Hacker News new | past | comments | ask | show | jobs | submit login
Metro (daringfireball.net)
313 points by aaronbrethorst on Sept 14, 2011 | hide | past | web | favorite | 158 comments

If I'm reading this right, Gruber's effectively saying, "Metro, and Metro alone (i.e., no desktop), is a good competitor to the iPad." That's wonderful; a real iPad competitor.

But what about all the rest of the computers? You know, those big things on your desktop? We all know just how smart most users are when it comes to figuring out how to use stuff on a computer. Its taken years to get to the point where your typical "user" can use a computer without too much difficulty. Part of this is due to the consistency of the Windows UI; folder windows, mouse motions, menubars, etc. Suddenly, we're going to throw something completely new at them, something which:

1) Has been shown to be difficult to do well for everyday use (touchscreen desktops)

2) Has a moderately non-intuitive interface (hidden UI elements until I swipe from a particular side, or maybe tap over here and here and here)

3) Has questionable benefits in a desktop computing environment where a keyboard is a perfectly appropriate device

And, once they learn all that, our user realizes that he still has to use the original UI for many programs! That's right, photoshop isn't going to be going Metro anytime soon, and neither is Matlab, or Autocad, or Excel, or video editing software, or any "Pro" program (for lack of a better word). Metro may be wonderful for tablets and other mobile devices, but it sure looks like it's going to be a drag in being forcibly married to the traditional Windows UI.

I suspect we'll be seeing a more and more pronounced split in computing and, therefore, interfaces. On one side (the majority) will be "consumers": web, email, Facebook, maybe some light photo editing. On the other side will be the "producers": people who use Photoshop, or Matlab, or AutoCAD. Office falls somewhere in the middle... Office itself may eventually be divided along those lines, with a lightweight, tablet-friendly spinoff.

So in many cases I think you're actually talking about different users: the majority of people may eventually not need anything beyond a tablet. I also suspect the rest of us will be annoyed and inconvenienced by this trend, since it won't be aimed at us.

I don't like these ideas Gruber and friends are repeating and repeating again. In the article, it is carved in marble: "the advantages that come from NOT allowing you to do so many [things]". The reasoning behind is that Steve Jobs or other said genius should be choosing for you what you are "allowed" to do.

It feels awfully 1984 for me.

I live in China and I can tell you it is painfull, sometime, when someone else chooses for you what you are allowed to do. I'm reading "Zen and the Art of Motorcycle Maintenance", and heck, this guy is right, you have those losers who fear and get used by technology, and the other guys who don't fear and use technology.

The cynic (borderline totalitarian) approach is to say that people should be divided between those clever hacker who have control and the rest, which should not be given "too much" control, because their are not clever enough, or don't need it, or what.

The humanist (democratic, cf "Enlightment") is the opposite: human being are clever, and should be given control on their life, on their tools, should be given choices, etc.

It seems to me that too many people have forgotten the we gather here, on HN, on the free, interesting and open parts of the Internet, because we like to control our shit, and ALSO we prefer the common people to be educated and control their shit. Especially when it come to so personal matters as their own files.

Sorry for the rant. I hope I'm mistaking, but from my recent reading of HN I feel there is a progressive shift toward more "Gruberness" which is questionable in its hypothesis and conclusions.

There's a pretty significant difference in kind between design constraints and political or economic constraints. The "freedom" to mistype a URL and then get phished, or to misconfigure a backup drive and then lose all of your data, or to save a file inside an arbitrary folder with an arbitrary name on an arbitrary drive and then be unable to find it again, is very different from the freedoms of the Enlightenment.

Computer owners who don't understand their computers do not feel free. They feel a constant sense of discomfort, even dread. They fear making mistakes, they worry that the box will break at any moment, and they're embarrassed that they have to ask techies -- often expensive techies -- for help.

There are equally as many people that run rough shod over everything because they'll just ask their techie friend/family member to fix whatever broke for them.

Whether one falls into this camp, or the camp that you're talking about is probably just up to personality types.

Inappropriately applied dichotomy, my brotha. "Figure it out yourself and lead" vs. "Manipulate someone else into doing it for me" ---- a lot of people would rather do it themselves if it were easy enough.

Maybe, but I was talking about the fear of screwing up that the parent was talking about. Some people have no fear of screwing up, even when doing things that probably should make them at least cautious, because they have the view that someone else will fix it for them.

If you cast the app universe as 'open' vs' 'closed' as some have tried to do or 'free' vs. 'controlled,' then yes, of course, open and free are better.

But there are a few other ways to look at it:

- Chaotic vs. orderly. - A pile of laundry vs. neatly folded - Unfiltered vs. filtered - Amateur vs. professional - Unedited vs. curated

Pick whichever way you want and your perspective might be different. It's really in how you define the context.

As far as I'm concerned the app store is a shopping mall and the proprietors have chosen to limit what kind of businesses can open shop there (no head shops or brothels). There are shoppers who are happy to go there and those who don't like the brand-name selections. There are businesses who are prospering happily and those who are struggling.

And then there are those who like to stand outside with pickets and complain that the management is squelching free speech.

Comparing it to dystopian nightmares and totalitarian regimes is quite a stretch.

I am ok with the idea that on a given app market, website, content provider, the proprietors can and should curate the content the way they decide.

What I am not ok with is that the makers of a machine I buy keep a too strict control on what I can do with this machine I bought. We are grownups, we should be allowed to use the things we buy in the way we please. As far as I know, if I do something illegal with a Porsche or an iPad, their respective makers are not responsible of that misbehavior, and therefore should not be entitled to forbid me from doing it.

> As far as I know, if I do something illegal with a Porsche or an iPad, their respective makers are not responsible of that misbehavior, and therefore should not be entitled to forbid me from doing it.

You are correct, but I am unclear as to what you could be referencing. You can add pirated MP3s, eBooks and movies to iOS devices, if you so choose. You can even jailbreak them to customise them even further, provided you don't also expect warranty service.

>You can even jailbreak them

Not if Apple had their way and were able to win that arms race.

>I am unclear as to what you could be referencing.

You're _playing_ the fool here I hope.

I put rockbox on my iPod classic, I love rockbox. Apple encrypted the device in a way to prevent aftermarket modifications and set development for that device back years while they worked to get around the restrictions apple put in.

That kind of practice is what he is referencing.

My understanding is that Apple is able to do whatever they want to their own products. They are under no obligation to support jailbreaks, Rockbox or other modifications. Encrypting the iPod has no noticeable downside for the average user, but for a hacker, it's a pain. But Apple is under no obligation to support an alternative firmware.

Making a product do something the standard setup will not allow is the whole point of the hacker space anyway, isn't it?

PS: I don't agree with Apple's decision to encrypt the standard (non-touch) iPod line. I wonder if it was part of their licensing terms with movie studios.

Yes, precisely.

It's kind of nutty to equate technological simplification and totalitarianism. Some people don't want to solder their own computers together or inspect their own meat.

We need, occasionally, to draw a line of abstraction over complicated stuff so we can build upwards. This is why we code in high level languages rather than toggling switches on front panels, even though it meant surrendering some "freedom".

Yes, layers of abstraction can be useful. I don't see how the iOS restrictions are a layer of abstraction, though.

And selling pre-built computers doesn't mean you have to glue the box shut. You just have to make sure people don't have to mess or even think about the internals unless they really want to.

Some iOS restrictions are obviously simply business driven, but fewer than you might think.

E.g. The "give us 30% and don't undercut" rule means that the user can buy from the app store without needing to price compare. The "no execution of outside code" rule means no platform subversion with the endless security issues this would bring.

If you stop thinking in terms of technical capabilities and start thinking in terms of the user's cognitive load, iOS is a huge step forward in terms of abstraction.

I don't like it either (hell, most of us here on HN won't like it... we love to tinker and break things and get under the hood), but that doesn't mean it's not the most likely direction.

You describe it as a struggle between totalitarian and democratic ideologies, with all the moral and ethical implications. I think it's really about a third, much more powerful force: giving people what they want.

This is tangential, but I think explains what I mean better than I could:


Thanks for the link. I am no historian but I have the impression that Human history would show that Orwell's fears are more grounded, as the totalitarian world he described has actually been implemented for real in many occurences, far away in the past (maybe under Pharaos? certainly under Qinshihuang) and closer (USSR, PRC). It is possible that Huxley's fears are a bit less likely to happen for real. Maybe Roman empire had a pinch of it (panem et circenses) but even then roman people (the free ones) had their freedom, and some used it to write books, create things, etc.

However I am extremely opposed to entertainment feeding, I don't watch TV, nor do I read magazines, so I agree with Huxley (and you?). But I don't think it is closely related to my impression that some people go too far in the "don't say yes to all" direction.

The good reason for tool creators to say no is not because they know better how the tool should be used, it is because they know that saying yes to request A would forbid the tool to be used in ways B and C, and other unprospected ways. Saying no is ok when it actually opens new doors.

>I think it's really about a third, much more powerful force: giving people what they want.

What people want depends, to a large degree, on their level of education, what they're exposed to, and what leaders and the media tell them they should want. It will be interesting seeing how pop computing and lock-down devices do globally in the long-term.

I'm not saying this is the case, but I think maybe the difficulty the tech crowd has with these new "easy" interfaces is exactly that, they are easy. If everybody can easily use a computer (tablet, phone etc), never looses their files, never picks up a nasty virus, then they don't need us any more.... We lose our exalted position of power as one of the enlightened who "know".

Don't get me wrong. I regularly drop back to the command prompt, and hate having the flexibility of a full system taken away from me. But I also hate that feeling I get when random relative calls up to ask for help with their latest tech disaster. I hate it that configuring operating systems so that they are reliable, robust and functional is almost impossible. I hate it that the operating systems still get in the way of just doing the job.

We're hackers, hobbyists and enthusiasts, we want the computer to be in our faces, that's what we do. The vast majority of users however, hate it when they have to put up with the crap we think is cool just to read an email, or write a document.

Can't remember who the quote is from but it goes something like this, Technology goes through three stages, simple with limited functionality, complex and unreliable, simple and reliable....

I have a feeling that we have started to reach the end of the second stage.

  > "the advantages that come from NOT allowing you
  > to do so many [things]".
I read this as referring to the examples he talks about later with regards to memory management and restricting the ability to background to preserve battery life. The idea being that tablets -- as devices -- a restricted in their resources, so the environment that the programs run in also have to be restrictive in allocating those resources. The bulk of current Windows programs are not designed to restrict their resource usage because they were never targeting such devices.

I don't think that the he was talking about a curated AppStore (with a 'totalitarian' gatekeeper), which seems to be what you're talking about.

I was more talking about a general trend in design and development that is, I think, based on a misunderstanding of Apple's success. The iPod was a success because it managed to break the constraints of a very small interface giving access to a hudge set of media files, and made something usable out of it. I don't own a Mac but I guess a fair part of MacOS success is owed to the BSD backend, and the terminal, which allowed to do real pro developpper work on it. Terminal, command line and scripting is hard for beginners, but it gives a lot of power.

Now we are told that iPad don't allow background process and that it is a good thing. If the only reason is battery life, then it is not good enough a reason for me (I have two spare Nexus batteries in my pocket, what is the problem?). Somewhere else we are told that we should not expect direct access to our own files because "files are so 1990", or because iTune will handle them. I don't agree with that view. There ARE files behind the pages or handles or icons I see on my screens, and they are MY files (pictures, docs, etc.) It is ok that the main use of this content is done through some cloudy vapored tool, but I need the possibility to grab the file, MY file, and make what I want with it (including, obviously, destroying it).

If you had final authority over the design of a consumer electronics device, and the engineers told you "we will have to ship this with a battery and two spares, and users will have to change batteries twice daily," would you say "That's good enough for GBOG Devices; start the factories"?

Constraints are important for innovation. By "not allowing" you do things, new ideas and solutions bubble up. And the market decides if/how much a particular product succeeds. Many devs on this board own both a full computer and an iPad. I do, and enjoy both! Glad to see Metro moving forward and competing with iOS.

I feel like if I was reading a Gruber article from 1995 he would tell me that I'd have to choose either a responsive, efficient system or a system capable of running with pre-emptive multitasking, but not both.

He is so incredibly apologetic to the corners that Apple engineers cut that I wonder if he really believes himself.

That seems to be in line with Steve Jobs' vision of the future, in which PCs become "trucks" while average people drive "cars." I do see early but pronounced signs of that split.

And Windows 8 is trying to be the SUV here.

Or a member of the more en vogue crossover family?

I own a crossover. And with the newness that is Metro, yours is a better analogy.

Thanks. Does a crossover give (most of) the benefits of an SUV without the bad gas mileage and environmentally unfriendly vibe?

I'm not power productivity suite user but I would have to believe close to all of Office's features can be implemented with the proper UI with the advanced features of the apps hidden away from the eyes of the average user.

If it's a question of specs or software maturity then it really is just a matter of time.

It is at this point that I feel obliged to point out that Apple already ships an office suite for the iPad and, bizarrely enough, even the iPhone: http://www.apple.com/apps/keynote/ - http://www.apple.com/apps/pages/ - http://www.apple.com/apps/numbers/

I enjoy these apps but they're certainly not on the level of Office or even iWork for the Mac. I think it's a good start of where we should be going and they've done well in creating a consistent UI across the iPad and iPhone.

I recently had to write a one-page letter on my iPhone and email it to somebody to print out, sign and hand to a third-party. It wasn't by choice, it was just an urgent request and I had nothing else with me.

Surprisingly, it wasn't nearly as painful as I expected. It didn't take me long to write the letter (the default templates helped a lot), making edits was fairly easy and I had no trouble being able to read over what I had written.

So for small edits to existing files and quick docs it's usable, though clearly not as good an experience as using the same suite on an iPad.

Good point. Still, this relies heavily on the use of palettes, which we haven't seen in Metro. Maybe they'll have some UI elements like that in Metro, I don't know.

>Office itself may eventually be divided along those lines, with a lightweight, tablet-friendly spinoff

It'd be interesting to see how such a thing would differ from Google Docs, and if it would justify the kind of prices MS likes to charge for Office. An Office app for lightly-spec'd computers makes sense, but you'd have to charge app prices for it and I don't see MS sacrificing their second biggest cash cow any time soon.

How about an offline version? That's something that Google doesn't really offer.

More and more these days I find a computer without Internet to be increasingly useless. Google Docs' value proposition is irrevocably tied to its online status. It offers features that a strictly offline office suite does not. For Microsoft to be able to sell an offline Office, on a mobile, 3g-integrated computer, for Office prices, would require some functionality that they haven't come close to implementing. You would have to make some argument as to why your offline Office Light (lite for you colonials who insist upon spelling that corresponds to pronunciation) is better than the free, connected office suite Google offers.

At the moment, Google Docs is Office Light. It offers a limited subset of Office's features, and doesn't do it particularly well. What it does do, extremely well, is make it easy to get the documents you create to your friends and co-workers. I can't see a limited-functionality, offline version of Office competing with that.

It won't be long before most people have only one computing device: a very powerful smartphone. When used as a mobile device, the interface looks like iPhone/Metro. When you are at your desk, you dock it and the interface that pops up on your 30 inch monitor is like Windows 7 (WIMP). And you can also dock it in a range of shells (for lack of a better word) from a 7-inch tablet (touchscreen UI) to a 17-inch laptop (WIMP UI). All these devices you are docking to are just relatively cheap shells. All your data and processing power is on the smartphone. In fact, even the smartphone will be just a shell, and the computer will be like a large SD card.

I actually prefer the more logical conclusion to this thread: Sun's SunRay concept... everything is in the network (cloud), you use key(s) to access/control them.

However, I'd extend that and say that along with the "key" is a smartphone that can also (in a limited fashion) view/manipulate your cloud data and send/receive messages.

The important thing to recognize is that there will always be need for "more and faster" than a smartphone will be capable of providing.

re: your last sentence -- agree, but I suspect most people will not need that much power

2 things will decide when/whether this is possible:

1) The rate at which cooling tech increases; How can you keep a miniature multicore cpu cool enough to process gigantic displays?

2) Power efficiency of the miniature multicore cpus.

That's right, photoshop isn't going to be going Metro anytime soon, and neither is Matlab, or Autocad, or Excel, or video editing software, or any "Pro" program (for lack of a better word).

I'd actually be surprised if there isn't a fully staffed team working on a Metro version of the Office suite.

But what would it look like? Consider Powerpoint. There needs to be a way to switch slide layout, add elements to the slide, work with animations & transitions, reorder slides, etc. And that's just some of the things that users will want to do all the time, on almost every slide. There's just no way to have that sort of rich interface in Metro. Heck, there's not even a menubar.

Remember, we're talking about regular users, here... these people (broadly speaking) never use keyboard shortcuts. Everything needs to be clickable (touchable) with the mouse (finger).

Oh, it's definitely possible - Keynote for iPad (http://www.apple.com/ipad/from-the-app-store/keynote.html). No a menu bar in sight (at least, not in the traditional sense).

> But what would it look like? Consider Powerpoint.


> There needs to be a way to switch slide layout, add elements to the slide, work with animations & transitions, reorder slides, etc. And that's just some of the things that users will want to do all the time, on almost every slide. There's just no way to have that sort of rich interface in Metro.

See link. It's perfectly possible to do all that without a classical menu bar. Does it do everything the desktop version does? No. Can it do pretty much everything useful of it? Sure.

>> Can it do pretty much everything useful of it?

No, not really. It's great for presenting, useful for making small tweaks to existing content, and painful for trying to create presentations that go beyond stuffing text in the provided templates.

This is also true of Numbers. Pages is pretty good though.

There's just no way to have that sort of rich interface in Metro.

You must've missed the "charm", sliding & docking side-windows part of the Metro demos at BUILD.

I'd hate to use this on a 24" display, have icons the size of my hand and using a mouse (or "Magic Touchpad" like device) to swipe somehow.

Thinking about it, isn’t window an odd word choice for what we call movable, stackable, resizeable content regions in a user interface? Other than being rectangular they’re not like real-world windows at all.

I can remember my high school computer studies teacher explaining they're called windows because they're split into panes. Microsoft Windows versions 1 and 2 didn't actually have stackable overlapping windows (you'd split the screen up into as many panes as you needed) so I suspect the name made sense then and has since stuck.

Good point - though I believe they were called windows on previous GUI systems.

EDIT: Though now that I think about it and, IIRC, Xerox's stuff didn't have overlapping windows either. Apple's Lisa and Mac did. I think Englebart's demo had windows but I'm not sure if they overlapped.

I recall a story (maybe an urban myth) about Steve Jobs visiting Xerox's labs and seeing their UI in action and thinking "we've got to have that". He went back to tell Woz all about it, telling him that Xerox's UI supported overlapping windows and that Apple needed to do the same.

Woz was amazed that it was possible to have overlapping windows on such a resource constrained machine. But if Xerox engineers could do it, he could to dammit! After a Herculean effort he got the overlapping windows to work nicely and they shipped.

It was only later that they found out that Xerox's engineers were blown away by Woz's achievement as they had studied the problem and decided that it wouldn't be possible to do overlapping windows. They never developed that functionality.

Turns out that Jobs had made a mistake in remembering what he had seen and Xerox never did have those overlapping windows :-)

Anyway, this story might be incorrect on some points or a complete fabrication, but I like it.

Not Woz, Bill Atkinson.


"Smalltalk didn't even have self-repairing windows - you had to click in them to get them to repaint, and programs couldn't draw into partially obscured windows. Bill Atkinson did not know this, so he invented regions as the basis of QuickDraw and the Window Manager so that he could quickly draw in covered windows and repaint portions of windows brought to the front."

The 'regions' used in QuickDraw are an amazing (for their time) data structure, a testament to the genius of Bill Atkinson. I strongly recommend reading the patent to see how cool and useful they were to the original Mac windowing system (http://patentdetails.patenttools.com/patent/04622545?searchi... click on the 'download pdf' link in the top left).

Steve Jobs know how important regions were: http://www.folklore.org/StoryView.py?project=Macintosh&s...

Obviously I meant 'Steve Jobs knew how important regions were'.

In a semi-related matter, does anyone know of the algorithm Bill used to draw RoundRects 'blisteringly fast'? http://www.folklore.org/StoryView.py?project=Macintosh&s...

I've been intending to write a blog post retrospective on the workings and algorithms of the quick draw library, but regions seemed too big a topic to start with. You've given me an idea though: I'll have have a go at RoundRects first and see if there is any interest.

Ah, excellent stuff. Thanks for the correction and digging that up.

It's a good story.

According to the Apple stories I've read, Wozniak had little to no involvement with the creation of Macintosh. He was in a plane crash in 1981. When he came back to the company in 1983 he was mostly involved with the (increasingly sidelined) Apple II team.

But overlapping windows first came to Apple with the Lisa, anyhow. Some of Bill Atkinson's photos here show them from December 1979 or early 1980. Story says Bill wasn't sure whether they predated Apple's Xerox visit, or not. http://www.folklore.org/StoryView.py?story=Busy_Being_Born.t...

Anyhow, it's a good story. If it has some basis in truth it might be with s/Woz/Atkinson/ (he was their UI/graphics guru programmer) and s/Macintosh/Lisa/. :)

Indeed, and due to various miracles, the source for QuickDraw and MacPaint is at:


He didn't actually mis-remember, though. The Alto, the OS used internally at Sun had overlapping windows but Xerox Star, the version that they sold outside the company couldn't handle them. http://www.osnews.com/story/20365/pt_X_the_Window

FWIW the Xerox Alto (1979) called them windows, even though you're right that they didn't overlap like the Mac/Lisa ones did. http://history-computer.com/Library/AltoUsersHandbook.pdf

EDIT: The Alto could run SmallTalk as one of its environments, though, and its windows overlapped as of SmallTalk-72 or -76 http://www.esug.org/data/HistoricalDocuments/SmalltalkHistor...

This makes a lot of sense. Emacs calls the OS-level-thing-to-draw-on a "frame", and then you have non-overlapping "windows" inside the frame.

Yes, but that's because Emacs' windows were invented for the console. So they had to use a different word for what they call "frames" later on.

I've heard this explanation, too. Also, I don't find the 'metaphoric' window all that big a stretch.

So similar to tiling window managers you'd find on GNU/Linux I guess.

Finally a Gruber post I haven't felt forced to flag. A pretty good analysis, excellent (as usual writing), interesting testable predictions plus just enough Apple flag-waving so you don't forget which religion he follows.

edit reading through the comments shows a commensurate level of better discussion than the usual Gruber response as well.

Err, he's wrong on ARM PCs not running desktop Windows, I don't see why this wrong post needs to be on top of HN. It's like there's like zombies that seem to have a need to auto upvote any post by Gruber. Maybe they're similarly misinformed.

Yes. Clearly someone is a zombie if they upvote a John Gruber article that contains one false speculation. It can't be that they found the rest of the article appealing.

Metro is really nice. It's great to see Gruber praising it like this: it is very praiseworthy. Also, it's easy to forget that Gruber is first-and-foremost a design geek.

I thought he was an Apple shill first and foremost.

Nope. You should read his rants about Helvetica…

But those aren't really from a design perspective either. There's little about what makes 'his' font work better, just potshots at the 'other' fonts that suck because they don't copy his closely enough (except where they do, then they're ripoff assholes).

He's a sportswriter, and his home teams are Apple, the Yankees, and Helvetica.

Rooting for dominant winners like that isn't so bad in itself, but he does it as if they were still downtrodden losers in need of defenders — that's the core of what makes him so insufferable

From the article:

"you can’t give iOS apps even the option to run continuously in the background without sacrificing battery life and foreground app performance. But that’s how Microsoft has positioned Metro for tablets — a modern touch interface that carries the full CPU and RAM consumption of Windows as we know it. That have-your-cake-and-eat-it-too attitude is what I didn’t get with Microsoft’s positioning Metro as its answer to the iPad."

This is wrong. From Anandtech: (http://www.anandtech.com/show/4771/microsoft-build-windows-8...)

"discarded applications will continue to stay open as a background application, having all of their memory pages intact but unable to schedule CPU time so long as they’re a background application. They’ll remain in this state until the OS decides to evict them, at which point they need to be able to gracefully shut down and resume when the user re-launches the application. Internally Microsoft calls this freezing and rehydrating an application."

Metro's approach sounds very similar to that of iOS and Android. Presumably this behavior will be adjustable so that background processes can be allowed on desktops without mobile power constraints. This is actually a really smart way to do things. Make how the OS handles background apps a setting rather than hard-coded architecture. e.g. If you're out and about using your tablet background apps get quashed so that you get decent battery life. When you go home and plug it into a dock you can leave a torrent downloading in the background while you browse the web or play games. Best of both worlds.

To me, "you can’t give iOS apps even the option to run continuously" is precisely the same as "discarded applications will continue to stay open [...] but unable to schedule CPU time" indeed.

At no point the article says Metro would not have such a feature. The point is merely that you can't have outstanding energy management with runaway background processes, which is the way desktop apps work today, so having a full-blown desktop running legacy apps on a mobile experience is compromising power management.

Side note regarding "you can leave a torrent downloading in the background": see task completion API on iOS, which fills the use case while eliminating undue battery use.

Which proves he didn't even bother to watch the keynote. There was even a piece of hardware that showed the power consumption during sleep. They also showed the new task manager and specifically pointed out how the Metro apps have been suspended (using 0% CPU, but still loaded in RAM).

He later goes on to suggest that perhaps there will be no "traditional" Windows on the tablet, only Metro, which he discusses having the same sorts of backgrounding-and-suspending capability iOS enjoys. From the article:

It’s worth noting that Metro is more than just a new look, and more than just putting touch first. Metro apps have similar restrictions to iOS apps. According to Jensen Harris, for example, Metro apps will get “about five seconds” after they’re no longer on-screen before the system puts them into a suspended state. There’s no file manager. Users no longer quit (or, in Windows parlance, exit) apps explicitly. These tradeoffs sound familiar?

...Did you (and the parent poster) simply read the Gruber article until you could find something to nitpick, then go with it?

"He later goes on to suggest that perhaps there will be no "traditional" Windows on the tablet, only Metro, which he discusses having the same sorts of backgrounding-and-suspending capability iOS enjoys."

Which, again, proves he didn't watch the keynote. At one point they brought up the scenario of using a tablet on the couch and then, after finding a bug, being able to pull up Visual Studio on the same device. Even the tablet they gave out came with VS installed on the device.

It's an Intel tablet. Gruber envisions a possible scenario where Metro-only is on ARM tablets.

That's a very narrow definition of a tablet. Regardless, they also showed at a different conference the actual Windows desktop with Office running on ARM. If it can run the full desktop it can Metro as well.

Does that apply to everything including Office and all that stuff?

Metro to me, and I hate to say it, reminds me of MS Bob, and frankly of the original Windows 3.1 except where in that case it was a layer on DOS, this seems like a gloss on Windows 7. I hope for the sake of innovation I'm wrong, but it is frankly more bizarre than Launchpad (and that's saying something).

Let me show: https://skitch.com/ethank/f3drj/windows-8-x64-preview

then https://skitch.com/ethank/f3drc/windows-8-x64-preview

and... https://skitch.com/ethank/f3dr1/windows-8-x64-preview

after you go back to metro: https://skitch.com/ethank/f3dri/windows-8-x64-preview

dumped back into Metro with no context how you got out.

Thus: feels like a skin or window manager, not the OS.

"like a gloss on Windows 7" - From what I've read[1] so far this is not the case. The APIs that 'Metro style' apps use are built deep into Windows. In fact it seems more like the opposite - legacy Windows is treated more like another app that you can switch to, if you're on x86/x64 machine.

[1] http://www.winsupersite.com/article/windows8/windows-8-devel...

I'm speaking from the beta. The act of going between a "traditional" looking UI and Metro and what is available in each is very disjointed.

And this is why kudos must be given to Microsoft for not making the distinction between "Developer Preview" and "Beta" more noticeable. Wasn't one of the reasons behind Vista's huge flop being that in the betas, driver support was incredibly lacklustre?

It's not a direct parallel, but this and the fact that companies like Google have completely upturned the definition of "beta", or well, pre-releases in general, seems to show how people come to their final judgement very easily.

I suppose that could be why Apple doesn't release things until they're well and truly ready for the limelight.

It's not a beta!

Yeah, exactly. It's alpha, and it feels that way. That the dev preview doesn't resume state well when you go from Metro -> Desktop -> Metro didn't surprise me in the slightest. I'm sure it's one of a litany of features and fixes they're planning on.

Win32 and Metro are both User Mode subsystems running on the same MinWin NT Kernel and you flip flop back and forth between them based on what you're doing.

Hitting the windows key brings up the start screen. It's supposed to do that, kind of like how the iPhone always takes you to the first page of apps.

Task switching is done by swiping on from the left or hitting Win+Tab. I just tried it, Windows+Tab brings you right back in the control panel where you left.

It could probably be clearer, but the Windows button takes you to the Windows front page, it's not a toggle between desktop and metro.

I think the act of flipping between them is not meant to be a common event. But I do agree they do need to clean up the edges, although I don't think its "critical" for v1 of Metro.

I disagree actually. I think cohesiveness should be a pre-requisite for the release of this.

Achieving cohesiveness in such a system will be very difficult though. There's a reason why Apple chose to keep OS X and iOS separate.

I do agree that cohesiveness is important to the user experience. It makes me a little disappointed that the extra step wasn't taken to make the Metro UI live on its own without the full desktop experience tagging along.

I don't think cohesiveness across these systems needs to be a prereq. In classic mode things should work great. In Metro mode things should work great. Bumpiness in transition, IMO, is acceptable for v1.

To put it another way, I'd much rather have bumpiness shipped on day X rather than a delay to clear up all the transition bumpiness and ship 6 months after X. Although I'd like to see updates that fix the transitions come out a 8-12 months after release.

Then why even put Metro on if they are striving for only half a level of perfection? Why not wait until they've got it solid before releasing?

What would make me very pleased is if they delayed 6 months to get it right rather than rush it out.

Why do you keep pushing on the idea that this is some sort of finished release of Windows 8? It's a developer preview, that's alpha stage at best.

>I think Metro will only run alongside the traditional Windows desktop on Intel PCs. On ARM devices, there will only be Metro.

It seems there will be a desktop mode for ARM tablets if you look 1:40 into the video here:


The desktop mode doesn't seem to immediately respond to touch like Metro. It will be interesting to see the final product and how well Windows 8 performs on an ARM processor in desktop mode.

Is that really hot off the press, or did Gruber just miss this?

Gruber really missed this and also how well Windows 8 runs on ARM in the article yesterday. Maybe he should stick to Apple stories where he's much more knowledgeable.

And yet his articles on Windows 8 get top billing on HN whereas the actual Windows 8 posts hardly got many votes. That's HN for you I guess.

> Tradeoffs. Mutually exclusive tradeoffs. Separate devices are required. ... And you can’t give iOS apps even the option to run continuously in the background without sacrificing battery life and foreground app performance

This is a typical Gruberism of false logic. He starts with the notion that all Apple's decisions are holy and right. He then derives conclusions by extrapolating from that.

There is no reason that tablets cannot achieve fantastic battery life while running background processes. They certainly need to be designed to achieve it. Existing Android tablets support background processes and multitasking and get comparable battery life to the iPad - typically we're talking a sacrifice of < 10% battery life to achieve an incredible expansion of utility. And this is not taking into account the fact that Android is a less efficient OS overall (utilizing less hardware acceleration, running most tasks inside a Dalvik VM instead of native, etc.) Even the iPad itself evidently does background processing as you can have it play music, give you calendar reminders and all kinds of other things happen in the background even when it is in sleep state.

Compared to desktop operating systems Android does very limited multitasking or, to put it in another way, it is cleverer about multitasking (iOS, too).

Remaining backwards compatible while picking up such a new model of multitasking is hard. Very hard. Apple made parts of it opt-in with Lion but that’s not the same as requiring everyone to adhere to the model and today hardly any app supports it.

Gruber simply sees this as a hard challenge that is, at least currently, not solvable if Windows is to stay backwards compatible. That’s all. He might be wrong but I think that position is quite sensible.

I think metro is very minimal and very clean. But it's also... cold. A bit too austere. The goal for an interface isn't to impress--it's to connect. Somehow, I've never felt a sense of connection with the Metro interfaces. It feels like design borrowed from the annual report of a faceless international corporation. All design; no personality. I like simple and clean. And I like minimalism. But I also think you have to be careful with it or you'll end up with something soulless. That, I think, is what has always subtly bothered me about Metro.

Metro is actually meant to be "soulless" by design. It's supposed to be distinctly digital. The idea is that your "connections" come from your actual social web, not from rendered wood textured bookshelves. At least, that's the idea.

That is why Apple is using fake wood backgrounds all over the place? I think they are very cheesy - prefer a "cold" interface, like a command line. Then again, maybe it would cheer me up to configure my shell to use Comic Sans...

It is beyond me how Apple, most of the time, designs minimalist products and interfaces, and suddenly comes cup with really kitsch stuff (iCal, Address Book, Photobooth in fullscreen mode, Time Machine). It's like Dr Jekyll and Mr Hyde.

Because it’s different people doing it.

I have long had the theory that the people at the top who could have done something against that (notably Steve Jobs) simply don’t care about that discrepancy, that’s why it can continue.

As compared to what? A traditional desktop like interface with application icons scattered over the background? As in... iOS?

One would think that the live tiles with continuously updated information show more "connection" with the user.

From a purely aesthetic point of view, I kind of feel the same as toddmorey. I find it too cold, too precise, too machine-like. It's like looking at a 3D rendition of a face versus an actual picture. A little bit of roundness here and there, subtle transparency, or at least shades of color could help.

Amen. This is probably coming from people who think that a computer which basically resembles a thin slab of aluminum is not too "cold". It's basically minimalist which is what Apple strives for these days and does a good job with. The really interesting design choices in Metro (to me) will be apps with a lot of text ala Outlook.

Indeed, it's essentially a live dashboard with a built-in program launcher. It shouldn't be the thing you connect to.

It's just a browsable report of the status of things you care about, with the UI more trying to get out of the way, and provide options, than be a "thing" in your world by itself.

On that note, It probably should (and perhaps will) become more subtle over time. I'd prefer it taking the same approach as classic mac os: stay in black & white so that the user's real data is in color. They won't do that, of course, but a subdued palette could really help. Of course, then the individual widgets need to really bring home their content's presentation.

Hotlinking means bandwidth theft, bru.

To be fair, that's probably there mostly to prevent other sites from embedding those images rather than to prevent linking from forum posts.

This is why Metro will make or break Windows. This is where the line is drawn in the sand. Since ARM is the future together with touch based devices, Microsoft will have to move over to ARM. The problem is they are very vulnerable on ARM, because they have no apps there, and wanting millions of apps on a tablet was kind of the whole reason you'd want Windows on a tablet.

But it won't work, because those apps won't be available on ARM, and even if they were, they wouldn't be designed for touch. So Microsoft is starting from scratch, and this time they have strong competition from both iOS and Android.

In this market, their Windows dominance doesn't matter as much, so they are on equal footing with the others. And I find that very exciting. If you notice, Microsoft is innovating only when it's the underdog in some way, not really when it dominates.

So I hope 5 years from now we'll get to see iOS, Android and Windows with about equal market share each for "personal computing devices", whatever that means 5 years from now.

Keep in mind a lot of .net applications will run on x86 or arm with no effort by the developer, ie no "simply recompile your application" either, it will just run on either platform.

But you're right that they're exposed with no applications, which is why they're basically giving windows 8 beta/vs11express away to any and all who are prepared to download it, they need to catch up, and quickly

Depending on how good a job they did moving their applications over to CLR languages, it wouldn't be that hard to port to ARM....in theory.

There will be ARM laptops, not just tablets, and they will run Microsoft Office in classic mode or nobody will buy them. Office on ARM has already been demoed on stage.

Contra-Gruber I'm betting ARM Windows 8 devices will have some mode that looks and feels just like Windows desktop. It may not be the default and it may be a dumbed down version of it but I can't imagine them shipping without it.

I believe it's already confirmed that Win8 ARM devices will indeed be Metro-only - no legacy Desktop whatsoever.

I don't think that's true. Can you point to a source? I've heard and seen only the opposite.

The impression I've gotten from all their presentations about MinWin and Metro is that Metro and Win32 are independent subsystems and that the Win32 subsystem is not being ported/supported on ARM thus no Win32 applications will run on the ARM based Win8 systems. I can't find where I read it but I remember seeing something about the Win32 subsystem possibly not even being installed by default on systems and only being installed if you tried to install a Win32 app.


"Furthermore, applications for the ARM version of Windows 8 will only be available through the "Windows Store" - and only apps compiled to use its "Metro" touch interface will appear there."

Nope. Straight from Gruber himself: http://daringfireball.net/linked/2011/09/14/for-realsies

technoslut's comment/link (above) would disagree.

>You can ask Mac apps to behave like iOS apps, which is what Lion’s Automatic Termination feature does, but it has to be opt-in.

Virtualization could enable you to run a legacy app, stop its processing instantly, bring up a new app, and save the 'background app' state to storage when it's convenient.

This seems to be where they are heading. I don't know how it would translate to ARM tablets, but intel wants in on tablets anyway.


The OS can stop any process at any time and resume it later. There's no need for virtualization here.

The added complexity of allowing an app inside a VM to see outside the VM probably outweighs the benefit of checkpoint/restart.

I had been reading statements like this as meaning that they wouldn’t be doing Rosetta-style emulation of x86 software on ARM...but that developers would be able to recompile traditional Windows apps for ARM. Now I’m thinking what they mean is more profound: that on ARM, Metro will be the only Windows interface.

I think that's a very black and white way of looking at it. Sure maybe MS will rule out C++ x86 apps targeting winforms apis, but there's no reason to assume that they will also exclude C# apps targeting WPF.

>I’m hung up on the question of how any OS that lets you do everything Windows does could compete with the iPad, because the iPad’s appeal and success is largely forged by the advantages that come from not allowing you to do so many of the things Mac OS X can do.

So you're saying the iPad is successful because it can't do stuff OS X can do? Sorry I don't understand -- that sounds a bit silly to me. I thought it was the portability and size and ease of access to apps (the ecosystem around the device) that makes the iPad successful. If we could have the hardware power of a desktop system on an iPad, while keeping the simplicity of use, I'm fairly sure we would all like that.

With every step Apple is moving OS X closer to iOS, and iOS closer to OS X. Will they ever combine the two? I see no reason why they couldn't eventually with the amazingly quick progression of the relevant technologies that we're seeing.

Now whether MS is doing the right thing by combining them now -- I'm OK with saying I have no idea until I actually play with the device. Maybe they can pull it off, maybe not.

I thought it was the portability and size and ease of access to apps (the ecosystem around the device) that makes the iPad successful.

All of these things are possible because of the restrictions in iOS.

For example, the lack of a filesystem forces apps to have a more simplified data access mechanism[1].

The good battery life and response time comes from the fact that the OS puts severe limits on what each app can do.

[1] - There's an escape hatch of course, with something like Dropbox

> the lack of a filesystem

What?! How are files stored on iOS if there's no filesystem? Maybe you mean no user visible filesystem?

What!? I thought you had an actual concern! Maybe you're just being needlessly pedantic?

I don't know why everyone (well, ok, I know why Gruber does that) focuses on Apple and iPad after seeing Windows 8. IOS is an established system, with a healthy ecosystem of developers and apps. IPhone and iPad dominate the mobile market with ease. That won't change anytime soon. Besides, Apple had showed in the past that they can survive living as a Microsoft competitor. They will do just fine. Who should be worried about Windows 8 is Google. Android tablets didn't set the world on fire, Chrome OS from the very beginning looks like a low-profile project. And here comes Windows 8 which offers the same thing as Chrome OS (cloud, HTML5/JS apps, etc.) and does that in a better fashion than Honeycomb. Plus, it will run Windows software. Apple users learned to live without Windows software, Google users - didn't. Considering the fact that Microsoft will offer an integration with Bing, Skype, Hotmail and your latest Nokia phone by default even the great Google services could be in trouble.

I feel like one of the few who sees the emperor with no clothes.

Let me first start by stating that I do believe that touchscreen devices will continue to revolutionize industries, as they already have. But, why in God's name are Microsoft and Apple trying to shoehorn the touch-screen onto the desktop?

Mac OS X Lion is probably the last OS X version that we'll see. With each version it's gotten closer and closer to behaving like iOS. It seems that Microsoft is doing the same with Windows 8. Think about this, the human-device interface with touchscreen devices and desktop is different. The whole paradigm is different. With touch, you have your finger. The other, you have mouse/keyboard. The user interfaces that cater to one don't cater to other very well. Why force it?

I believe mobile is the future. But, I'm not sure this is the best evolution for desktop interfaces.

TL;DR: Metro UI looks nice, but merging touchscreen UI with desktop UI is a mistake... a la same Windows 8 for all devices.

EDIT: Please share your thoughts.

Have you ever experienced coworkers discussing something in front of the computer and touching the monitor screen time and again to point out things? It is not the most popular use case, but it does show that touching a screen to interact with a desktop is a natural inclination. Even with a desktop.

I don't agree that it's a natural inclination for working. For talking to someone else, okay, but touching the monitor for actually doing work is a fast route to gorilla arm.

I don't think Microsoft is merging both UIs at all. I think MS wants to pretend that Metro is something you'll use with keyboard and mouse, but I bet you won't, and they know that.

My thinking is that MS wants to sell you a hybrid tablet device. By day, you have it docked, and you spend most of your time in Desktop with mouse and keyboard. When it's going home time, you pull it out the dock, and it becomes a touchscreen tablet where you spend most of your time in Metro.

Have you actually used a touch screen on a Desktop? I'm not being snarky here; I genuinely believe that you should actively try to use a touch screen with a Win8 desktop before deciding one way or the other. I do it every day, and while it is not great for everything, there are some places where it is really nice.

For example, I love having a touch monitor when I'm reading a long Word document. Why hold a mouse and use the scroll wheel as I read a 35 page spec when I can sit there casually and flip through the document as I read? Or, if I have two hands on they keyboard typing and need to switch to another window, it can be nice to just tap the screen, rather than grab the mouse, navigate to the correct place on the screen and then click the button. Touch isn't perfect for everything, and I use the mouse plenty, but it's nice to have both. The key for me, however, was that I initially had to force myself to use the touch screen - I just wasn't used to it. Once I got in the habit of using it, however, I've found it has it's place and can be nice.

That would make sense on laptops and other small devices, but not so much if you're sitting 2-3 ft away from 2-3 large screens. Which is not uncommon for programmers and designers these days.

Right now, from where I'm sitting, I can't touch my desktop monitor without leaning forward uncomfortably. Even if large touchscreens become cheaper over the next few years, my arms aren't going to get any longer, and my field of vision isn't going to get any wider (larger monitor = sit further away). In this situation, touch isn't simply imperfect, it's physiologically impossible. It'll be even more impossible if your "screen" is a 50" plasma TV on the opposite wall.

So it seems that @jonpaul does have a point.

The other practical problem I see: fingerprints. I never touch my laptop screen because I'm looking at it all day and even at full brightness you can see previous fingerprints.

Haha, you're right. I have dry skin, so I make heavy use of hand cream... which doesn't play nicely with touchscreens at all. The smudges are especially annoying on glossy screens.

I'm not sure how Apple is trying to shoehorn touchscreens in Lion. They have made a very clear distinction between OSX and iOS, and instead added more gestures to the touchpads instead of trying to make a Mac with a touch screen.

I agree though, that trying to mix both is a mistake (one that's been tried for 10 years without much success).

Would his opinion have been any different if metro was the only mode available while undocked and classic PC mode was only available while docked?

Metro looks like it competes with iPad and the rest seems to be a version of Win7 under the hood. It feels to me like cmd is to windows as windows is to metro; something under the hood for power users.

I can imagine taking my computer on the bus and reading hacker news in metro and then when I get to work I plug into a dock and open visual studio. That seems to be the vision and I think I like it.

I'd have to control what was going on in the background of my computer when I was undocked, but if I'm enough of a user to set up background server like processes then I should be clever enough to understand that heavy background processing will eat my battery if I don't act responsibly when I unplug it.

The new more powerful ways to do diagnostics are exactly the type of tools I'd want to be able to control power; so it really feels like MS has a similar vision.

It will only be a matter of time before mobile phones start replacing laptops for desktop-like use cases. People will likely use their phones as if they were laptops with wirelessly-connected keyboard, mouse, and screen. I can totally see businesses start buying Windows-ready phones for their employees than the alternatives.

Eric Raymond describes this sort of transition here: http://esr.ibiblio.org/?p=1759

I think the more likely interpretation is that they're "betting on the future"; i.e. that tablet and other "portable" devices will continually get more powerful and that letting services and applications run in the background will continually become less and less of a problem. Seems like a pretty safe bet to me.

This isn't really a ground breaking revelation but something that MS has been hinting at for months if not years. Win32 was never suppose to be so tightly integrated in the Kernel, it was suppose to be a user mode subsystem just POSIX.

MinWin was a kernel cleanup effort that was suppose to be part of Longhorn and then Win7. If MinWin is finished in Win8 then Win32 would just be a subsystem along side Metro and neither would be dependent on the other or necessary for the other to operate.

I can't find the reference but I remember reading somewhere a few months back that Win32 may not even be installed by default in Win8 and that it would only be installed when you attempted to load an application that needed Win32.

I'm guessing Microsoft doesn't know yet how it's going to address this quandary.

As a long time third-party Windows developer, I don't call the rectangular regions "Windows" I call them frames or containers.

Also a few questions could be answered by downloading the Developer Preview. The Windows 7 desktop still exists, (thankfully for these of us that need to get our jobs done), but how a user will use Metro on a desktop or laptop will probably be dependent on their tasks and desires. As database application developer and admin, I'll be sticking to the desktop and the CLI a majority of the time.

Bias is one thing, but this relatively picky, especially for Gruber who can't stand when people pick apart Apple pre-dev.

I love Apple, but I really respect what Microsoft is trying to do here. The whole "You did great, kid, but maybe next time..." routine is a little much. It smells like a smear campaign more than an inquiring mind. This article stinks of fear.

>You can’t ... remotely log into an iPad.

Tell that to Aaron Barr. I know the iPad wiping story isn't confirmed, but it would be pretty easy to refute were such a thing impossible.

I remember Apple claiming that the iPhone runs OS X.

OS X and iOS are both running the Darwin OS with a different User Land UI running on top.

It does, iOS is a minimal OSX with a customised GUI.

Gruber's wrong on this. The ARM version will include the regular Windows desktop, Office etc. Microsoft has shown this multiple times. They have demoed even ARM desktops with a mouse and a keyboard!

Video back from January: http://www.youtube.com/watch?v=rvzJmRBS84w (watch from 2 mins in).

I'm not sure the interesting split is ARM vs x86. There is clearly a lot of interest in putting ARM chips in laptops and desktops, so why shouldn't the desktop version of Windows 8 support ARM?

At the same time Intel is trying to make mobile chips, so it might just be that the mobile version of Windows 8 will be Metro-only, regardless of whether it's running on x86 or ARM.

Just because they demoed it that way doesn't mean they'll ship it that way.

Who needs Windows that cannot run their old lovely trojan.exe and virus.exe^W^W^W old-proprietary-crap.exe and in-house-crap.exe that was developed 10 years ago and all contacts of the developers, leave alone sources, got lost or never been made?

The ability to run an old win32 desktop crap.exe is what Windows is all about. Only a complete idiot will choice it as a platform for a new, build from the ground up project, or, god forbid, a server.

And there is enough ways to run a web-browser, especially plug-in-less one. It is called Android. ^_^

Registration is open for Startup School 2019. Classes start July 22nd.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact