn't work.
- Swift playground in Xcode crashes regularly.
- Mobile Safari regularly crashes randomly on backspace in text areas.
- App Store installs corrupted apps but they don't show as corrupted until reboot, and then future downloads fail.
- Mail.app synchronously hangs the UI when processing new email notifications (probable not using a background queue).
Update: I forgot the worst. FileVault 2 + HFS+ corrupting metadata and becoming un-fsck-able (Repair). TimeMachine restore and 5 hours loss per incident.
Also Xcode 6.1 took a week to fix a docset feed bug that they duplicated a docset dmg url. I found the exact correct dmg url, and it still took 3 days to fix. Everyone had broken Xcode Tools docsets until they fixed it. (!)
iBooks can use 100% CPU trying to connect to the (nonexistent) internet when reading an ePub.
All-in-all, these aren't show stoppers, just a collection of things to whine about. It would cause more happiness if they were fixed, but they're mostly navigable.
Definitely not, it's 2 years old and memtest86+, prime95, iozone and iperf for 24 hours each. These are all independent soft errors. Furthermore, I'm continually running stuff that is doing gpg signing/verification and sha256 checksums in VMs... They would fail. Also, there would be kernel panics... I've had zero.
- App Store likely has a problem with storeaccountd / installer framework from the exception traces I've seen (debug menu on).
- HFS+ can corrupt itself irreparably (it's not ZFS). BTDTBTTS.
The funniest one is an auto correct bug. If you make a text area that is limited to n characters, and auto correct changes the word you are typing to be n + 1 characters Safari completely crashes. I learned this one several times filling out a very long web form and just gave up and moved to Chrome.
I bet there's multiple issues in the text input / autocorrect buffer bounds checking. It's possibly an exploitable security issue too, or at least a nice area for research.
To add one to your list, HealthKit hasn't worked on my iPhone 6 for months -- loads no step counts, pulls no data from third-party apps (despite settings), and renders some apps completely unusable (e.g. Jawbone UP when trying to use the M8 coprocessor as the pedometer).
HealthKit seems a bit half-baked. I can see it being really useful in the future once everything is "Healthkit enabled™," and data magically flows in, but it certainly has many bugs.
As an example, for a while I was manually entering my weight every couple days. After two weeks or so of doing this, the data simply vanished without a trace. This happened again after one of the early updates. After that, I simply gave up. Yes, a bug report was sent, but who knows what really happens once it goes through the 'Gates of Radar.'
The SOP is basically never give any feedback to users. You won't know until it's either fixed or it's not months later.
If OSX/iOS were more open source (beyond the kernel and some libs), I think it would be easier to get external developers contributing valuable fixes and offload some support to the community.
It doesn't make much sense to keep the OS totally closed, because the business model is in selling good HW. Sure, control the platform end-to-end ensures quality, but delegating some can be mutually-beneficial.
It's like a small business owner "control-freaking" so they don't put out napkins for fear people will take too many, making it a pain for everyone to accomplish a common task. There is a reasonable amount of control / final say-so that's necessary, but there is "penny-wise, pound-foolish" also.
Funny how open source hasn't resulted in a desktop Linux or Android or ChromeOS or BSD that you want to use, yet. Not for want of effort or money or forks.
But maybe if Apple open source Mac OS X, which you do want to use despite the annoyances, then open source will magically help it become perfect?
On the privacy front:
iOS Spotlight still falls back to searching the web without permission, leaking privacy info. And all the Privacy and Spotlight settings are set correctly.
For grad school, working at the dept in which one would wish to enter is a popular, lower risk approach. The admissions for grad school tends to be either multi-tiered (uni & dept) or defers entirely to the department's admission folks. This tends to be one or two people for smaller programs. If they know you and you work in the same dept, you're probaly already friends by this time, so you'll at least have more intel on what's expected, if not some bias in your favor.
Seems a bit conspiratorial sensationalist. There's probably some subtle bias (humans are fallible, impossible to be purely objective) common to reviewers that could be brought to light with large-scale stastical analysis. Admissions have a tough job because they are inundated with candidates that appear good but the signals of quality, if known, would tend to be easily simulated... Defeating their use.
I don't care that much what admissions offices do, to be frank about it. If they want to put bigoted "holistic"/nonacademic bullshit into their admissions processes, then that's about #43,207 on my concerns list. I don't have kids so I'm at least 18 years from this being my issue and the whole fucked-up mix will be different by then (better? worse? who knows?) I don't think the fault is mostly with the admissions offices themselves. The academic signals they get are not always reliable; high-school grades are hard to standardize, and the SAT is pretty preppable and doesn't go high enough on the math. (The SAT-M should have harder problems that extend it out to ~1000-1100. Or AMC/AIME scores can be given weight in admissions.) But right now, the SAT is too preppable and the relative lightness of academic merit in the admissions process isn't the officers' fault. They are, as you've noted, working with crummy data. In order to believe they're doing better than they would at a dartboard, they have to convince themselves to "see things" in a bunch of 17-year-old strangers.
What would be of value is to reduce the importance of these brands in the world at-large. It's not that I give a damn either way about Stanford or what it does. I do think the pedigree whoring that has crept into "tech" has been to its detriment. Twenty years ago, Silicon Valley was much less pedigree-obsessed. You didn't need a Stanford degree to raise capital. These days, we see pedigree being the most important factor in the Valley determining who gets to be a founder and who is merely "Engineer #2" on 0.1% of a $5M company, and we see absolute shit founders like Spiegel and Duplan tapping into that private welfare system. A bit of reversion-to-truth in the power given to brands might be a good thing... not only for the world, but also for elite universities, which would be pushed in the direction of using academic merit again, because the mystique they get in admitting rich idiots and making admissions appear "holistic" would be blown. That would be good for them, because it would force them to admit better students.
Get more than one (1) IITian on a ski/gambling trip, and it's rank and test war stories the whole way. :) (I aced (800) the SAT I math section without studying a single minute, no prep classes... Not a lie.)
One thing I noticed is how anything that disess Ivy/Pac 10 always gets down-voted. Doubting the value of pedigree is tantamount to Siné out of Charlie Hebdo for poking at a certain religion I guess. People might be down-voting because they hate the circumstances and then do that the same as down-voting a video containing something newsworthy but horrible on (video platform).
Perhaps I'll try an evidenced-based comment sometime with generous, neutral language and let the audience make up their mind.
On the plus side of (undergraduate) academia, it signals completion of a large task.. Which has some value.
(Bias: I've worked at a big brand uni on both the research and business sides.)
English might be German, French, etc. influenced, but it's nowhere as straightforward as German. Also, it is no surprise that it's about as difficult to learn as a foreign language as Xhosa.
It's childish to downvote something unwritten that needs saying as important as historical discrimination. Even though the context of society at large was also extremely discriminatory, yet the loss of and failure of not capitalizing on bright minds and talent for any attribute was and is a disgrace. It's unpleasant to say or think about, but it's necessary to be broadly and deeply educated in reality. Also, one could argue SAT, ACT and similar standard tests often reinforces subtle biases along the lines of IQ tests.
Otherwise, the test itself is curious and reasonable for an entrance examination.
(Personal note: My parents refused to take me to interview to private primary schools that required pictures (We couldn't be more WASP) because they were firmly against supporting the inculcated exclusivity of attributes ahead of merit.)
SATs, ACTs, alumni preference and other current admissions policies reinforce biases too, but any admissions standard will. (If it helps to get one into college, wealthy parents will find a way to optimize their kids around it) It's a matter of degree, and in the past, top schools were much worse.
