I was skeptical of the new home button and app overview - (why fix something that already works pretty well?) - but it actually looks good and seems to actually improve the experience.
I was the same. The moment I saw the three icons gone I rolled my eyes. I thought it was pretty user friendly as is. That said once demoed I am interested to try it. I would like to have an option to toggle back though if desired because...
When I see UI changes like this my initial concern is not for myself but for my 75 year old mother. It has taken her years to be able to move through Android with some efficiency and the basic consistency has been a good thing. A fundamental UI change like this isn't as easy for users like that to adapt to. I am already dreading the first time she has a Google P based phone. I hope to be proved wrong.
How are UIs on third-party apps affected? If you turn on gestures, does it mean that on a third-party app, one has to use gesture now to do what could be done before using a back button?
If you turn on the gesture mode, it's still largely the same UI, it's just that the "App switcher" button is gone. The Back button is still there, in the same place as before (although it does disappear on the Home screen now, which is neat).
Tapping Home and tapping Back still do exactly the same thing. Long-pressing Home still opens Google Assistant.
Opening the app switcher UI is now done by dragging Home up (instead of tapping App Switcher).
Quickly switching to the last used app is now done by dragging the Home button right (instead of double-tapping App Switcher).
...Say you connect your headphones to your device, Android will surface an action to resume your favorite Spotify playlist...
...but what if we could surface part of the app itself...
What is it with the repeated usage of 'surface' here, instead of the more obvious and common word 'show'? Is it something to do with Android API naming, or something else?
As far as I can tell, it's an Android thing. They even call their display manager thing (I don't know the right term for it) SurfaceFlinger: https://source.android.com/devices/graphics/
The surface in SurfaceFlinger is in the context of "drawing surface", a term of art in computer graphics to refer to a window, bitmap, or other thing that can be drawn on/blitted to.
it seems they have also reused these two familiar words in a way that could add some confusion to the Android developer's life:
* "bundle" (as in "app bundles", a feature that lets Google Play deliver an installer with a stripped down set of resources to the end user's device)
* "action" (as in the new feature of the UI known as an "action" which some have alternatively describe as a "visible intent" or a "deep link with parameters." or something)
It's corporate lingo. More specifically, PM/UX speak. "I'll ping you on gChat the slides on what to surface when users tap on the map. Then we can ideate together."
Only a slight exaggeration, based on my observation of specimens in the wild (many years at Google).
My interpretation of this is that “Surface” carries the connotation of “bring to the surface”, i.e. it’s some existing object that is being highlighted or displayed more prominently. This is terminology that often gets used with search results (e.g. the headline “Google’s new search mode surfaces results from apps on your Android device”).
"Surface [verb]: To make easily readable or noticeable on a web page or app
If you'd rather avoid it: Highlight; showcase; present; display"
"It’s easy to see how handy this word is for web developers, page designers, and business strategists, because it efficiently conveys both the goal and the act of making something appear online in a more prominent way. Although its meaning is transparent, it’s still new and specialized vocabulary, and probably will sound odd to most people."
Corpspeak. Usually on part of Account Managers and sales folks, where every wheel (including the English language itself) must be reinvented in a cycle of no more than 5 years...
...or you're not "moving the needle."
I've been in agency for a decade or so, and yeah, this is pretty much how it works. Don't even get into the use of the word "agile" around here.
I doubt there's an exact definition, but to me "show" can mean that it will bring up an entirely new view (some existing UI element), whereas "surface" means it will be inserted into the context already being discussed.
That is, "surface" means the data is brought to you in your current context, whereas "view" means you are taking to a new context that includes the data.
In this discussion, they are saying that when you search for "lyft" in on-device google search, it will incorporate Lyft info into the search results, as opposed to giving you an option to launch the Lyft app.
So when they say "we will surface this data" they mean we will show it, not that we will allow the normal software to show it like normal.
While I appreciate the concept of learning things like "common next steps" to make a device more usable, going by the history of devices like Google Home or the Google Assistant, I also wonder how much of this is going to depend on "give us all your usage history" settings and simply stop working with those settings disabled. The Google Assistant is already unusable without a pile of search history settings enabled; what will "smart" next actions and similar require?
I don't appreciate it, and hope it can be optionally disabled. I don't find it makes the device more usable.
One example, YouTube suggestions. I watched a "Family Guy" clip, didn't upvote it or comment or anything and suddenly I have 10 channels of Family Guy content "recommended" as if I'm suddenly YouTube's biggest Family Guy addict because I watched one two-minute video.
I don't want my phone anticipating what I want to do, I want to tell it what I want to do. Because when it tries to anticipate, it's usually wrong, and that's annoying.
Perhaps this is nonsense and anecdotal but what's concerning for me is not that they're simply wrong and are showing me irrelevant material, but it's that they can become a feedback loop of dissonance and can be reflection of a shift in viewing habits or pollution by what YouTube thinks is a similar demographic to you, but are quite different. When these recommendations are combined with a pseudo-trusted body (in most people's mind) that can be manipulated, it becomes quite a fruitful environment for propaganda.
Still here! Definitely stay away from preview release on your main phone, but once Google slaps beta on a build stuff is pretty stable (have not tried in car Bluetooth yet -- knock on wood).
As a dev, days like these are the ones where I get my moneys worth buying into Googles hardware :)
Yes! It feels.. very polished. The new notifications and settings area are just awesome. Also the app drawer is pretty great. And the animations... smooth as butter!
If you do OTA (over the air) updates, no, you don't need to ever wipe your device. At the end of the beta they'll just push the retail version and de-enroll your device.
This is my third beta (8 and 8.1) and I never wiped.
I felt the same and had mentally prepared for a factory reset, and all the pains that accompany having to re-setup a phone from scratch.
I upgraded anyway.
I have literally had nothing break or change for the worse, and moreover, the phone (Pixel XL) is noticeably more responsive. I was such a ludicrously smooth upgrade I have been actively looking for colleagues to compliment on the work.
Updated. Runs smoothly without any apparent bugs so far.
You have to enable the new gestures manually, it seems. Also, none of the new app features demonstrated in the conference seems to have made it to this beta. :(
I thought Project Treble was supposed to make the updating process easier from an OEM point of view. If so I wish they supported the Nexus 5X and 6P for one more version.
Treble requires SoC vendors to provide drivers with explicit support. SoCs from Nexus devices aren't supported by Qualcomm so they simply can't be upgraded.
Then Google is free to give me an alternative device for development.
With the emulator broken, and the Nexus 5X unsupported, I am unable to assure that my apps work on Android P, and therefore I have no idea at all if they’ll crash or not.
But if they’ll crash, Google will punish me by downranking my score. A punishment for something I can’t do shit about.
Want me to continue? I’ve got a list of dozens of issues.
The problem isn’t speed (that’s something I could live with, but on Linux with KVM it’s faster than all my real devices anyway), but that it’s literally broken.
1.5 years. The Nexus 5X was still sold by Google 1.5 years ago.
And all the alternatives are significantly more expensive.
I’m building open source apps.
If I put ads in my apps or track my users I’d make thousands and could buy a phone easily.
But I’m trying to do the right thing, and have a monthly budget of $6 (current donation level via Patreon).
If you sell me a phone with Android P for one year of my app income (aka 72€), I’d take it. I’ll even put some of my own money on top, 250€, as I paid for the Nexus 5X, I am willing to pay.
It was released three years ago, which is the date you should use for figuring out the life span of a device. You should expect to get less time out of a phone that's already been out for a year and a half when you buy it.
In my country a used 1st generation Pixel can be had for less than the 5x was when it came out (~$300 vs $379). It's not ideal to have to buy used instead of new but it is doable. The Nokia appears to cost about the same price that the 5x did but im not sure it's easily available everywhere.
Anyways, your situation is unfortunate but I don't consider Google's actions to be egregious in this instance.
Very nice. I am especially excited about these features:
> App Timer lets you set time limits on apps, and will nudge you when you’re close to your limit and then gray out the icon to remind you of your goal.
> Wind Down will switch on Night Light when it gets dark, and it will turn on Do Not Disturb and fade the screen to grayscale at your chosen bedtime to help you remember to get to sleep at the time you want.
I actually have a separate device for fun/addictive apps, and my phone's charging spot is not in the bedroom. But these kinds of features will help orders of magnitude more people than my very manual interventions ever will.
I'm glad Google is seeing themselves as on the side of the user here. Rather than enabling the "up and to the right" religion that has a lot of tech companies chasing DAU and UAM metrics by any means necessary.
I don't particularly care about the release cycle, but somehow they manage to add small features every iteration that attract me. Those sound pretty compelling.
The Treble project is pretty darn exciting too, if only I had such a phone.
You know, I'm pretty sure anyone could have built these apps on Android
You can overwrite the screen in Android. You can replace the home screen in Android. You can scrape running apps in Android. There's a high degree of flexibility
Unfortunately this is probably one of those problems where the cost of building the app is relatively high (because the domain is complicated to get right!) and it would be hard to make too much money on this
it's hard less because of Android and more because you have to support hacked up vendor distros of Android if you want to see any adoption, looking at you Samshit Touchwang
Building an app is way too much barrier to entry... What I wish is something like, right-click or swipe down to open console, enter some script (or copy-paste/import a highly starred from a public repository), close console, ta-da. If the API was flexible enough, the scripts would mostly be short and obviously non-exploiting (i.e. a script for graying-out the screen shouldn't use the internet API) so they would be very easy to review, modify, and manage otherwise.
Hmm for this specific case I think that there might be a lot of of edge cases, just a gut feeling
But I would totally be for some much simpler packaging scheme for native apps/"scripts". It's kinda insane how much money is going into Android from Google but all the dev tools are basically just Java stuff. I really wonder what Android could be if a more lisp-y approach existed as well
A few steps further you will invent something like Emacs. Or maybe Atom. Or browser extensions.
BTW do you remember why making a browser extension is a bit involved, and publishing, even more so? Trusting someone else's code. Copy -pasting it from a message board may end up in a problem.
CF.lumen can do this on rooted devices. I have it set up to apply a gradually-increasing f.lux-like effect as the day turns to night, and then switch to a red-channel-only color mode past midnight.
Some sort of sensible API I guess... Or something automatically generated, like Excel macros, that you can then inspect, understand and modify... If it came with "infinite" undo and some sort of playground/sandbox, people could literally experiment with very low cost.
Apple is known to shamelessly copy features that work and are consumer friendly- I have no doubt that they'll be keeping a close eye on how consumers respond to these updates.
As a longtime iPhone user, I'm very happy that the competitors copy shamelessly from each other. The more they take the good parts from competition and refine it, the quicker maturity can be improved the board.
We're in the era where basic phone capabilities have stabilized. Aspects of competing offerings converging, because it has become more well-defined what the basics of phone is, and the basics of how it's supposed to behave.
Less "innovation" (trial and error), more copying of basics, and marginal improvements is disappointing to techies, but it's good for almost everyone else.
Of course they will pull ahead of each other somewhat now and then. In the grand scheme of things, they're statistically at the same level of development, and becoming more so over time (when it comes to the software).
I don't know if notches are the best example. If Google has been keeping track of display technology, and I have no reason to doubt that they haven't, they had probably saw this coming miles away and had been preparing. Sure, it's not all open and public information, but Google has enough clout and HW ongoings to warrant meetings with display manufacturers regularly. Thus it's not a case of copying, but rather how product lifecycles and price points line up with the development cycle of the technology.
In this case, though, these seem to be, well, heavily inspired by iOS's Bed Time, Night Shift, and DND features. Mostly showed up in iOS9/10 a couple of years ago. The shameless copying goes both ways ;)
your comment reminded me of a terrific article I came across a couple years ago, was thinking "ethical apps"... but that's not it... searched for hippocratic oath software, oh yeah -- Tristan Harris (ex-Googler) and "Time Well Spent" and Humane Tech https://humanetech.com/
(ironic that the website renders poorly on my iPhone given the mission and creds of its founder, but I digress)
help you remember to get to sleep at the time you want
On one side I think this is great, on the other side: I'm not sure what to think of needing to be reminded when to go to sleep. It seems kind of wrong if one needs to be actively reminded of something as simple and straightforward as that.
I think i have a very healty sleep system where i basicially follow no rules. I sleep when I really get tired and wake/stay up when i want.
My last sleep cycle was 11 1/2 hours, because a day before i only got like 90minutes (thats not common for me, but happens - usually i sleep more like 6 hours).
That only works if you dont have a fixed day time job.
Especially when i have appointments i sometimes really need to plan ahead and go to bed earlier so a reminder works.
Just a reminder that research basically found that “catch up on sleep” doesn’t really work. Bodies are astonishing machines, capable of adapting to the most extreme conditions, but they do work better when they can repeat the same routines day after day.
As a fellow 6-hour-sleeper, I think modern society makes it a bit hard for people like me. The reality is that my body wants to sleep in 3-hour cycles, and doesn’t really appreciate doing three cycles in a row, but it still wants to sleep about 8 or 9 hours out of 24. That would work fine in past ages: wake up early, tend to animals and fields, then have a mid-day nap - at a time when it’s too hot to work anyway (at least where my genes mostly come from). Then get back in the game, do social stuff, and be at peak awareness by dusk / early evening, when the danger of common violence gets higher.
These days, when the dominant culture is clearly more “northern” in nature, one is still pushed to be an early riser, but to then do all work and social activity in a continuous chunk, no napping allowed. Because I sleep only 6 hours per night, after a few days I’ll be tired. The only option I have is to go to bed absurdly early, wake up after 6h, do some work in the middle of the night then go back to bed. That’s very antisocial in nature.
> Just a reminder that research basically found that “catch up on sleep” doesn’t really work
Yeah, i dont do that. I just used to be the case I had a night of less sleep, a stress and eventful day and then a night with a long sleep phase. I dont try to catch up or anything. I just slept this without a break (My Gear sport recorded no interruptions and very few restless phases).
It's not so much to help you remember. We all know that if I wake up at 7 and I function best on 8 hours of sleep, that I have to be asleep by 11.
But if I'm so engaged as to lose track of time, or simply undisciplined (just one more dose of political outrage/cute animal picture/epic fail schadenfreude on reddit!), the app closing down will help a lot.
Definitely. We live in very artificial environments.
When I spend a week camping, I don't need to be reminded, because it's really dark and there's been nothing exciting to do for a while. But at home, lights are bright and there are exciting things available at the push of a button. (Thanks to push notifications from things like the news and social media plus autoplay from video providers, they're often available without pushing a button at all.)
I thought the valuable part would be waking up early without the stress of an alarm. But actually the best part has been the auto-dimming in the evening. My brain get the hint. Instead of staying up and hacking on something or binge-watching TV, I am very likely to turn in early enough that I now get plenty of sleep.
I wish the App Timer was like a TV timer, where the TV just turns off after the timer is up. I'd like something that kills a certain app that is running on my phone. It sounds like that isn't what the App Timer will do.
I imagine you would need root on your phone to kill apps from another app.
I'm annoyed that my new device isn't covered here, as well. :( I'll clearly survive, but there seems to be no solid reason for why some of these devices aren't more supported.
"machine learning" is the new "cloud". If you slowly expand the definition of ML to include absolutely anything, then it makes it easier for your marketing teams to attach "powered by machine learning! woohoo!" to try to make everything sound cooler.
The typography in Android P looks strange now with the addition of Product Sans alongside Roboto. The old graphic design rule of "never pair multiple sans serifs" is hardly absolute – different sans families can often be successfully paired as header and body, for example – but in this case, there are headings rendered in Product Sans with barely-smaller subheads in Roboto, which just looks jarring.
Looks like Project Treble actually helps with faster updates - The beta is available on 11 devices from different OEMs. Essential said they were able to put out the beta for their hardware in less than a month's time. I did not want to put it on mine though after reading the known issues.
And I'm not mad at all. I'm glad that both iOS and Android are implementing one of WebOS' best feature. Using the Android P beta now and it seems to be most of the way there.
I don't. Most users probably don't. This isn't going anywhere. It was easily one of the most praised features of WebOS by both users and tech reviewers. There's a reason both Android and iOS are going this way.
Lots of slick UI stuff, but no meaningful updates to things that are well overdue:
- Encryption of specific secrets for each app.
- Fine grained control of what information is shared with apps.
- Ability to selectively deny particular data (and have the app keep working, e.g by giving it an empty contacts list, fake phone ID, etc).
- Ad and js blocking capable browser.
Why should installing e.g. the facebook app, allow it to track my position at all times? Just freaky.
this applies only to installed apps from the store, not system apps like googles or oem apps. you cant disable background activity of any apps. dont want an always listening ai helper from google? too bad for you.
you ideally want something like Xprivacy, which is unfortunately no longer supported
Install something like Netguard (same dev behind Xprivacy) -- which functions as a firewall (vpn) b/w apps and the internet. Turn on logging. Watch all the apps you have installed phoning home and services like graph.facebook.com every few moments.
Another good point -- it sucks stupid amounts of data and radio power. People are always puzzled by how quickly they go through their data, some advertising systems can suck ~500MB/month of cellular data. Some telcos have specifically started showing usage graphs identifying traffic as advertising, but it can be hard to see which app or web page is the culprit.
Just don't use them. Unlike in iOS, you can replace all those apps with other apps that you prefer. No rooting or jailbreaking required. If you don't like Google Assistant, you can use Alexa or Cortana or something entirely different. If you don't like Siri on iOS, tough luck.
You can't disable the only launcher on your phone. Just install another launcher, set that launcher as your default, and the end result is exactly what you would want (no Google launcher running). This is not possible at all on iOS.
Also, you've once again conflated "stock" with Google phone. If you use more precise terminology, you won't be so confused.
I have a different launcher installed, and in use. I backported the changes Google did in their launcher to Launcher3, built that, installed it, and am using it.
And by installing a different launcher, the Google apps never run, which is the effect you wanted to achieve. You gain nothing by disabling the apps after that.
If you were able to disable the apps and then uninstall the the new launcher, your phone would be unusable until you did a factory reset from recovery.
Incorrect. The Google apps are still used and running, e.g. for hotword detection even when the launcher is changed. And even with hotword detection off.
Not on the Pixel. On my Pixel, I can disable the Google app and if I use a different launcher, the Pixel launcher won't run at all. So they're getting more customizable over time.
To clarify, Firefox for Android has full extension support, so you can install NoScript, uBlock, and what have you just as you would on a desktop browser.
I find it quite slow, slower than chrome downloading all the ads. But it is an option I guess. It won't get wider uptake though, the default platform browser is the main game.
And this is a perfect example of the bait-and-switch: Location now goes through Google's location services, so you're not saying "FindMyKeysApp may use my GPS location", you're saying "FindMyKeysApp AND GOOGLE may use my GPS location."
For most apps (unless I'm wrong?) there's no way to let the app use my location without it going through Google's location services and sending data back to the Borg Cube.
go to settings, gps mode, select "always high accuracy"
presto - no google - only actual GPS signals used. But good luck - you'll see just how terrible GPS is indoors and in urban canyons. Without wifi and BLE augmentation it truly sucks.
1) Using visible WiFi networks to determine location sends data to Google to use their geolocation database. Your suggestion here does stop this (assuming you meant to DE-select "always high accuracy" and instead select "device only" which uses the GPS without sending any data offboard.)
2) Google pushes the use of Google Play location services API over Android's location services system API. These services don't work unless you grant Google Play Services permission to use your location. Since Google Play Services also needs network access, you effectively have to grant Google access to your location in order to use any location services whatsoever in any application.
It's the second point to which I was referring in my post above.
That is the same with iOS, but iOS doesn't even give you the option of disabling AGPS. Google Location Services' AGPS is anonymized, exactly like iOS's implementation.
You could use microg, then nothing goes to Google. https://lineage.microg.org – This together with yalp gives you a good experience and keeps you private.
Yeah, I've been on LineageOS since it was CyanogenMod, haven't yet needed Google Play for my (admittedly simplistic) requirements but it's good to know that microg is coming along so nicely in case I get stuck.
But it doesn't. Google Location Services is what gives permission to use Google's AGPS service. Meanwhile, on iOS, there is no way to disable that at all.
If you enable location just once (e.g. to check in to a location) it has that permission until you laboriously find the control in per app settings, and apps can use your location any time from then on, including in the background.
I recently switched to iOS for this reason. Yes there are permissions - but Android still periodically activated location services despite my settings.
Googles insistence on capturing as much data as possible about me is just creepy. I disabled voice search, location history, and everything else I could. Switching to DuckDuckGo made a big difference, but still the data gets swallowed up.
I know that Apple is only slightly better - but I’ll take whatever privacy I can get.
It was said in another thread, Apple's financial interests align better with their users -- Google keep insisting on stalking their own users. I'm so happy I don't have a Gmail account.
I think commenter was pointing that an Android app can fetch location even from the background (albeit w/ limits on frequency due to power consumption). It just needs to get that one Location permission from the user.
iOS has a fine-grained permission setting for this - you can decide whether to allow the app to access location or not when it's not in foreground, i.e "While Using the App".
Denying / delivering fake data is #1 on my wishlist. Wechat (and pretty much all apps from China) needs to have absolutely minimum permissions. Currently I have to put them on a separate phone...
It was hard (and in some cases it still is) but I have started to just not the use apps and services that need blanket permissions. Like I didn't have TrueCaller on my phone until I moved to iOS where it surprisingly works w/o Contact permission. The last time I had checked it would simply refuse to work on Android w/o Contact permission. And I didn't give Uber location permission till it started to ask for "While Using the App" permission. Until then I used to manually enter the address.
Similarly, I have decided not to use many Dropbox apps that need R/W permission to entire Dropbox instead of any one folder that I can choose.
I think we kind of need to show these apps and services their place, so to speak.
Doesn't Android pretty much have all that already?
> Encryption of specific secrets for each app
Not sure what you mean by this, but apps are already sandboxed and can't access each other's data. What specific threat are you trying to protect against?
> Fine grained control of what information is shared with apps
Already exists.
> Ability to selectively deny particular data (and have the app keep working, e.g by giving it an empty contacts list, fake phone ID, etc)
Already exists. Not with fake data though; you just deny the permission and the app has to deal with it.
> Ad and js blocking capable browser
There already are _several_ browsers on Android that do that. Unless you mean you want it in Chrome specifically?
> Why should installing e.g. the facebook app, allow it to track my position at all times?
It doesn't. That only occurs when you grant it the location permission.
Fine grained permissions do exist, your phone is simply not being updated to that version.
You can update permissions on types of notifications, location, storage, contacts etc. all independently of each other. O is great for that and I assume P will only be better.
It does not. You have to explore XPrivacy to know what fine grained looks like. I can not explain. Also Android does not show all the permissions that XPrivacy can.
There is not fine grained permissions. There is something that is better than it used to be, but fine grained is not what I'd call it. For instance, I believe the following should be possible for photos alone.
* Give an app access to only save, not read my photos
* Give an app ability to read one specific photo (think upload a photo to facebook)
* Take pictures without always having access to the camera.
I believe the API for that actually exists, it's just that no app uses it because the UI/UX for that is messy. It's much easier to just requests access to all files.
All of those options exists and are implemented - apps can invoke OS photo picker which exposes a single photo. Apps can store photos to storage without requesting read access to other directories. Apps can request a photo intent which opens camera without exposing anything to them.
You just choose to use apps which instead request you give them permissions to whole photo storage.
That is by definition not fine grained permissions that I'm in control of. Of course apps can do that. But we know that will never be the default way of doing it. I want the OS to be in charge, not only asking the developers to play along nicely, because they will go where the money is.
Saying that I can just choose to use apps that do it "the right way" is the same as saying "you can just not use a smart phone". I can see the value in smart phones and in popular apps. What I'm asking for it a pragmatic choice where the OS puts me in charge, because it's not working to just ask the developers to play nicely.
On desktop maybe, on mobile it has been deliberately crippled. Can't share photos/posts via messenger for example, can't save photos either. Best you can do is screen cap, crop, resend. Using the website also means you are logged in, so you get tracked all over the web.
This looks awful. I've come to realize that any feature with "smart" anywhere in the name, is something I want disabled.
Smart quotes? Only exist to fuck up copypasta code and SQL statements.
Smart spell check? There's a field of internet hilarity about how bad this is.
Smart restore? Best way to lose my data and fuck up my install.
Smart assistant? I've yet to meet even one person who uses GAssistant/Siri/Cortana for anything more complex than "schedule a meeting" or "set an alarm."
Smart handwriting recognition? Refuses to detect the word "fuck". And my handwriting was perfectly legible anyway.
I want my software to be TOOLS. Predictable, ideally interoperable and chainable, but above all: simple. The best pieces of Ux make technology feel like a simple tool. Cut/paste. Messaging conversation-style, with a simple "send" button. Grep.
I find my best note taking software is a text editor, with all formatting turned off. I use ad blockers to disable as many recommendation engines as I can. I use duckduckgo because it just answers my query, without trying to tailor the answers to what the machine (or the advertiser) thinks I want. To mee, simple tools that behave consistently are infinitely preferable to the game of "guess how the computer is misunderstanding the request." Maybe I'm a curmudgeon.
I work with ML/AI, and there are great benefits in lots of applications. Just not in second guessing what people want to do. Even other PEOPLE are terrible at that.
Please don't make me dance around the quirks of your "smart" AI. I promise, I am smarter. I have better context. I know what I want to do. Just get out of the fucking way and let me do it in the simplest, most direct way.
> Smart assistant? I've yet to meet even one person who uses GAssistant/Siri/Cortana for anything more complex than "schedule a meeting" or "set an alarm."
Ten years ago, when we launched the first Android phone—the T-Mobile G1—it was with a simple but bold idea: to build a mobile platform that’s free and open to everyone.
To what quantitative fraction is Android still free and open (as in speech), and how much of it has been moved into closed-source containers?
All of those apps exist in AOSP and have many open source equivalents not developed by Google. The platform that those apps run on is just as open now as it was in the G1 days.
The Pixel 2 is not an AOSP device, just like the Galaxy 8 is not an AOSP device. Why would you think you get the source for all the third party apps?
> Im the days of the G1, the actual Google Talk and Google Search apps were open, the Google Books app was open too!
Those apps as they were are still open (the Google Talk app was/is actually a multiprotocol messenger), though I don't remember a Google Books in AOSP. Even better, there are now superior equivalents to all of these that are also open source.
Because the G1 was an AOSP device, and so was the Nexus One, and so was the Nexus S.
I was barely in middle school when the G1 came out, I saw Android morph from the promised solution as entirely open source platform to what we have today, where the platform and base apps are all proprietary.
No, none of those were AOSP devices either. They all included Android Market and Gmail, as the most obvious examples.
The platform remains entirely open. That's how Chinese people are able to use Android even though Google does not do business there. You confuse the platform with the apps and services that run on it.
By now you can’t access the stepcounter, or movement data, or most raw sensors, you can’t run in the background, you can’t access location (except for raw GPS, not even A-GPS) without proprietary libraries.
Hell, you can’t even use the latest OpenGL without proprietary libraries.
> If you want to do WiFi Geolocation, you now need to use Google Play Services.
This was always using Google's proprietary data, even in the G1 days. They split it out into Play Services, so Chinese phones using another provider would be possible.
> If you want to use background notifications, you now need to use Google Play Services.
Background notifications weren't even possible on the G1 when it launched. When background notifications were first made available, they used a Google service and were never part of AOSP for that reason. Chinese phones use other push services inside the firewall.
> If you want to use shared library of openssl, you now need to use Google Play Services.
This is because Google now updates that library for security. If someone else updates the shared library (because you're in China, for example), you can call that version. Or you can just bundle the library in your app.
> By now you can’t access the stepcounter, or movement data.
None of those existed on the G1 and are simply sensor fusion implementations on the open sensor APIs.
> Hell, you can’t even use the latest OpenGL without proprietary libraries.
> This was always using Google's proprietary data, even in the G1 days. They split it out into Play Services, so Chinese phones using another provider would be possible.
Incorrect, in the past your apps could access the raw data and feed it into Mozilla’s Geolocation APIs. This is not possible anymore (but returning in limited form with 9.0).
> Background notifications weren't even possible on the G1 when it launched. When background notifications were first made available, they used a Google service and were never part of AOSP for that reason. Chinese phones use other push services inside the firewall.
Incorrect. Until Android 7.0, apps could run in background and open their own sockets for background notifications.
> None of those existed on the G1 and are simply sensor fusion implementations on the open sensor APIs.
Incorrect – the open sensor APIs are only available for foreground apps. You can not collect any stepcounter data unless your app is constantly showing a notification in foreground, or is using the Google Fit APIs.
Not on any devices, you’ll need to load bindings through Google Play Services.
> You need to understand the difference between the platform and the services that Google provides on top of it.
I seriously suggest you actually go and try to build apps for non-Google Play devices at least a single fricken time, or try to build ROMs, before you comment. You’re just making a fool out of yourself.
Now you are confusing what the platform allows you to do (which has decreased to help battery life) with what the open source platform has implemented for you (which has increased drastically). Every single statement in your post is an example of that. For example, push services, which were never open source, can still be implemented by third parties as a system app. This has always been the case.
> I seriously suggest you actually go and try to build apps for non-Google Play devices at least a single fricken time, or try to build ROMs, before you comment.
I've built apps for the Amazon App Store and built AOSP from source multiple times. That's not the issue here. The issue is that you were confusing the fact that more of the things on Pixel devices are not open source (true, but Pixel doesn't claim to be an AOSP device) with fewer things being available in open source (which is false) and are now confusing fewer things being allowed to third party apps (true) with fewer things being available in open source (still false).
> The platform that those apps run on is just as open now as it was in the G1 days.
It never was open.
Sure, Google ups a bunch of source code to AOSP after the release of a new version but there is no actual OSS-style development going on. All development is behind closed doors, only done by Google and the public cannot contribute.
I cannot check out the latest development version of Android from github, fix some issues and send a PR. It's all a one-way street.
Agreed. It might have made sense to do development behind closed doors in the early days with the mobile patent war raging, but now that most advances are in apps and services on top of the platform instead of in the platform itself and now that Apple is way behind in features and usability and has no way to catch up, they might as well switch to the Chrome development model, or at least the Fuchsia model where the roadmap and design discussion is closed but the code is entirely open.
My team of ten went from majority iPhone to all but two on Pixels with the last two saying they will switch on their next upgrade. It's not due to cost — the iPhone X or the Pixel costs nothing to the employee. It's due to sheer usability.
Also since Android 7 they started locking down what was possible to access with the NDK, any attempt to link against anything not listed as official NDK API will terminate the application.
With Android P they are doing the same to Java apps that make use of reflection to access non-public APIs or work around bugs due to lack of updates.
To me, Android is definitely not FOSS but it is still open in the sense that matters.
As far as I know, any company can pay to use it on their phones. I think members of the Open Handset Alliance don't have to pay. I'm not sure what it takes to join that, though.
Anyone can download Android Studio on any major operating system and make an android app. There are no limitations as to what features those apps can use. There is no fee to put an app in the app store. People can still sideload their own apps.
Eh, I find that the current adaptive brightness works very poorly. Sometimes slight shifts in my sitting or holding angle will cause the screen to dim.
Coming from iOS, with all the disclosures about how Android leaks to Facebook much more I expected to see permissions locked down a lot more.
There's a difference between "I want to put a location in my status update" to "I want to upload my location 24/7 to Facebook's servers so they can profile the crap out of me".
You can already disable SMS, phone and location permissions (and others) for each app in the previous version of Android (Oreo). I'm sure this has it too.
> You can already disable SMS, phone and location permissions (and others) for each app
This doesn't address the comment you replied to in any way. Surely you see that there's a gap in reasoning between "I want to be able to add my location, just not have my location always broadcast to them" and "you can disable location entirely". Your suggestion specifically violates the "I want to be able to add my location" part of personal control over information.
Or are you suggesting that it's reasonable that a user should have to manually switch apps between white/black lists for behaviors? Personally I would find that idea contemptible.
Bwah? The permissions model is the same on Android and iOS, if you opt into permissions, you have to "manually switch between white/black lists to opt them out"
The difference is on iOS you can specify to share your location only when the app is running in the foreground.
They could both still be improved - it would be nice if you could do one-time permissions, like access the camera, but only this one time (So the permission would be removed when you lock the device or leave the app.)
"Android P strengthens privacy by limiting the ability of background apps to access user input and sensor data. If your app is running in the background on a device running Android P, the system applies the following restrictions to your app:
- Your app cannot access the microphone or camera.
- Sensors that use the continuous reporting mode, such as accelerometers and gyroscopes, don't receive events.
- Sensors that use the on-change or one-shot reporting modes don't receive events.
If your app needs to detect sensor events on devices running Android P, use a foreground service."
> ML Kit offers developers on-device APIs for text recognition, face detection, image labeling and more
This is what I always thought Google should do. Operating systems have completely stagnated, but a truly new and differentiating set of features would be to deliver AI as a platform - not a Google-only feature but a set of open APIs that anybody can use or implement. I hope they continue down this track.
315 comments
[ 3.6 ms ] story [ 263 ms ] threadWhen I see UI changes like this my initial concern is not for myself but for my 75 year old mother. It has taken her years to be able to move through Android with some efficiency and the basic consistency has been a good thing. A fundamental UI change like this isn't as easy for users like that to adapt to. I am already dreading the first time she has a Google P based phone. I hope to be proved wrong.
The new swipe-up-from-Home UI is off by default. I had to turn it on in Gestures. By default, it's the old three-button UI: Back, Home, App switcher.
Tapping Home and tapping Back still do exactly the same thing. Long-pressing Home still opens Google Assistant.
Opening the app switcher UI is now done by dragging Home up (instead of tapping App Switcher).
Quickly switching to the last used app is now done by dragging the Home button right (instead of double-tapping App Switcher).
...but what if we could surface part of the app itself...
What is it with the repeated usage of 'surface' here, instead of the more obvious and common word 'show'? Is it something to do with Android API naming, or something else?
* "bundle" (as in "app bundles", a feature that lets Google Play deliver an installer with a stripped down set of resources to the end user's device)
* "action" (as in the new feature of the UI known as an "action" which some have alternatively describe as a "visible intent" or a "deep link with parameters." or something)
https://developer.android.com/reference/android/app/slice/Sl...
Only a slight exaggeration, based on my observation of specimens in the wild (many years at Google).
"Surface [verb]: To make easily readable or noticeable on a web page or app If you'd rather avoid it: Highlight; showcase; present; display"
"It’s easy to see how handy this word is for web developers, page designers, and business strategists, because it efficiently conveys both the goal and the act of making something appear online in a more prominent way. Although its meaning is transparent, it’s still new and specialized vocabulary, and probably will sound odd to most people."
...or you're not "moving the needle."
I've been in agency for a decade or so, and yeah, this is pretty much how it works. Don't even get into the use of the word "agile" around here.
https://books.google.com/ngrams/graph?content=surfaced&year_...
That is, "surface" means the data is brought to you in your current context, whereas "view" means you are taking to a new context that includes the data.
In this discussion, they are saying that when you search for "lyft" in on-device google search, it will incorporate Lyft info into the search results, as opposed to giving you an option to launch the Lyft app.
So when they say "we will surface this data" they mean we will show it, not that we will allow the normal software to show it like normal.
You can view it yourself by dialing ##INFO## aka ##4636## and then selecting "Usage data".
You’ll see all apps you’ve used, for how long you’ve used them overall, how many times, and when the last time was.
EDIT:
This goddamn markup on HN is utterly broken, I wish it would someday get fixed. But no, of course not.One example, YouTube suggestions. I watched a "Family Guy" clip, didn't upvote it or comment or anything and suddenly I have 10 channels of Family Guy content "recommended" as if I'm suddenly YouTube's biggest Family Guy addict because I watched one two-minute video.
I don't want my phone anticipating what I want to do, I want to tell it what I want to do. Because when it tries to anticipate, it's usually wrong, and that's annoying.
https://developer.android.com/preview/
Is there an easy way to manage this from a worktation? Completely backup my current state and restore it afterwords?
As a dev, days like these are the ones where I get my moneys worth buying into Googles hardware :)
This is my third beta (8 and 8.1) and I never wiped.
If you flashed the beta manually or want to downgrade, then yes, you need to wipe and flash.
I upgraded anyway.
I have literally had nothing break or change for the worse, and moreover, the phone (Pixel XL) is noticeably more responsive. I was such a ludicrously smooth upgrade I have been actively looking for colleagues to compliment on the work.
You have to enable the new gestures manually, it seems. Also, none of the new app features demonstrated in the conference seems to have made it to this beta. :(
Especially when the only available alternative starts at 3 times the price. $900 for the smallest Pixel in Germany, fuck that.
With enough money, it was apparently possible.
With the emulator broken, and the Nexus 5X unsupported, I am unable to assure that my apps work on Android P, and therefore I have no idea at all if they’ll crash or not.
But if they’ll crash, Google will punish me by downranking my score. A punishment for something I can’t do shit about.
It’s absolute madness.
Except it's not? Not even remotely true anymore. Starts up instantly now even, they just showed that off. And if you're opposed to Intel HAXM you can now use Hyper-V https://blogs.msdn.microsoft.com/visualstudio/2018/05/08/hyp...
If I type Ctrl-B, the emulator inputs a gesture.
I can’t reliably emulate certain network issues.
Want me to continue? I’ve got a list of dozens of issues.
The problem isn’t speed (that’s something I could live with, but on Linux with KVM it’s faster than all my real devices anyway), but that it’s literally broken.
It's not unreasonable to expect a developer to buy a new device after three years, in my opinion.
And all the alternatives are significantly more expensive.
I’m building open source apps.
If I put ads in my apps or track my users I’d make thousands and could buy a phone easily.
But I’m trying to do the right thing, and have a monthly budget of $6 (current donation level via Patreon).
If you sell me a phone with Android P for one year of my app income (aka 72€), I’d take it. I’ll even put some of my own money on top, 250€, as I paid for the Nexus 5X, I am willing to pay.
In my country a used 1st generation Pixel can be had for less than the 5x was when it came out (~$300 vs $379). It's not ideal to have to buy used instead of new but it is doable. The Nokia appears to cost about the same price that the 5x did but im not sure it's easily available everywhere.
Anyways, your situation is unfortunate but I don't consider Google's actions to be egregious in this instance.
I can't just throw 300€ every year at Google for the privilege of developing apps, even iOS is cheaper at 400€ every 4 years plus 99€ a year.
If Google actually cared, they would impose support requirements on OEMs that wanted to play on Android playground.
OEMs are expected to push updates just like before, having a certified Treble device is meaningless, if the OEM does not care to provide OS updates.
how authoritarian could Google be about forcing this on, say, Samsung?
and if they granted Samsung an exception, what would the reaction be from other manufacturers?
Given the results of Bada and Tizen, Google would have little to worry about if Samsung ever decides to go rogue.
Well, sure, but you need to invent time travel before that.
> App Timer lets you set time limits on apps, and will nudge you when you’re close to your limit and then gray out the icon to remind you of your goal.
> Wind Down will switch on Night Light when it gets dark, and it will turn on Do Not Disturb and fade the screen to grayscale at your chosen bedtime to help you remember to get to sleep at the time you want.
I actually have a separate device for fun/addictive apps, and my phone's charging spot is not in the bedroom. But these kinds of features will help orders of magnitude more people than my very manual interventions ever will.
I'm glad Google is seeing themselves as on the side of the user here. Rather than enabling the "up and to the right" religion that has a lot of tech companies chasing DAU and UAM metrics by any means necessary.
The Treble project is pretty darn exciting too, if only I had such a phone.
You can overwrite the screen in Android. You can replace the home screen in Android. You can scrape running apps in Android. There's a high degree of flexibility
Unfortunately this is probably one of those problems where the cost of building the app is relatively high (because the domain is complicated to get right!) and it would be hard to make too much money on this
But I would totally be for some much simpler packaging scheme for native apps/"scripts". It's kinda insane how much money is going into Android from Google but all the dev tools are basically just Java stuff. I really wonder what Android could be if a more lisp-y approach existed as well
The developer options in android are a hidden feature for pity's sake.
It looks dead in that repository though.
BTW do you remember why making a browser extension is a bit involved, and publishing, even more so? Trusting someone else's code. Copy -pasting it from a message board may end up in a problem.
We're in the era where basic phone capabilities have stabilized. Aspects of competing offerings converging, because it has become more well-defined what the basics of phone is, and the basics of how it's supposed to behave.
Less "innovation" (trial and error), more copying of basics, and marginal improvements is disappointing to techies, but it's good for almost everyone else.
why should they be ashamed to add a useful feature that somebody else pioneered?
you as a consumer only stand to benefit after all. though i'm not a user of apple devices, i see absolutely not fault in that.
- App Timer is in App Guided Access
- Wind Down is essentially Night Shift + Do not Disturb
(ironic that the website renders poorly on my iPhone given the mission and creds of its founder, but I digress)
The article I first read was https://www.theatlantic.com/magazine/archive/2016/11/the-bin...
Recommended reading.
On one side I think this is great, on the other side: I'm not sure what to think of needing to be reminded when to go to sleep. It seems kind of wrong if one needs to be actively reminded of something as simple and straightforward as that.
I think i have a very healty sleep system where i basicially follow no rules. I sleep when I really get tired and wake/stay up when i want.
My last sleep cycle was 11 1/2 hours, because a day before i only got like 90minutes (thats not common for me, but happens - usually i sleep more like 6 hours).
That only works if you dont have a fixed day time job.
Especially when i have appointments i sometimes really need to plan ahead and go to bed earlier so a reminder works.
As a fellow 6-hour-sleeper, I think modern society makes it a bit hard for people like me. The reality is that my body wants to sleep in 3-hour cycles, and doesn’t really appreciate doing three cycles in a row, but it still wants to sleep about 8 or 9 hours out of 24. That would work fine in past ages: wake up early, tend to animals and fields, then have a mid-day nap - at a time when it’s too hot to work anyway (at least where my genes mostly come from). Then get back in the game, do social stuff, and be at peak awareness by dusk / early evening, when the danger of common violence gets higher.
These days, when the dominant culture is clearly more “northern” in nature, one is still pushed to be an early riser, but to then do all work and social activity in a continuous chunk, no napping allowed. Because I sleep only 6 hours per night, after a few days I’ll be tired. The only option I have is to go to bed absurdly early, wake up after 6h, do some work in the middle of the night then go back to bed. That’s very antisocial in nature.
Yeah, i dont do that. I just used to be the case I had a night of less sleep, a stress and eventful day and then a night with a long sleep phase. I dont try to catch up or anything. I just slept this without a break (My Gear sport recorded no interruptions and very few restless phases).
Mid-day nap culture is still very much alive in really hot countries, where it's practically impossible to work mid-day.
But if I'm so engaged as to lose track of time, or simply undisciplined (just one more dose of political outrage/cute animal picture/epic fail schadenfreude on reddit!), the app closing down will help a lot.
I think you mean "reminded". Sorry, I don't mean to be rude.
When I spend a week camping, I don't need to be reminded, because it's really dark and there's been nothing exciting to do for a while. But at home, lights are bright and there are exciting things available at the push of a button. (Thanks to push notifications from things like the news and social media plus autoplay from video providers, they're often available without pushing a button at all.)
A few years back I build an automated lighting system that is sort of like Night Mode for my house: https://github.com/wpietri/sunrise
I thought the valuable part would be waking up early without the stress of an alarm. But actually the best part has been the auto-dimming in the evening. My brain get the hint. Instead of staying up and hacking on something or binge-watching TV, I am very likely to turn in early enough that I now get plenty of sleep.
I imagine you would need root on your phone to kill apps from another app.
If I open a task switcher, I want to see a lot of my open tasks. Not one, not three, not even four, but dozens.
How long am I supposed to be swiping if there's hundreds of open tasks?
And scrolling with the home button is even worse.
Let's hope Android Q quickly gets rid of this again.
Having both a zoomed in view and a grid view is standard in every single gallery and camera app, adding it to the recents menu shouldn’t be that hard.
It's a phone with a small as fuck screen, not a workstation with a big monitor. Why do you even have 'hundreds of open tasks' on a phone ?
Why should installing e.g. the facebook app, allow it to track my position at all times? Just freaky.
No, they won't willingly destroy a major revenue source.
Install something like Netguard (same dev behind Xprivacy) -- which functions as a firewall (vpn) b/w apps and the internet. Turn on logging. Watch all the apps you have installed phoning home and services like graph.facebook.com every few moments.
https://www.theverge.com/2018/5/7/17328758/amazon-alexa-defa...
I can’t even disable this Google bloatware on a stock phone (Nexus 5X)
Also, you've once again conflated "stock" with Google phone. If you use more precise terminology, you won't be so confused.
Yet I can still not disable these Google apps.
EDIT: Screenshot https://i.k8r.eu/4ZLdMA
If you were able to disable the apps and then uninstall the the new launcher, your phone would be unusable until you did a factory reset from recovery.
And they take up space in my launcher.
Please stop spreading FUD :(
you mean something like ios keychain? what does this do that can't be achieved using disk encryption + sandbox?
>Ad and js blocking capable browser.
firefox for android
Because once the phone is booted and unlocked by the user there is no further protection. Do you keep your phone turned off?
Also the lack of device pairing means it is usually easy to dump the encrypted image and crack the pass code quite quickly.
https://blog.elcomsoft.com/2017/10/ios-vs-android-physical-d...
> firefox for android
I find it quite slow, slower than chrome downloading all the ads. But it is an option I guess. It won't get wider uptake though, the default platform browser is the main game.
https://developer.android.com/distribute/best-practices/deve...
For most apps (unless I'm wrong?) there's no way to let the app use my location without it going through Google's location services and sending data back to the Borg Cube.
presto - no google - only actual GPS signals used. But good luck - you'll see just how terrible GPS is indoors and in urban canyons. Without wifi and BLE augmentation it truly sucks.
1) Using visible WiFi networks to determine location sends data to Google to use their geolocation database. Your suggestion here does stop this (assuming you meant to DE-select "always high accuracy" and instead select "device only" which uses the GPS without sending any data offboard.)
2) Google pushes the use of Google Play location services API over Android's location services system API. These services don't work unless you grant Google Play Services permission to use your location. Since Google Play Services also needs network access, you effectively have to grant Google access to your location in order to use any location services whatsoever in any application.
It's the second point to which I was referring in my post above.
Googles insistence on capturing as much data as possible about me is just creepy. I disabled voice search, location history, and everything else I could. Switching to DuckDuckGo made a big difference, but still the data gets swallowed up.
I know that Apple is only slightly better - but I’ll take whatever privacy I can get.
iOS has a fine-grained permission setting for this - you can decide whether to allow the app to access location or not when it's not in foreground, i.e "While Using the App".
Similarly, I have decided not to use many Dropbox apps that need R/W permission to entire Dropbox instead of any one folder that I can choose.
I think we kind of need to show these apps and services their place, so to speak.
> Encryption of specific secrets for each app
Not sure what you mean by this, but apps are already sandboxed and can't access each other's data. What specific threat are you trying to protect against?
> Fine grained control of what information is shared with apps
Already exists.
> Ability to selectively deny particular data (and have the app keep working, e.g by giving it an empty contacts list, fake phone ID, etc)
Already exists. Not with fake data though; you just deny the permission and the app has to deal with it.
> Ad and js blocking capable browser
There already are _several_ browsers on Android that do that. Unless you mean you want it in Chrome specifically?
> Why should installing e.g. the facebook app, allow it to track my position at all times?
It doesn't. That only occurs when you grant it the location permission.
This is the main reason I root my phone and use XPrivacy. After using XPrivacy, Android permission seems stupid. I can't go back to that.
* Bunch of permissions are grouped together into CATEGORIES and the app gets all if I allow that category.
You can update permissions on types of notifications, location, storage, contacts etc. all independently of each other. O is great for that and I assume P will only be better.
* Give an app access to only save, not read my photos * Give an app ability to read one specific photo (think upload a photo to facebook) * Take pictures without always having access to the camera.
That is fine grained.
You just choose to use apps which instead request you give them permissions to whole photo storage.
Saying that I can just choose to use apps that do it "the right way" is the same as saying "you can just not use a smart phone". I can see the value in smart phones and in popular apps. What I'm asking for it a pragmatic choice where the OS puts me in charge, because it's not working to just ask the developers to play nicely.
Ublock has been fantastic on Firefox for me, even though Firefox isn't as good as Chrome.
Smart quotes? Only exist to fuck up copypasta code and SQL statements.
Smart spell check? There's a field of internet hilarity about how bad this is.
Smart restore? Best way to lose my data and fuck up my install.
Smart assistant? I've yet to meet even one person who uses GAssistant/Siri/Cortana for anything more complex than "schedule a meeting" or "set an alarm."
Smart handwriting recognition? Refuses to detect the word "fuck". And my handwriting was perfectly legible anyway.
I want my software to be TOOLS. Predictable, ideally interoperable and chainable, but above all: simple. The best pieces of Ux make technology feel like a simple tool. Cut/paste. Messaging conversation-style, with a simple "send" button. Grep.
I find my best note taking software is a text editor, with all formatting turned off. I use ad blockers to disable as many recommendation engines as I can. I use duckduckgo because it just answers my query, without trying to tailor the answers to what the machine (or the advertiser) thinks I want. To mee, simple tools that behave consistently are infinitely preferable to the game of "guess how the computer is misunderstanding the request." Maybe I'm a curmudgeon.
I work with ML/AI, and there are great benefits in lots of applications. Just not in second guessing what people want to do. Even other PEOPLE are terrible at that.
Please don't make me dance around the quirks of your "smart" AI. I promise, I am smarter. I have better context. I know what I want to do. Just get out of the fucking way and let me do it in the simplest, most direct way.
Children. No really watch them use an iPhone.
Related to this, I recently found the option that controls whether the system from recommends or corrects to offensive words. It's buried under:
Settings > System > Languages & input > Virtual Keyboard > Gboard (more than likely for most) > Text correction > Block offensive words.
The same option exists for Google voice input. Wonder if there's something similar for your issue.
To what quantitative fraction is Android still free and open (as in speech), and how much of it has been moved into closed-source containers?
Since the G1, the following parts are just some of the many that were removed from the open project and replaced with proprietary ones:
Launcher, Dialer (the actual phone app), Contacts, Calendar, Email, on-device Search (last open in 2.3.7), and many more.
Everything that the user touches or sees is closed source.
The versions in AOSP are heavily cut down on functionality and barely even work.
Im the days of the G1, the actual Google Talk and Google Search apps were open, the Google Books app was open too!
That was the Android we bought into: one where a default install was entirely open, and living without any closed source apps was easily possible.
Today, that's long gone. Play Services here, SafetyNet there, Google®™© apps in the middle.
> Im the days of the G1, the actual Google Talk and Google Search apps were open, the Google Books app was open too!
Those apps as they were are still open (the Google Talk app was/is actually a multiprotocol messenger), though I don't remember a Google Books in AOSP. Even better, there are now superior equivalents to all of these that are also open source.
I was barely in middle school when the G1 came out, I saw Android morph from the promised solution as entirely open source platform to what we have today, where the platform and base apps are all proprietary.
The platform remains entirely open. That's how Chinese people are able to use Android even though Google does not do business there. You confuse the platform with the apps and services that run on it.
That’d be an argument, if Google removed all of those APIs that were in AOSP and forced people to use their proprietary solutions.
If you want to do WiFi Geolocation, you now need to use Google Play Services.
If you want to use shared library of openssl, you now need to use Google Play Services.
If you want to use background notifications, you now need to use Google Play Services.
This is an official Google diagram describing which parts of Android are open source, and which are proprietary: https://cdn.arstechnica.net/wp-content/uploads/2013/08/plays... That was 2013, since it’s gotten even far worse. This was 2016: https://www.googlewatchblog.de/wp-content/uploads/android-st... Since then, it’s gotten even worse.
By now you can’t access the stepcounter, or movement data, or most raw sensors, you can’t run in the background, you can’t access location (except for raw GPS, not even A-GPS) without proprietary libraries.
Hell, you can’t even use the latest OpenGL without proprietary libraries.
This was always using Google's proprietary data, even in the G1 days. They split it out into Play Services, so Chinese phones using another provider would be possible.
> If you want to use background notifications, you now need to use Google Play Services.
Background notifications weren't even possible on the G1 when it launched. When background notifications were first made available, they used a Google service and were never part of AOSP for that reason. Chinese phones use other push services inside the firewall.
> If you want to use shared library of openssl, you now need to use Google Play Services.
This is because Google now updates that library for security. If someone else updates the shared library (because you're in China, for example), you can call that version. Or you can just bundle the library in your app.
> By now you can’t access the stepcounter, or movement data.
None of those existed on the G1 and are simply sensor fusion implementations on the open sensor APIs.
> Hell, you can’t even use the latest OpenGL without proprietary libraries.
Yes, you can. https://developer.android.com/guide/topics/graphics/opengl
You need to understand the difference between the platform and the services that Google provides on top of it.
Incorrect, in the past your apps could access the raw data and feed it into Mozilla’s Geolocation APIs. This is not possible anymore (but returning in limited form with 9.0).
> Background notifications weren't even possible on the G1 when it launched. When background notifications were first made available, they used a Google service and were never part of AOSP for that reason. Chinese phones use other push services inside the firewall.
Incorrect. Until Android 7.0, apps could run in background and open their own sockets for background notifications.
> None of those existed on the G1 and are simply sensor fusion implementations on the open sensor APIs.
Incorrect – the open sensor APIs are only available for foreground apps. You can not collect any stepcounter data unless your app is constantly showing a notification in foreground, or is using the Google Fit APIs.
> Yes, you can. https://developer.android.com/guide/topics/graphics/opengl
Not on any devices, you’ll need to load bindings through Google Play Services.
> You need to understand the difference between the platform and the services that Google provides on top of it.
I seriously suggest you actually go and try to build apps for non-Google Play devices at least a single fricken time, or try to build ROMs, before you comment. You’re just making a fool out of yourself.
> I seriously suggest you actually go and try to build apps for non-Google Play devices at least a single fricken time, or try to build ROMs, before you comment.
I've built apps for the Amazon App Store and built AOSP from source multiple times. That's not the issue here. The issue is that you were confusing the fact that more of the things on Pixel devices are not open source (true, but Pixel doesn't claim to be an AOSP device) with fewer things being available in open source (which is false) and are now confusing fewer things being allowed to third party apps (true) with fewer things being available in open source (still false).
It never was open.
Sure, Google ups a bunch of source code to AOSP after the release of a new version but there is no actual OSS-style development going on. All development is behind closed doors, only done by Google and the public cannot contribute.
I cannot check out the latest development version of Android from github, fix some issues and send a PR. It's all a one-way street.
LOLWUT ? Apple is years ahead of Android.
With Android P they are doing the same to Java apps that make use of reflection to access non-public APIs or work around bugs due to lack of updates.
As far as I know, any company can pay to use it on their phones. I think members of the Open Handset Alliance don't have to pay. I'm not sure what it takes to join that, though.
Anyone can download Android Studio on any major operating system and make an android app. There are no limitations as to what features those apps can use. There is no fee to put an app in the app store. People can still sideload their own apps.
There's a difference between "I want to put a location in my status update" to "I want to upload my location 24/7 to Facebook's servers so they can profile the crap out of me".
iOS makes this difference. Android doesn't.
This doesn't address the comment you replied to in any way. Surely you see that there's a gap in reasoning between "I want to be able to add my location, just not have my location always broadcast to them" and "you can disable location entirely". Your suggestion specifically violates the "I want to be able to add my location" part of personal control over information.
Or are you suggesting that it's reasonable that a user should have to manually switch apps between white/black lists for behaviors? Personally I would find that idea contemptible.
They could both still be improved - it would be nice if you could do one-time permissions, like access the camera, but only this one time (So the permission would be removed when you lock the device or leave the app.)
https://developer.android.com/preview/behavior-changes#input...
"Android P strengthens privacy by limiting the ability of background apps to access user input and sensor data. If your app is running in the background on a device running Android P, the system applies the following restrictions to your app:
- Your app cannot access the microphone or camera.
- Sensors that use the continuous reporting mode, such as accelerometers and gyroscopes, don't receive events.
- Sensors that use the on-change or one-shot reporting modes don't receive events.
If your app needs to detect sensor events on devices running Android P, use a foreground service."
This is what I always thought Google should do. Operating systems have completely stagnated, but a truly new and differentiating set of features would be to deliver AI as a platform - not a Google-only feature but a set of open APIs that anybody can use or implement. I hope they continue down this track.