People are happy using Chromebooks as development platforms because they're $250 for an ARM laptop with long battery life, and you can hack them to run Ubuntu, not because Chrome OS has any merit on its own.
Things that are cute on a $250 machine are not so cute on a $1,300 machine...
I have two Chromebooks. One with 3G, that runs the stable ChromeOS, and the other wifi only, runs ChromiumOS, latest ToT, mostly because I'm working on getting ChromiumOS to run on our hardware, so it's a bit easy to compare and contrast when things are working and when they aren't. It's dead simple to boot off an sdcard or USB stick. The Chrome/Chromium did a REALLY good job at making the information available.
I'll probably buy the Pixel as well. The more I use ChromeOS, the more I realize how much time I waste on other crap that doesn't really help me get done what I want to do. And using ChromiumOS, there is the "dev_install" command which will set up a Gentoo chroot in /usr/local, that points to the machine that you built it on, and if you're running the devserver, you can run emerge pkgname, it will shoot it off to the server to build it, and once it's done compiling, installs in the chroot. There is also gmerge which will install ChromiumOS packages for you (e.g. A new version of the browser or maybe a newer kernel, without rebuilding everything) OR you can build a new image, hit up the Help page in settings and it will generate a diff on the server and update the entire install to the latest image. It's really an interesting setup, and I really like it.
Everything goes through gerrit, so you can even help out with version bumps and or patches fairly easily. The documentation is really good, and if you happen to hang out on IRC, the ChromiumOS developers are extremely helpful when something isn't clear, and even update the documentation for clarity.
I believe it's much easier to hack this one to run Linux than any ARM-based Chromebook. It certainly has enough memory, storage and processing power to be a decent development machine (provided you avoid heavy IDEs).
It's likely possible to even run Windows on this machine.
It's already braindead easy to put Linux on the ARM Samsung Chromebook. One of the first things I did was to assemble my own personal Linux distro and besides having to copy some blob files out of the original ChromeOS image, it's all very straight-forward stuff and I pulled up about a hundred packages I wanted including an X.org stack without fuss. The compile speeds are adequate if you aren't cross-compiling, I've certainly put up with worse.
I personally have little interest in the Pixel based on the specs. I think X86 is excessively 'big iron' now for a majority of needs and I find the lack of USB3 is mystifying. The screen looks interesting, but it's nothing I actually /need/ and certainly not worth another thousand bucks. I've personally taken to just using X86 for storage/cross-compile servers for the rest of my cheap ARM/MIPS/etc. crap and I've made it a point to stop buying expensive hardware. What $250 buys you now is actually pretty ridiculously awesome.
* Drop Crouton[2] onto Chromebook to get a full dev stack and unfortunate Ubuntu/XFCE environment.
* Set up chroot and start building other people's crap.
* Write to SD Card/internal storage and reboot.
Which step here is hard? Tedious to roll your own I'd give you, but you don't even need to as there's stuff like ArchLinuxARM[3] which skips the middle two steps.
Not gonna lie, I just spent several minutes trying to figure out what context sensitive notifications for Android had to do with hacking the chromebook.
Either after flipping the dev-mode switch and running a chroot or such (thus being outside CrOS's use case), or using it entirely as a thin client (which doesn't need this kind of horsepower).
I really hope that running a chroot or vm becomes a fully supported part of chrome os along with an x server. Working with a crouton ubuntu chroot on my (free) chromebox feels a lot like using OS X circa 10.3 or so and I would consider using it as my full time dev box if autoupdate worked in dev mode. I think i would actually prefer it to modern os x if I could tweak some keyboard shortcuts and run an x sever (even an annoying one like apples) mixed in with aura instead of having to switch back and forth.
I'll assume that's sarcasm but I haven't personally noticed that trend. The only way I made my cr-48 useful was by flashing the BIOS, adding a bigger SD, and running ubuntu on it.
Chrome has a wonderful SSH client [0], and with a $5/mo Digital Ocean [1] (or other [2]) VPS for all your UNIX work, it's a perfectly capable dev machine.
Plenty of people already do this with the $250 Chromebook, and even iPad. The $1,299 isn't for the 32GB SSD. It's for the 239ppi 2:3 multitouch Gorilla Glass display. It's for the 4GB of RAM. It's for the dual core Intel i5. It's for the impressively made aluminum chassis. It's for the 1TB of cloud storage. It's for everything a high-end cloud device should be.
If you're logging on to a remote server to do development,you're not taking advantage of the dual core i5 or 4GB of RAM. You're logged into an image that's giving you maybe 1 GHz of CPU and 512MB of RAM. You're buying a $1,300 machine to do development on a $300 machine, over a shitty internet connection.
It depends what you're working on. If you're doing heavy client-side work (e.g. WebGL) then a beefy client is preferable to a beefy server.
I agree though that I probably wouldn't buy this for development work. But I do think that browser's are still better on beefy hardware than on gimped hardware.
But if you also have ~20 tabs open in your browser, and are streaming music etc... I am doing a thin client setup of sorts, and by the end of the day things start slowing down quite a bit.
I do agree that $1300 is too much, at that price range you might as well get a mac. If it was 600-800 i'd probably jump on it. However, I think the curve ball here is the touch screen. I'd be willing to bet that Google's strategy for this laptop is to do a smaller production run, and get these in the hands of tech/programmer types, and then use the lessons learned to develop a model targeted toward a wider audience.
> I'd be willing to bet that Google's strategy for this laptop is to do a smaller production run, and get these in the hands of tech/programmer types
If that's the case, pull an Apple and ship great dev tools on the installation disk. As far as I can tell, there are no top-notch Chrome OS development tools, just various ways to remote into a real machine to do that work.
Pretty sure Square disagrees, and they know what they're talking about when it comes to XCode. See PonyDebugger, a tool they wrote to pipe iOS apps' network connections and debug information through Chrome to take advantage of the developer tools: https://github.com/square/PonyDebugger
Now, you'd never write code in the Chrome Developer Tools (except maybe the occasional one-liner in the REPL). But the Chrome Developer Tools are great — best in class even — for profiling, visualizing, and debugging.
You do for things like WebGL, client-heavy web apps, and other wonderful upcoming web technologies. The point is that this laptop should last you until it physically wears out, not until it becomes obsolete.
I think it might work reverse to this. You don't buy your think client and then figure out how to live with it. Rather, Google is betting that in the future people will be using thin clients and cloud services because of all the other reasons they are attractive.
Once you've already decided you want to live in the cloud, then your question becomes, what is the best possible device I could buy for doing that? And that is what the Pixel is for.
Thin client, thick client, cloud, local, etc, these are all only relevant to technical people. Focus on use-cases. What does the Pixel let you do that a Macbook Pro does not? Have access to all your documents from anywhere (over internet connections of various levels of unreliability?) No, because you can do all that on a Macbook Pro. Does taking away the ability to do things locally improve the user experience in some other way? Arguably, it improves maintainability and makes the UI easier to use.
Is that the target market for the Pixel? People whose main concern is reducing maintenance burden and having an easier to use interface? Okay, now how many of those people are better served by an iPad?
> Arguably, it improves maintainability and makes the UI easier to use
When your entire OS is cloud based, I think you get some synergies - you can reliably sit down at any computer and everything "just works" with all your state exactly how you left when you got up from the previous one.
So yes, you could purchase a Mac Pro, but because it's not cloud based from the ground up there are massive gobs of local state such that you continually need to worry about sync'ing things here there and everywhere, installing apps everywhere, etc. So in world where the default thinking for every application is to store local state, removing the capability for local state is necessary to achieve a true thin client.
> Is that the target market for the Pixel? People whose main concern is reducing maintenance burden and having an easier to use interface?
I think that's not really the selling point of this high end machine. This is Google's statement that cloud based computing is better even for people who do demanding, complex tasks (I'm not saying they are right, but I think it's what they believe).
> Okay, now how many of those people are better served by an iPad?
As I said, you're underestimating Google's ambition for what can be done with a thin client. An iPad is not suitable for highly intensive desktop tasks at all. A chromebook has all the physical features of a high end professional laptop, but is 100% cloud based, and thus can be used as a real computer for real tasks.
> A chromebook has all the physical features of a high end professional laptop, but is 100% cloud based, and thus can be used as a real computer for real tasks.
Except it can't. There is a theoretical possible future in which everyone has 40 mbps unmetered LTE and powerful apps exist in web form, but if anything the trend has been the opposite in recent years (carriers eliminating unlimited data plans, apps moving back to native on tablets and phones).
Yep, I completely agree - I'm hugely sceptical about whether this future where we can all rely on high speed network access 100% of the time will ever arrive. But I don't think Google is - they are betting this is coming. And in the meantime they are trying to provide just enough local state (HTML5 style) to get you by during the outages.
> Okay, now how many of those people are better served by an iPad?
None of the ones who need to write documents and emails, or edit spreadsheets.
For spreadsheets, it's still less than perfect because Google Spreadsheets are still worse than Excel wrt scripting and managing large datasets. But security, ease of use, and lower maintenance burden might make up for it depending on what your needs are. For my parents, this would be perfect.
Unless you wanted the gorgeous 2560x1700 display... Many people have paid close to the whole computer's price just for the screen (I'm typing this on a 2560x1440 screen that I paid $999 for).
I don't see the point of having a pixel density that high; almost no one has sharp enough eyesight to distinguish individual pixels at even lower densities than that, and enough people have poor enough eyesight that they won't be able to read text on software that doesn't scale its fonts properly.
That resolution is appropriate for a monitor with twice the diagonal size.
Also, what's up with the 3:2 aspect ratio? That seems awfully odd.
almost no one has sharp enough eyesight to distinguish individual pixels at even lower densities than that
That's exactly the point. Apple markets its high-PPI displays as being higher-resolution than the human retina[0]. It does need some help from software to not cause usability problems like font scaling and tiny images, but we've had the hardware capability to do this for a while. It's past time we improved on this.
Also, what's up with the 3:2 aspect ratio? That seems awfully odd.
If I had to guess, I bet a lot of the people involved in bringing this machine to market wanted 4:3, but someone in marketing and/or a focus group said that would be perceived as old-fashioned, so they got as close as they felt they could get away with. I'd really like to see a writeup on the reasoning from someone involved though.
[0] I believe the accuracy of said claim is disputed.
> That's exactly the point. Apple markets its high-PPI displays as being higher-resolution than the human retina.
But the resolution doesn't need to be any higher than the maximum resolution of a human retina. By definition, you'd be unable to see the difference; you're just left with the scaling problems of an extremely high resolution on a very small display.
Apple's successive retina devices have decreased in pixels per inch, but have increased in pixels per degree at their intended viewing distance as larger high-PPI devices were introduced. Of course, sometimes people use devices at distances other than those the manufacturer intended, and some people have better eyes than others.
The point is, current high-PPI screens haven't actually banished the pixel from human perception just yet. There are still gains to be made, but they aren't nearly as significant as quadrupling pixel counts over previous devices.
You mean the iPad? I remember the iPhone having a very positive reaction.
As a general rule, most of the time when people say "this is going to be a failure" they are right. The iPad is the exception, not the rule. And Apple has market-making ability that Google can only dream of.
Well, I mean, given that most new product launches and most new business ventures fail, of course the naysayers will be right more often. That doesn't mean they actually have insight into why products fail.
So, that having been said and because this is the internet, I'll rush in with my probably foolish and ill-advised explanation for with this won't be a game changer: In this case, I think the price vs. capabilities will make people compare this unfavorably with other ultrabooks/macbook airs. For the price, there isn't much here that can't be copied, easily, and quickly – and if I really want the cloud connectivity, I can get that on my ultrabook.
What's the unique value proposition in this product? I just don't see it being compelling – except that it's not an Apple product (don't get me wrong, I'm typing this on a Macbook, but I recognize that some people don't like the fruit vendor).
> That doesn't mean they actually have insight into why products fail.
I think it's gut reaction: I wouldn't use this product or recommend it to my friends. I think most of the time, that gut reaction is on point. Sometimes it's wrong, like the iPad, because you don't see the use for it until you try it, but that's the exception.
>Well, I mean, given that most new product launches and most new business ventures fail, of course the naysayers will be right more often. That doesn't mean they actually have insight into why products fail.
I regularly use the SSH client. I wouldn't call it wonderful; it works, but that's about it. Highlight a wrapped line? GNOME Terminal will Do The Right Thing, Chrome's SSH client won't and gives you a newline. See a URL? GNOME Terminal will let you click to open; Chrome's SSH client won't.
Not even remotely true. Super high pixel density screens are an unalloyed win. Everything you do on one looks better, most particularly, text rendering, which is absolutely superb on my rMBP. I would expect Google's screen to be comparable.
I spend 12 hours a day looking at screens. Sharpness makes that more pleasant. So does a wide viewing angle, good contrast ratio, accurate color reproduction and adequate brightness. This screen likely does pretty well on all of those factors. Considering how much time programmers spend interfacing with screens and keyboards, it might make sense to spend a couple days' pay on good ones.
Many of us consider this screen's aspect ratio a big improvement over the 16:9 of most PCs for working with code/text as well.
Text sharpness on a retina screen is amazing. You know how lots of people turn of anti-aliasing for code because they don't like the blurriness it introduces? Anti-aliasing on a 250 dpi screen looks as sharp as your 1-pixel thick coding fonts at 72 DPI used to.
I got a cr-48 in november 2010. The new chromeOS remaps ctrl-o to "open window" even when in crosh, so you can't save files in nano or pico. I stopped using it altogether. It's gathering dust.
I'm still using mine (got it December 2010), but I replaced the ChromeOS with a Linux. It actually works great, especially for a free laptop that's over two years old. I agree that the last several updates to ChromeOS made things worse rather than better, though.
I guess you are talking about the $250 chromebook, I am using a lenovo x120, having all my dev setup in digitalocean 1GB server. Do you use X or just fine with command line/emacs? Just curious to know how other people are using ec2/linode/digitalocean for development.
These are really cool tools that I didn't realize existed! I think I may take advantage of one of them!
However as the other reply said, it seems silly to buy a big powerful incredible machine just to do all of your actual work on a virtual server. Why not just buy a cheap machine and an expensive server instead of vice versa?
I want good performance and it feels wrong to buy a beefy, sexy machine and gimp it with a 1 core 512MB development bottleneck.
AKA lesson learned: the pixel is not for cost-conscious developers.
If you're doing CPU/GPU intensive web development this would come in handy. WebGL is pretty much a no-go on the ARM based Chromebooks but would perform fine here.
I paid $200 for an Acer C7, $80 for 16GB of RAM, $100 for a 128GB SSD, and $50 for larger capacity battery to end up with a portable linux box (running ubuntu 13.04 now) for under $450 that is light, fast to boot, and has reasonable connectivity (wifi, gige, 3x usb ports).
This took me about an hour of reading, 45 minutes to go from sealed boxes to linux, and a small phillips screwdriver (one screw on the back to get to ram+hard drive).
So I have to buy another machine to develop and test my stuff on? What's the point of a powerful laptop if it's just a dumb terminal to connect to my actual development machine?
Why not just buy a $300 super cheap laptop and do the exact same thing? Sure I'll have less resolution on the machine, but seeing as the developing is happening on another machine all together, it seems like the performance would be similar.
Well, if you're a developer, presumably this isn't the first computer you have owned. I'm using a 5 year old cheapy HP laptop as my dev server and it works wonderfully.
There are other alternatives, of course, Cloud9 premium account gives you full shell access, Action.IO gives you full shell access as well.
Regarding the idea of using an old laptop as a dev server, have you run into any issues as a result of having the laptop on all the time (such as overheating)?
Also, how would you set it up so that you can log in to your laptop from outside your LAN? I have a few old machines lying around and it would be great if I could put them to use as servers.
Shouldn't have overheating issues as long as you sit it so that the vents can breath. I leave mine in the basement where it's a little bit cooler. My current uptime is 45 days, I only restart for kernel upgrades and never shut it down.
I purchased dynamic DNS, which IIRC is around $35/year, well worth it for the convenience of logging into my home server whereever I am.
You could put the original Chromebooks in Dev mode and have a terminal with a build environment, or just install your own Linux build on it. Or even Windows. Not the fastest machine in the world, but not the slowest either. I quite liked it. This feels wildly overpriced to me though.
None of which use the $1250+ machine to actually "run or test the code you write." If you are just using a browser for coding, why pay that much money for a machine to run a browser?
One could argue that if a browser is your only environment, that you'll be spending the most time in, then why wouldn't you want the best option for running said browser? The way I see it, if you're going to be working in a completely remote/cloud capable environment, then a powerful, fast, beautiful screened machine to work from makes the most sense.
My current job gave me a Dell Latitude E series machine for work. I work exclusively in AWS. The screen on this laptop is so horrifically bad that I practically get sick looking at it. Compared to my personal (2010 era) macbook pro, or the Retina MBP's I've used at my last two jobs, there is no comparison. I'd much rather have a lightweight powerful little machine that I can ssh from and have some sort of IDE or text editor to work on my AWS projects then worry about a big bulky crappy laptop that is overkill for my work.
Obviously I realize this is MY current use case, however I'd be willing to be that the majority of dev/devops type folks on HN have a similar type of use case. AWS, Heroku, just about any PaaS, Saas, or Iaas, etc... would take advantage of a machine like this beautifully.
Oh and the last time I had an Air (right before upgrading to the rMBP), My use case? Chrome, Chrome, Terminal, RDP, Chrome. That was at a PC Game dev/publisher company where I administrated all windows based desktops, mac laptops as well as all servers of misc variety...
Sooo.....Is this machine useful? I think it is, and I am excited to check it out and may look to getting one.
One more thing, I too am one of the lucky Cr-48 owners and LOVED the concept. Biggest reason I hacked it to load ubuntu? SSH. Had the plugin existed back then? ChromeOS ALL THE WAY BABY!!!
I spent some time with cloud9 convincing myself it was usable and bought a chromebook only to find that aswell as chromeos being a massive let down c9 ran like a dog on it. Main thing is it was ridiculously slow, the same actions taking noticably long on the chromebook compare to my laptop.
It's interesting to see all the people on here saying that there's no point to having a beefier machine to "just browse the web" and then see your comment which would appear to indicate that there is a need for more performant web-browsing devices.
Its ancient and these days massively underpowered but i could sit and do sid by side operations in cloud 9 and the laptop would always complete noticably faster. For example i would be debuggin and i would ask to go to hte next step and in the laptop it was immediate, on the chromebook i might wait a second.
Very very odd and very unexpected but it made cloud9 totally unusable for me. Chromeos is also horrible for nearly everything else i do, even reading pdfs, so i cannot think of any reason to shell out for another machine with it installed.
Totally dig hosted IDEs. I think it's the way of the future. But why do you need a powerful computer to use them? How does it help? I'm genuinely asking.
The display...and your hosted IDEs don't do all their processing on a remote host. I believe a lot of the compilation/interpretation/services are provided locally via JavaScript.
Am I just imagining all the posts I've seen on HN from people talking about being extremely happy using Chromebooks as development platforms?