Slashdot Mirror
Future of the PC on NPR's Science Friday
EMS writes "NPR's
Science Friday news radio program today has
a discussion on the Future of the PC at 3-4pm EST.
Guest speakers Tim O'Reilly, Steven Levy,
and Scott Bradner discuss how PC's are
evolving in a "decentralized, network driven world". Contributing forces to be discussed
include Linux and network appliances.
" It's worth checking out NPR site later on-they typciall put these up in archives.
Putting it another way, a particular bit of hardware becomes less important, and instead the data and services become more important. So, say within a business, you can log onto any (or most, at least) computers around, and you get the same data, same functions, same customised desktop at each and every place. Or, instead of taking a laptop or handheld, you login remotely (at the hotel, airport, general kiosk, whatever) and access your stuff like that. Or you just borrow a handheld/laptop from anywhere, and providing it has the required client stuff, then it takes little effort to 'customise' it, because your particular settings and files are exportable.
An example of this sort of thing is Sun's recently released i-Planet software. All you need is a bit of hardware, with Java and internet access, and (without taking any software with you at all, just as password) you can remotely, and securely, get access to your mail, files, etc from it. (of course, you do need the required server stuff setup first.)
Windows is very badly suited to doing this sort of thing (in general, not just the i-Planet example), though Microsoft have been cludging in some hacks around this.
To workers/members of the public, it probably won't make that big a difference for a while, unless they do lots of 'remote' work. It will make a much bigger difference to the IT departments, and sys-admin though...
It'll be interesting to see how NPR frames the conversation with these folks, especially since Steven Levy was the primary author of Newsweek's recent cover story on "The New Digital Galaxy", a fairly mainstreamed, fluffified and narrow view of the future of PCs and pervasive computing. I know Levy is capable of much better, but he does a good job of bringing it to the masses wrapped in a candy coating. The issue also included an article by Bill Gates on " Why the PC Will Not Die". Possibly good background reading for the NPR bit.
I need to point out that most ISPs have a tough enough time managing just providing TCP/IP service. I don't think they are really prepared to handle all of the needs you outlined above. That sounds a lot like those SANs everybody has been talking about recently (Storage area networks for those of you who havn't). It sounds great in theory, but what it boils downto is a way to convert hardware limitations into software limitations and admin overhead. The scheme you propose above allows the users to rent as much CPU time/disk space/memory/etc... as they want (and as much as the ISP has to offer that is not in use by the thousands of other people currently using the system) whereas before they would have had to buy the hardware. However the ISP will have to buy the hardware to support that, along with the broadband connections, and the connections to the users's homes. In addition the ISP will have to keep upgrading to keep in line with the times (just imagine when a new game comes out that requires a lot of resources) and find a way to administer and manage all of these resources.
Anyway, I think this is going to cause a large amount of pain with end-users (especially when they get those monthly bills) and administrators alike, and will take a long time to be implemented, if ever.
I read the internet for the articles.
Go to Here to listen to the show. (requires RealAudio)
The real Webmaven is user ID 27463. I don't rate an imposter, because my ID is such a lame-ass high number.
-Games - Dedicated 3D rendering hardware is so cheap it can be embedded in $100 consoles - why not in TVs, VCRs, etc.
And what is it about dedicated console gaming that sucks? It's targeted at the lowest common denominator of player; it's nonupgradable; it's non-networked; platforms are tightly, centrally controlled (Hello, Sony and Tendo!)... I could bitch ages about consoles, but the point is that the last thing I want to do is introduce these factors into my choice of a TV (and why do I want to play games using a TV's embedded hardware? The lifespan of the display is vastly longer than the lifecycle of a game platform.)
-Surfing the Net - Your HDTV purhcased for $1000 in the year 2005 will have this built in.
Whoopee. And that will integrate with junkbuster coupled with my gig of shared local cache... how, again? That cache will reside locally on a... what? And I'll be able to script, automate, and customize browsing with... what? I'll publish my weblog and surfing habits... how?
A navigation interface for browsing and information collection is vastly different from a navigation and preference interface for a large public display device (like a TV). Or will I tune the characteristics of my display using some awful local web form / Java applet / ActiveX control / whatsit?
Sorry, but a TV with browsing functionality that I want *is* a PC with a really huge display. Never mind that I have no use for browsing in my home theater! Integration rocks, convergence sucks.
-Local storage (I gotta lotta stuff and I don't trust it on the internet) - So you buy a storage unit for your local wireless network. The minute you plug it in it announces its presence to all networked devices and voila, your VCR has a place to store movies, your TV a place to archive web pages, etc...
Fair enough. That's what I do with a PC server now. I rack up a bunch of disk in a box, throw it in the back room and hide it behind the couch. Every other networkable device uses it for persistant storage (even the Sun3/50 and the Amiga 500!). The household gig of squid sits on it.
Every CD I buy gets encoded to MP3 and stored on the disk array. It serves all but one audio output device in the house (the theater -- and only because I don't want disk & fan noise from a PC in there; I'm likely to fix that with a diskless MP3 client of some kind). If storage were cheap enough, and closed formats weren't, I'd archive my DVD collection, too, and stream that around the house.
I may not capture video to it, but that's partly because I don't capture video. (I don't even own a VCR!) If I wanted to store video, that's where it would go, then spool it out to DAT on a LRU basis of some kind.
A nitpick: if I'm that concerned about not sharing the my local storage, the hell if I'm going to translate it to RF broadcast and squirt it around the house!
But why do you need local storage if your ISP can offer terabytes of movies/audio/programs all over a broadband connection?
Because I barely trust my ISP to get packets to me, let alone store my data. I want my data on media I physically control, with access control at my sole discretion. Period.
Video/audio on demand is a waste of my time. Why should I download (even "instantly") data that I can store locally on LD, DVD, disk array, or whatnot? Even assuming infinite free bandwidth, the "I own this" factor is pretty strong here, too. This is the reason Divx failed -- people pay money to own something, and they own it. A transaction-based universe is overrated and assumes some things about privacy that I don't particularly like.
-Word processing, DTP, graphic design (general purpose applications) - broad band will most likely make it possible to rent the usage of such apps from your ISP and run them using a thin client. Might sound expensive now, but hardware and bandwidth prices keep falling. Do you want to buy photoshop with x plugins for $1000, or rent it at 50 cents an hour?
I apologize, but I could poke holes in the vision of a transaction-based outsourcing world all day. For one, I barely trust my ISP with my traffic, let alone with renting functionality with me. Why rent Photoshop at 50c an hour when I can "apt-get gimp" and be running in 20 seconds, for nothing?
For that matter, suppose I'm cleaning up my nude girlfriend snapshot library (assuming you don't banish secure, physical local storage with your fantasy of infinite bandwidth to outsourced servers, of course), do I want my usage tracked in that way?
"Oh, look, he's fired up Photoshop to touch up his girlfriend's nipples again!"
Do I want to rent an email client every time I interact with someone? What about PGP? Do I have to rent time on some master server to decrypt my mail? And is it then somehow transmitted to my thin client in a trusted way?
Assuming the existance and prevalence of Universal Thin Client Hardware (a long, long shot to say the least!) that runs whatever software I'm renting, that still isn't about the death of the PC in the first place. Thin clients have certain tasks their great at (manipulating data on remote servers using rich interfaces) and some things they suck at (being a PC and doing PC tasks).
Finally corporate drones will stop printing their email and schedules and just be able to carry their damned computer everywhere they go.
That would rock. I'd love to see that.
But mobile clients have limits given their lack of local storage, the requirement for wireless networking (which doesn't work well over long hauls, from in caves, on airplanes, or with data you don't want transmitted), and (presumably) limits on integration with other devices.
Compiling, rendering, other processor intensive tasks - Two words: server farms.
I'm with you 100%. Infinite bandwidth will make it much, much more cost effective to buy or borrow CPU time for intensive tasks from somewhere else.
Although if I'm rendering nudes of that girlfriend, I might still want some local CPU...
Some dork will always want a tremendously expensive box that does it all, but it will become less and less economically justified when most of the common devices in your home do 90% of your silicon based processing already, and the other 10% can be cheaply bought on demand.
I want a small number of cheapish boxes that are highly flexible, which sit at the center of a network of even cheaper dedicated devices. I want to have a high degree of control and flexibility over the application-specific sattelite units through a powerful programmable system. I want my sattelite devices to be able to access the greater resources (local storage/caching/connectivity/IPC and device coordination) of a central network of servers and fat clients.
Your home can be a network of microsmart devices that talk to each other. Mine'll be a network of smart, programmable PCs that integrate with dumb dedicated devices.
You distribute your computing tasks -- I'll distribute usability, accessibility, and flexibility.
I agree that a lot of what you say'll happen will happen, and probably in less than the 30 years you quote.
All I'm saying is that I like my way of doing things better.
-josh
Nice name. =)