125 comments

[ 1.9 ms ] story [ 195 ms ] thread
Let me guess

"It's not secure."

I really love the Windows 10 S page, which almost off the bat says "Microsoft Edge is your default browser since it’s more secure than Chrome or Firefox." with a footnote, referencing a study that focused on social engineering and phishing attacks.
This is disingenuous. The report focuses on different browsers' ability to automatically detect and block URLs of phishing pages and malware downloads. All three browsers have some kind of automatic URL blocking, and this report claims that Edge is the best.

At a glance, the results don't look immediately credible to me: I didn't see an explanation of how they chose the specific URLs, why there is a difference between the browsers, or the false-positive rate. But the study is, at least, actually comparing browser "security", and isn't comparing user behavior or anything like that.

Disclosure: I've never used Edge.

I didn't think that they were studying user behavior, but saying Edge has better security, full stop, based only on a study that only examines the URL blocking capabilities (and not e.g. protection from malicious JavaScript) is pretty silly.
I don't understand this decision. I get the argument that it's most likely developers who want the Linux Subsystem but as soon as you say that only some apps in the store work in 10 S you start the trail of platform fragmentation.
Easy, MS loves Linux.*

Too much loves turned into rape.

too much love normally leads to sacrifice, you are confusing between love and lust.
Microsoft lusts for Linux. They want the attention Linux gets. They want the tooling Linux has.

And they want the developers, before all of them switch to Linux or Mac. They desperately want developers to stay. They need it.

You are mixing developers with UNIX developers.

There are many kinds of developers in the world, some of us dislike being mixed with UNIX developers all the time on HN.

Do you realize you can upgrade easily to full version of windows with a little bit of extra cost? It perfectly makes sense. S is kinda limited sandboxed version of Windows of only educational environment use.
It is worth noting that this is a subset of the educational environment.

Primary and Secondary schools may find this of limited value due to existing licensing agreements. I used to work in a province which licensed much of the educational software used by public schools and first nations schools (and the scope of licensing ranged from school computers only, to teacher computers, to student computers). While this would probably be addressable with time, the transition would be a rough one.

Post secondary education is also a mixed basket. The majority of students may get by with a restricted range of software, yet many departments depend upon specialized applications. This would create barriers for students regardless of whether the software is Windows native (but not approved in Microsoft's store) or Unix software that can run in WSL.

While this isn't a huge issue for students with their own computers, who can upgrade to for the nominal price, it is likely going to create many issues for institutions.

the education is "do not compute without permission of the company"
Start the trail? The desktop vs. Metro/UWP fragmentation started a long time ago.
> most likely UNIX developers who want the Linux Subsystem

Fixed it for you.

Use Windows since Windows 3.0, never missed having a POSIX shell.

In fact, I am yet to install WSL on my Windows 10 machines.

I've used Windows since 3.1. The old DOS shell is woefully lacking compared to even Bourne (let alone bash or ksh or what have you).

PowerShell is a godsend, on the other hand. Even the better POSIX shells out there are woefully lacking in comparison to PowerShell. It's what the Unix shell should've been from the beginning, and I'd sooner use it over WSL any day (despite my home being a Windows-free environment of Linux and OpenBSD boxen).

On the third hand, I've recently taken up Tcl, and it's in turn a godsend. There are some things from PowerShell which I miss in Tcl, but they're minor in comparison to having Tk at my fingertips and a vast array of tribal knowledge encoded in the Tcl wiki. It also integrates reasonably well with both Windows and Unix.

I am using F# for when Powershell gets too complicated.

For multiplatform scripts, Python, as F# isn't available in all systems.

Love Tcl, I used it on a startup as our application server main language, back in the first .com wave.

Admit it. You don't want people to load external code.
It has been explicitly mentioned by Microsoft that they don't wan't external code being run on Windows 10 S.
Two weeks post launch there will be jailbreaks.

Happens every time.

(comment deleted)
Not being able to run arbitrary external code is kind of the whole point when building a security-oriented system.

You remove (not just hide or disable, but remove) everything that's not absolutely necessary, and the ideal result is a system that cannot run anything but a very small list of whitelisted binaries.

that may be the 'ideal' result but judging by the number of 0-day exploits it will not be the reality.

Even Apple let you side load apps onto an iOS device if you're a developer... Microsoft is definitely setting it up for a stumble (again) similar to WinRT.

You can probably side load uwp apps if you want to.
You can also sideload app all you want, just upgrade for 10S to home/pro/..
The way to protect against 0-day exploits is to have a smaller attack surface. If you remove stuff from the OS, you're not vulnerable to 0-days in that stuff anymore.

They're not intended to be developer machines - imagine the recent UK NHS WannaCry situation; it would be very, very nice if most of these NHS systems were running a locked down not-general purpose OS. If you're deploying a hundred "office PC appliances", then you'd prefer if they could not be simply sideloaded with stuff to make them developer machines, you'd want a clear separation between that - e.g. installing the normal Win 10 instead of Win 10S.

Who will do the whitelist? The end user or ms?

Solitary confinement has the side effect of being both very secure, while being a torture.

I guess if they really wanted to keep WSL, they could just "sandbox" it by disabling access to the Windows filesystem (that is, no /mnt/c and so on).
The linux subsystem is implemented via Windows kernel modules. "Just sandboxing" has proved to be a pretty hard problem for both kernels, so I'm pretty skeptical that simply removing /mnt/c would suddenly result in a reasonable secure system. Especially since the WSL code is quite young.
Gotta love this. not exactly "embrace,extend,extinguish" but,well,ahem,cough
Removing WSL from Windows 10 S is going to extinguish Linux? And what was the extension?
Offering it on cheap laptops?
true.

they (at msft) love linux nowadays, but only if you get a premium windows version..

It's not like you can't install Linux on low-end Windows laptops. The machine I take to hackathons is a very light, very cheap, very replaceable Asus that came with Windows 8.

The next one will probably come with Windows 10 S and never actually boot into it (like the Windows 8 before it).

You can, it's very easy. That's probably why 10S is there, to stopgap chunk of the population before they see how easy it is.

Was your first laptop a cheap 10S? or chrome? or did you install it yourself on a thinkpad?

It's all marketshare with them.

It's probably harder to install Linux on a Chromebook than on a 10S laptop.
I hate the fact how Windows S is shaping up to be a completely walled garden. Not even CMD and Powershell?? I hope they still spend enough time on actual versions of Windows which are used by power users.
But all the power users left at either 8 or 10
I didn't leave yet. I think Windows is fine. Runs pretty stable. In the last decade, I had like 2 bluescreens on windows, and one was caused by accidentally bumping really hard into my PC, causing a kernel panic because of a voltage drop, and the other was a broken driver. Linux still has a long way to go when it comes to UI, drivers and video games.
YMMV. Across several of my phones, my Linux servers and my Windows desktops; My desktops seems to be 'recovering from graphic card crashed' once every few weeks.

It also depends how you count. Numerically, the games produced for Android far outnumber the AAA titles.

I'm not gonna lie, I think most Android games are very bad and give 1 hour of enjoyment at the most. Gaming on a touch screen is some of the worst UX you can have imho.
Oh I agree with you on that one. They are stable games, just not on the production level with your average PC game.

About the UX, not sure if it's just me, but I want more character or object control when playing a game. I guess I have played on a keyboard+mouse for far too long. Phones/consoles just don't have the same level of satisfaction for me.

> Linux still has a long way to go when it comes (...) video games.

That's why I have a console.

.NET/Java/C++ developer happily using Windows 10 to type this message.
Some did it before Vista...
How do you run bash inside Chrome?
Is all automation for Windows 10 S supposed to be performed only via domain policies? For the Windows admin crowd out there, how realistic is that?
This is like Surface RT all over again.
I hope it's at least possible to install something else on them...

It's a shame the Surface RT didn't allow that.

It's a Pro feature and Microsoft tends to ship a lean distribution tailored to the customers needs in order to provide superior user experience.
The user experience gets better with more options.

A 'superiour' user experience implies lots of options. I.e. luxury car with all the gadgets, etc.

A stripped down application, void of options, doesn't provide a 'superiour user experience's. If anything it provides cheap entry model.

> The user experience gets better with more options.

Not really. Better UX comes from better options, not more, poorly thought out ones. The options available fit the purpose of the device. A luxury car has features tailored for comfort. A real luxury car focuses on the back seat. A sports car has featured tailored for "feeling" and controlling the car.

Having more options brings a superior experience only if your goal is to do multiple things, some never done before, sometimes in unexpected ways. Sometimes you use `sed`, others `awk` and there are days `cut` will cut it.

I'll only partially agree to that.

In the segment of operating systems, windows differentiated and promoted 'home' 'pro' 'ultimate' , etc. They already paved the terminology and more importantly trained the users to expect it. Many users after, got really annoyed with Windows 7 ultimate being upgraded to only win 10 pro.

So in the case of Microsoft Windows, superior means something different than Apples IOS where superior means less (removing of the physical escape key, audio ports etc).

Apple is a great example of a "luxury computer". Their machines are solid, have decent battery life and don't bother with gimmicks like touchscreens which bring ambiguity to the UI (tooltip shows when you hover, but you can't do that via the touchscreen, only by dragging a pointer through a trackpad - argh). Their most recent bold move was the touchbar, that brought in fingerprint scanning and set of soft keys (esc included) that vary according to the task you are trying to accomplish. The lack of a physical esc key doesn't bother me at all.
Backpedaling is Microsoft's favorite pastime. Promise a huge convergence in platform support, alienate a lot of developers in the process, then artificially refragment it again.
>But the Windows Subsystem for Linux is very specifically targeted at developers, who should be using Windows 10 Pro. And because of the way it’s implemented, it’s almost certainly not in keeping with the reliability, performance, and battery life advantages of Windows 10 S.

The stratification of users like this always irritates me. I don't think "developers" should be a special class of users; today's curious kids are tomorrow's developers! But if you lock them out of their own systems, they won't even have a chance to get started.

kids who get these in schools will never get to tinker. this whole trend is a tragedy
I agree. Most consumer electronics (phone and tablets) are non developer friendly - you cannot develop anything on them.

I find this quote by Russell Kirchner to be really good:

> They’re trying to get everyone to use iPads and when people use iPads they end up just using technology to consume things instead of making things

From https://impossiblehq.com/an-unexpected-ass-kicking/

There is also interesting observation made by Cal Newport in his book Deep Work [1] that teaching kids how to use iPads is counterintuitive as they - consumer products - are designed to be super easy to use and one should rather learn how to master hard things (e.g. programming).

[1]: http://calnewport.com/books/deep-work/

Contrast that with an Apple II or a Commodore 64. You turn it on and you are inside a REPL in less than a second.
As someone who didn't grow up with systems like that, I really wish I had. I was never really introduced to programming until high school, but I feel like I would've loved it at a younger age.

GUIs are nice in many ways, but in general they seem to encourage consumption much more than they do creation. I feel like there has to be some way to build a GUI with the same kind of expressiveness as those early PCs, but I don't know how.

I recall seeing some big name in the computing world calling GUIs point-and-drool interfaces.

And i can't help wonder if we ended up in this situation because the DOS CLI sucked, and the Mac didn't have a CLI at all.

Thus anyone exposed to home computing back in the day things CLI is something nasty and borderline useless, and this mentality has transfered over to desktop _nix, where the CLI is quite the potent interface.

Thus we have big name DE projects trying to insulate the users from the CLI, and coming up with ever more elaborate schemes of daemons and protocols to do what a few quick commands in a virtual terminal could solve.

Yet time and again we see that the power-users and similar reject these elaborate GUIs and their plumbing for simple tiling WMs on top of an X11 session started from a CLI login to run a pile of virtual terminals (and the odd web browser).

The DOS CLI didn't really suck any worse than the CLI of contemporary machines such as Apple ][ or Commodore or TRS-80.
Today, that's the browser.

Hit F12 and you are inside a REPL (the console).

Hit "View Source" and you can learn how a website works.

The fact that developer tools ship with most browsers is really cool and encourages learning in my opinion.

May have been the case back when sites were static HTML.

Tells you jack all how a modern site do things, as so much is hidden away at the server end.

> you cannot develop anything on them.

Yes you can, iTunes has quite a few development environments to play with.

Swift Playgrounds and Pythonista are much more powerful than what I could do on my Timex 2068.

I'm not sure.

It's artificially restricted to be barely usable. If kids aren't apathetic about the device but actually willing to use it, it's hard to imagine a chance they would cope with the limitations (they seem way too severe, compared with the usual locking) and won't research how to exploit it to get around them and get the device's capabilities to the fullest.

Also, piracy and kids go together like bacon and eggs.
You can also get an Arduino or Raspberry Pi at a far more accessible price these days. It's not that kids are losing access to 'tinkerable' devices, it's just that not all of the many computing devices we use these days are freely reprogrammable.
The question is - will parents buy their kids these things, and will they actively use them? I think the much more frequent case is that kids playing computer games stumble upon this command line thing and are really curious what's going on.
Hopefully, they can just download a Linux distro and reinstall their laptop.
Sure, they can "just" do that - if they know how or are interested enough in the machine to learn, which is kind of my point.

If kids grow up with machines that restrict their behavior, they'll come to accept those restrictions as the norm. Some will be curious enough to learn to install a new OS and such, but they'll be the exception and not the rule.

I wouldn't expect so. Surely Secure Boot will be enabled with no option, at least on ARM platforms, for users to disable it. (This would be keeping with Microsoft's requirements for prior versions of Windows on ARM.)
MS' goal is for kids not to learn the Linux subsystem, and focus on Visual Studio instead.

I thought Visual Basic was the way to program on Win 95-2001 until very late. It landed me a crappy career start, despite starting at 7yo. Getting Windows, even for free, is a straight loss in the long term.

The guys getting paid to develop VB.NET at some of our customer sites are quite happy with their salaries.
Could it be that there used to be an oversupply of low skilled VB programmers and that nowadays JS is the new VB?

I remember hacking simple VB apps as a self-taught teenager and I could probably get a job in it if there was an opportunity and if I didn't go to uni to learn other things instead.

Today I suppose VB is not much different from COBOL - just an old language used in few systems here and there maintained by dedicated specialists.

Or are there still new things built in VB?

On Lifesciences there are lots of researchers using VB.NET to write small applications to process data coming out from microplate readers.

The industry for these type of devices is mostly Windows based, with drivers implemented in COM or Win32 APIs.

So they get access to the SDKs and write small applications to process their data in ways that fit their workflows.

Honestly, I don't think windows is that bad of a system to hack on. Especially for kids, they're not going to be recompiling Linux, they're just going to download python and do "from turtle import *"
This is about Windows 10 S, though. Windows today is an okay system to hack on - and I say that as a Linux user who prefers to have even more control over my computer than Windows allows.

Windows today at least lets you compile a program and run it - Windows 10 S will block you from running a program you wrote yourself because it's not "trusted".

10s is aimed for education.

There will be programming apps in the store.

Yes, some silly turtle, scratch and stuff like that. I can tell you that as a 10 yo I wanted to build "real" software with real tools (even if it turned out shit by actual real software standards), not toys.

You would certainly find a hater in me if I was to go to school today.

(comment deleted)
My workplace Windows 10 Pro doesn't support it either. I use cygwin and VM's so I haven't bothered to ask our IT about it.
Cygwin was the way to have a Unix-like environment on Windows before the Linux subsystem existed.

Is Cygwin available in the Windows app store?

I don't get why developers even bother - just don't use Windows as dev OS - if you really need windows use it in VirtualBox - I spent years as .NET dev on Linux desktop that way. Recently tried otherwise with docker for windows and one day after some magical windows update HyperV just killed my docker VM. For me Windows is just not reliable enough to be dev machine primary OS - it's not worth my time.
I've used (and loved) VMware Workstation and Virtual Box, but WSL is a whole new level of awesome. There is something about the pureness of it's integration with windows (well, since creators update) that just makes it a joy to use as a client OS.

Now, you still need virtualization for a ServerOS - WSL is all running under Bash.exe, so when that exits - everything is gone.

But, for client OS - It's the bomb.

WSL is the best thing since peanut butter. Recently moved all of my dev command line work to the Subsystem. Even better, VS Code integrates it perfectly well.
Is it really any better than cygwin? Cygwin was the first thing I installed on a Windows machine, going back well over a decade. It did a fine job running bash and all the utility and X11 apps that I cared to use.
Tons better (I say this as a dedicated Cygwin user since at least 2003). Cygwin, bash, and tmux are the first three things I install on any windows system, but with Cygwin, you are limited to whatever Cygwin has repackaged for you. With WSL, 95% of the stuff in Ubuntu is just an "apt-get install" away. Only stuff that relies on special network access, or /proc and other system directories seems to have issues, and they are quickly working away on that. (Serial Support just landed).

And now that they are planning on supporting Fedora, Suse, and Ubuntu environments, I can yum install, or just drop in-place any of the Redhat binaries as well.

That's the big difference - I don't need cygwin compiled apps anymore - I just drop in the Linux variants and I am good to go.

sorry without Docker support in 2017 WSL is no go for me
I don't get why developers even bother

I word: laptops.

I've used a laptop as my primary dev environment for years. During that time I ran Debian on ThinkPads. Compared to Windows you paid a price in terms of higher power consumption and randomly breaking sleep and hibernate, not to mention poorer plug-and-play multi monitor/dock support. Add in unpredictable driver support and my newest laptop just runs Windows 10. If I need to use Linux I'll do so via WSL or VMs.

Good for you - I refuse to use laptop for dev since I like multi-monitor env - and this is much easier and cheaper on desktop.

On the other hand how often you do real work without AC ?

Frequently. It's a laptop, that's kinda the point!

Aircraft and buses are common locations, but even when I'm coding on the couch I appreciate not being tethered to an outlet.

And the sleep/hibernate problems are equally if not more important. I need to be able to close my laptop and reopen it without it consistently crashing. Debian Sid had a good stretch for a while there but right around the move to systemd things destabilized again and I'm just done with it.

Got it - still I've had to use laptop - I would go with one of Chromebooks with Crouton or System76 laptops
A chromebook as a dev environment is something I've never heard of. My dev laptop has 16Gb of memory and a 512Gb SSD with a Core i7 onboard. It gets 8 hours of active dev work on a charge. No Chromebook comes close to those specs.

More generally, Linux just fails on a laptop. See the comments here:

https://www.reddit.com/r/linux/comments/48y2db/system76_lapt...

And that's for a laptop designed to work with Linux.

Again, I have deep familiarity with running Linux on a laptop, having done so for over 10 years. Debian running on my old Lenovo T410 worked okay. Most of the time. Well, for a while, anyway. Assuming I could live with 2-3 hours of battery life and apt-get update resulting in random hardware breakage.

In general I do agree that it's hard to have good laptop/linux developer experience. Chromebooks are usually not very strong machines however I've seen people working this way on Pixels.

Heard lot of good about using better laptop with this https://www.neverware.com/ - still it's unclear how reliable it's

I don't see any justification to use win 10s on any system for any reason. It's inferior in every way. Why does it exist?
It's targeting the chromebook market of a curated marketplace for apps.
To be honest I have the same opinion about Chromebooks ¯\_(ツ)_/¯
(comment deleted)
Chromebooks have the advantage of keeping state in the cloud. The other day I started editing a document at the office on my desktop and finished it on the train back home. No need to even think where the document was - it's always available no matter the device I'm on.

Office 365 is kind of the same, I know. We can think of these machines as Office 365 clients just as much as we can think of Chromebooks as Google clients.

But, at least for Chromebooks, you can seamlessly add a fully functional Linux environment.

To create a captive audience for the Windows Store, so desktop developers decide it's worth their time to make UWP apps.
Or to put their win32 app in the Windows Store using the Desktop Bridge (a.k.a. Project Centennial)
As others have mentioned they want a Chromebook/iOS alternative. They could have achieved this with Win 10 already but my guess is that with Win 10 S and its 50$ upgrade fee, they want to artificially drive the value of Win 10 regular up (or they are afraid to cannibalize Win 10). Anyway, the whole move looks very much like something "old" Microsoft would've done. IMHO, this was just a matter of time.
Nobody here complains about Apple's OSX vs iOS, where iOS has no shell
iOS is a companion device. Except for Federico Viticci, anybody technical has a Linux/Windows/MacOS system to go along with their tablet.
People using Apple have the mentality of slaves. Of course no one complains.
Is this reddit?
I’m trying to get my karma back to 1337.
For a good reason. iOS is not marketed as "macOS for beginners" and doesn't run Mac App Store software.
It isn't? I'm pretty sure that's exactly how the new iPad Pro with keyboard is marketed.
Do you want me to? :)

Every single time I fire up Pythonista on an iOS device I soon realise that any meaningful coding I do requires access to external data. And then I give up, until next time some years down the line.

There is StaSh — a Bash-like shell for Pythonista (scp, git, ptinstaller, mail, pip, etc) e.g., you can run: `pip install youtube-dl`.

Pythonista for iOS can be used for automation (Workflow for programmers).

No because this is HN, and Apple reigns on SV.
We've all heard it, these comments are boring.
(comment deleted)
This shouldn't be surprising, as 10S is meant to compete with Chromebook/iOS devices, where system is completely locked and you can install apps only approved gateways. This may not be an ideal system for people here, but schools love this and also makes good for home systems by non-technical users. Lesser chances of Virus/Worms, corrupted settings, easy to refresh. Enabling a backdoor to it, would have defeated the purpose
To clarify: Chromebooks are (last I checked) unlockable.
Windows has outlived it's usefulness for everything but games and office apps.
One of the fun thing to do on ChromeOS is crouton, which opens door to full blown Linux applications, now before you accuse me of 'pro'/developer user, my 12 year old nephew has more Linux software under crouton with his Chromebook than I do.
Here’s the blog post linked in the article.

https://blogs.msdn.microsoft.com/commandline/2017/05/18/will...

> Windows 10 S is not well-suited for many app developers/hackers, admins & IT pro’s!

> If you want to run all your dev tools, distros, shells, etc. on a machine running Windows 10 S – like the sweeet new Surface Laptop – then upgrade it to full Windows 10. You’ll then be able to run Linux distro’s, Cmd/PowerShell, install dev tools, debuggers, profilers, packet sniffers, etc.

That sums it up.