11.23
Did you hear about the new approaching paradigm in computing technology? It’s all about the thin client cloud computing. I am getting the feeling we’ve been here before.
It’s really simple. Look at the architecture of your average CPU. As well as the registers you commonly have three levels of cache integrated into the CPU. Then you have system RAM. If performance is a concern the aim should be to have to use system RAM as little as possible, and to never touch the hard drive if it can be avoided. It takes about a dozen wasted CPU cycles before the data is fetched from system RAM. It takes a few million before data is fetched from the HDD.
Yet it’s considered a revolutionary idea to stick everything on the network and deal with utterly horrible latencies plus being at the mercy of your internet connection. You certainly won’t be able to use ChromeOS as proposed on a train or plane.
Engineers have been fighting latency problems for decades with all sorts of architectural inventions such as on chip memory controllers, mad sized caches, ramdisks etc. Yet software companies always have this fascination with sticking half your computer in a datacentre in Mumbai. It makes no logical sense.
I believe the technical term for all this ‘thin client’ mania is a ‘Solution seeking a problem’.
The Inner Platform Effect
ChromeOS also looks to me to be yet another example of the Inner Platform Effect. “The computer is the browser” is the claim. You don’t need a complicated OS when all you need is a browser! But Javascript and XMLHTTPRequest are simply not going to cut it for long. You’re going to need more advanced languages, better server communication, better graphics libraries (XHTML + CSS is only going to take you so far). I am sure eventually local app caching (installing) will be introduced and then the aim will be complete – A web browser extended out to be an OS, running on top of an OS.
The problem that should have been solved was the trust issues with app installation on a normal OS. If you can make sure an app can’t mess with anything it shouldn’t be able to via sandboxing and access control. It seems to me taking the one advantage web apps have (being unable to mess with the core system) and building a whole OS on top of it is much less sensible than retrofitting that ability onto a normal OS.
Thin Client
Since my main laptop died I have been using my wife’s Samsung NC-10 as my main development machine. Seriously. And I haven’t had any issues with it, (except having to press Fn+PgUp/PgDn to get Home/End). I upgraded the RAM to 2gb (From 512mb) and put Windows 7 on it and unless modern gaming is your aim this will do as a main computer for 99% of people. And it’s a netbook. And it cost around £300.
ChromeOS is not going to be out for another year at least. When it is released I’ll be able to buy (according to Moore’s Law) a new netbook at the same price as this one, but twice as fast. Google are intent on ignoring this fact, thinking they can sell a range of low-cost, low-powered laptops by making the whole computer a jumped-up web-browser. The problem is todays low-cost, low-powered netbooks are fast enough to run a full OS, and the ones a year from now are going to be faster than most people’s current computers.
I just really don’t see the public accepting an essentially crippled PC which only runs web-apps, and which only works when you have a network connection, all in the aim of saving a tiny fraction of the cost price of the computer.
Q: But all you’re documents will be automatically backed up to the cloud!
A: Dropbox, or any of a wide range of services can do this. Plus you get a copy yourself!
Q: But all your apps are available and auto-updated!
A: I think what Google meant to create was Steam for apps.
I have a great idea – it’s called ‘Thin Cupboard Eating’, where instead of wasting space storing things in your house, be it clothes, food or books, instead you save money, buy a smaller house, and simply go to the shops every time you are hungry or want to wear something. Genius! Maybe I can sell Google the idea for GoogleHouse!
Not understand advantages of cloud computating. Creates dum terminal. Easily replace. Scalability concerns not by engineering anymore because network connection all that latancy.
Google leverageing open source linux kernel for power of community. Create low cost solution at high price and control data.
That wraps it up about as well as I’ve read anywhere else. Everytime we hear about ChromeOS it becomes a little less shiny-looking. By the time it arrives only the most die-hard of the freetards (like Nichols) will still be cheering for it, and only because it rides on top of Linux.
If this had been a Microsoft initiative it’d be laughed at from one side of the web to the other by now.
LH is down again:
bX-1x36wx
As much as I usually find the tech press laughable, I have been impressed with the number of articles I’ve seen that are critical of ChromeOS.
What Google seems to fail to remember is that Apple pushed web app development for the iPhone, and it failed miserably. If the public doesn’t want the relatively simple web apps that a phone would bring, they certainly aren’t going to want them to replace their desktop functionality.
It blows my mind that we’ve achieved such amazing technology and there are those out there that just want to scrap it so that they can go back to, essentially, dumb terminals all at the whim of a central computer inevitably run by an asshole
Chromes OS = Microsoft Bob + Dumb Terminal + Etch A Sketch
Nothing like a dumb thin client where everything is the way someone else envisioned it.
Thanks for “Inner Platform Effect”. I wasn’t aware this concept was given a name, but it’s readily visible everywhere.
I use Dropbox and thought I was storing data in the “cloud”, but you’re right in that it’s more of a hybrid approach. At any time I can uninstall the client, and the data’s still mine. Sure they still have my files, but I knew that going in. But, if they want to somehow leverage my scattered “TODO” lists and take a crack at my KeePass database that only has forum logins in it anyway, all the more power to them.
Honestly, this is as far as I can see the “cloud” going: online repositories that cache local activity. A total shift in the direction of online-only is nonsensical since the hybrid approach yields all the advantages and no downsides.
Most apps have gotten a lot better at notifying you when an update is available. Users who are conscientious about updating don’t need an OS to do it for them (I use Secunia Personal Software Inspector to check them occasionally myself: http://secunia.com/vulnerability_scanning/personal/ ) and those who aren’t won’t appreciate the service.
It’s just not a selling point.
An important point that many people miss is that browsers cannot compete with desktop when it comes to rich UI features. Despite the fact that HTML+CSS+AJAX, even + Silverlight / Flash can do great things, desktop is always ahead. Just consider W7 for example – there you can do easily things, which will cost many weeks of efforts for a web developer, and ultimately work under a strict selection of browsers. And I am not speaking of powerful content creation apps, where desktop has always been a winner.
It’s touted that Chrome OS will bring netbook prices down … but if you pay 200$ for a Chrome OS netbook, you won’t actually get a cheap netbook, what you’ll actually get is an “expensive internet browser”.
“I believe the technical term for all this ‘thin client’ mania is a ‘Solution seeking a problem’.”
The actual technical term is “A solution that *creates* a problem.”
Just trying to be helpful here, no offense intended.