March 25, 2017, 12:42:36 pm
News:
Pages: [1]
Print
Author Topic: Is Android the IBM PC of Our Generation?  (Read 2331 times)
JoeBorn
Neuros Audio Team
Administrator
Hero Member
*****
Posts: 1376



View Profile WWW
« on: October 22, 2010, 04:56:25 pm »

I wrote up a blog posting to try to articulate the reason we're in the Android business and why we see it as potentially so significant: open.neurostechnology.com/content/future-hardware
Logged

jborn (at) neurostechnology.com
#neuros on freenode.net
heyrick
Global Moderator
Sr. Member
*****
Posts: 338



View Profile WWW
« Reply #1 on: October 22, 2010, 10:34:10 pm »

[writing this here, feel free to paste it to your blog if you feel it is suitably relevant  Smiley ]

I wrote on my blog, after a discussion with friends, that to a degree this is the reason why Microsoft became all-powerful. If it wasn't for a part-bail-out by Microsoft over antitrust, Apple would have floundered and probably died a decade ago.

I mean, think about it. What desktop machines do Microsoft make?
Exactly.
They don't. They make an operating system, and leave the machines to everybody else. You can buy a Dell, you can buy a Hewlett Packard, you can buy an unnamed box with an MSI board inside. They'll all run Windows. They'll all run more or less the same software (minor differences, MMX vs 3DNow!, that sort of thing, but then different video cards need slightly different drivers...).

Now let's look at the alternative. A company makes a machine, provides the OS. There have been some amazing innovations, and some dreadful mistakes, but I can give you a list that reads like a roll-call of '80s computer giants - Acorn, Amiga, (Apple, to a degree), Amstrad, Sinclair - all of whom wished to create the entire package, but was not able to keep up with the reality of two large companies doing what they were good at - one with software, one with hardware, walking a tandem dance that led us through the last decade to where?

What lessons have been learned? As greyback says, I think Neuros was both brave and innovative to look to making the OSD an open design. I believe the world was ready for it, but I don't think the corporations involved were. Though I believe more and more people will demand open source designs, for so many awesomely cool gadgets seem half finished, as if run out to market "first" is the most important thing, and perhaps it is to the bean counters. For the end users, the opportunity to resolve personal quibbles (or download a firmware where somebody else has done so) means a LOT.

I'll give you an example. Our mains electricity is "noisy". Sometimes clicks and fuzzes, sometimes EPIC I-dunno-what that takes out the Internet and long wave radio. My Internet is provided by an Orange Livebox which works, but the performance is not spectacular. So I got for a few euros a WAG200G router. Turns out it was nuked by a lightning strike, but this just blew the protection circuitry and one of the ethernet ports (I use WiFi). A bit of TLC with a soldering iron fixed this. So I set it up, get myself on-line. It is marginally better than the Livebox. I find and install the OpenWAG firmware. Now I can telnet into the router. Now I have an extra set of menus for additional features like network activity, plus special options built into the ADSL chip - one if which is a 2dB margin for long lines on France Telecom hardware. Tick this, and you know what, the router has quite a trouble on very noisy interference and I might get a mere 10k/sec, but the Livebox would have given up a long time before. I've had it keep me on-line when the stats said I only had 1dB of signal to play with! All thanks to open source.

It is for this reason I wonder about the iPad. An Apple creation running Apple software. It's the whole package, but shouldn't it be a package people can modify? Isn't this the way forward in the future? If you need no other prompting, look at OpenOffice and Ubuntu almost ready to be Granny-Friendly (i.e. non-geek), and not only that, but very likely wiping the floor with Windows. Because, ultimately, the hardware will become less and less relevant - just as the PC has. Nobody quotes processor speeds anymore. Few people bother to say how much memory or cores their machine offers. It's not that relevant. What is relevant is what it runs. What you can do with it. What opportunities it provides. Look at smartphones. Are they phones, or are they mini-computers with useful "apps"?

I think Android builds upon the strengths of all of this. Multiple broadly-compatible systems, plus an almost-open operating system. This is evolution in action.

Best wishes,

Rick.
Logged
heyrick
Global Moderator
Sr. Member
*****
Posts: 338



View Profile WWW
« Reply #2 on: October 22, 2010, 10:50:23 pm »

And another thing...

I wonder if those techies who baulk at the idea of using Android in lots of new places are doing so because it doesn't make sense, or because they don't want to learn something new. Look at the deployment of Unix clones (usually some form of Linux). It turns up in all sorts of embedded systems - chances are your ADSL box is running it. Why? Surely it would be much better to write a custom bit of code, right? I mean, it's only got to connect a phone line with broadband to a bunch of ethernet ports, right? Well, why go to all the trouble to write, test, debug the core operating system of your router when you can use standard parts and a well tested and widely deployed operating system. Sure, you might need 32Mb RAM instead of 4Mb, but if you're churning out thousands of the units, you can probably negotiate the price where the additional expenses for increased hardware is more than offset by the cost saving of not having to design an operating system. Back in the Z80 and 6502 days, everybody did custom stuff. It was the order of the day. Now? To do that sort of thing would be madness.
I fully expect we will see, before long, an "AndroidLite" which runs well on extremely low-end hardware, like perhaps a budget SoC. This can be deployed in toasters, central heating timers, whatever - and the developer needs only throw together a pretty interface and the application to run it, the rest will have been taken care of.

You know what? I am in the process of designing a "smart" burglar alarm using a 6502 and some custom code. Basically because I have these parts lying around. You give me a little SoC for maybe $40ish [don't need Beagleboard complexity] with I/O interface for keyboard, LCD, and sensors, plus access to an easy to use IDE, I'll chuck away my old hardware and develop for Android tomorrow. And the worst bit? I'd probably have it running in a matter of days. This is where it makes sense to use an established lightweight OS, even in places where it might seem illogical.

Best wishes,

Rick.
Logged
JoeBorn
Neuros Audio Team
Administrator
Hero Member
*****
Posts: 1376



View Profile WWW
« Reply #3 on: October 27, 2010, 12:21:23 pm »

Yep, I don't think there's any question that you get it.  I don't think its so much resistance to learning something new that causes resistance, I think its a combination of the close up view that custom is smaller, simper and more efficient, the lack of appeal of working with "someone elses" high level system, and the fact that low level hackers often don't like high level systems at all (for many of them, that's why they became low level hackers)
Logged

jborn (at) neurostechnology.com
#neuros on freenode.net
Pages: [1]
Print
Jump to: