The $50 Telnet Machine |
| or: Ubiquitous Access for the Masses. |
|
| Every Internet service is accessable using only a VT-100 session over Telnet. Why spend $2000 on a home PC, or even $500 for an NC, when all you need is a terminal? |
The $50 Telnet Machine |
| or: Ubiquitous Access for the Masses. |
|
|
| Every Internet service is accessable using only a VT-100 session over Telnet. Why spend $2000 on a home PC, or even $500 for an NC, when all you need is a terminal? |
|
Everyone is hopping on the '$500 Network Computer' bandwagon. Not that I object - I'm looking forward to the little suckers. But the price worries me: $500. American. That makes them several thousand here. (Ok, maybe only $AUS800) But, that's still without a monitor or modem or keyboard, meaning the total cost arrives at around $1400 for a working setup. That's too much. This means that the NC is being targeted mostly at businesses. Well, it makes sense there: In a strange way it's a step back away from the 'personal' computer and back to the mainframe/terminal structure. Businesses are finding that it costs lots of money to maintain each desktop computer, so they have invented lots of schemes and workarounds to prevent this. The NC is a good idea in the curcumstances. However, as a scheme to give net access to the masses, it will fail miserably. Nearly everone agrees with the concept of ubiquitous access - the idea that everyone on the planet should have a presence on the net - because it's vital for the social and commercial growth of the Internet. It's no use having a business if no-one can get to your shop. Everyone having net access will have an effect on society similar to everyone having a telephone. But, $500 is too expensive. I know families whose budget leeway is maybe $20 a week, after the plain necessities of life. To expect them to forgive any and all luxuries they have for six months is plain silly. There are millions of people not on the internet because it's simply too expensive. So, what's a reasonable price to put on hardware? A cost that nearly everone can afford?
$50? I've been on the internet for nearly six years. And for most of that time, the one program I could not do without was Telnet. The internet is decades old, but Mosiac only really hit the streets two years ago! Gopher was around a year or two before that. But Telnet is what held the internet together. All you needed was Telnet and a unix account somewehere, anywhere, and you could do anything. You still can, just nobody seems to remember that anymore. Let's list the big things the internet does. First, the list of all things you can do with just a Telnet session and a unix account:
So, yes, most of the developments that made the internet 'great', or at least popular, aren't available. But perhaps it's already passed that boundary of popularity, and people will now accept a clunky interface to get at the functionality, and the content, that the internet now provides. And if we want everyone to have that access, we need an inexpensive solution which provides those basic needs.
By using technology that is ten years old and literally cheap as chips. Take a Z-80 microprocessor, 128K ROM, 256K ram, a UART, and some glue logic, and you have a complete, working computer. It's small. It's slow. But it works. A schematic is included below. The only thing not shown is the 'video interface'. Mostly because I haven't worked it out yet, and partially because there are two versions. The first solution is a built-in LCD display. B&W only. This is probably the more expensive option of two two, perhaps raising the cost of the device up to $60 or $70. The other option is one very familiar to a whole generation of microcomputer users - you plug it into the family TV. Of course, the device I've defined is built with ten-year old off-the-shelf technology. You could probably do the whole thing today with one microcontroller chip. The point is, however, that for $50, it can be done. Good grief, of you were going to go mass production, you could probably put the entire thing onto one single chip. This device could fit comfortably in the palm of your hand with modern surface-mount chips. Of course, for a usable keyboard, you'd probably make it a little bigger. The cost for this can be summed up thus:
These are estimated bulk purchase prices, because the only way this is going to work is if someone pulls together a few hundred thousand dollars of capital to start the project. Someone like, let's say, the Government. Or better yet, a company with that much to spend on consumer goodwill. Because there's no profit to be made on these things. None at all. The only 'profit' to be made is the stimulation of an entire information economy for an entire country. The secondary effects would be staggering. Let's take a couple of big examples:
Telstra spends tens of millions every year on printing the yellow pages and the white pages. They now have an electronics yellow and white pages on the internet. If they somehow get everyone in Australia on the net, then they can save that much money every year, while providing a more timely services, with more frequent updates. The Australian Taxation Office spends a similar amount of money distributing, collecting, processing, and returning Tax returns. And that's just the administrative effort: Electronic submission and reply could slash that incredibly. And, as an added service, intelligent forms could catch most of the entry mistakes that people make, improving the accuracy rate. And, we'd all get our returns back sooner. The Government doesn't hold referendums often because of the enormous costs of running one. If every second house had a net connection, then electronic referendums would provide a much cheaper alternative, as nearly every house becomes a voting station. Think of the impact ubiquitous email could make - the CES is dead, thanks to the government. How about a virtual one? From experience, being unemployed can be expensive. Just think of the cost of printing and mailing a dozen resume's each week. Small companies have big troubles contacting their customer base. Ubiquitous access give them that ability back using mailing lists and automatic response. Open learning becomes more interactive. The barriers to starting a new newspaper or magazine is removed. Community billboards can go online. Minority groups gain an active voice. Regular Democratic polling is possible. Junk mail you actually want to read. On-line legal advice. Library searches, and access to teaching materials. No more leaving messages on those damn answering machines. All available to everyone in the country for the same cost as having the electricity on. Of course, in terms of people who loose, Australia Post takes a big tumble. (Although not as big as you think. Personal mail is still a minority item to AP. They make most of their money in freight.) Printers take a tumble. Efficiencies mean that people may loose jobs. (Although what usually happens is that the extra people take on new roles and improve the level of service.) And big business looses some of their edge over smaller companies. Imagine a subsidy provided by the people who win - $10 million a year, perhaps, for a few years. This could put one, free, into millions of homes. Sell reduced-priced 'badged' devices with company logos on them, using corporate sponsorship. (Then again, looking at a brightly coloured 'Coke' telnet box could seriously upset my sanity, and the Golden Arches could push me over the edge. ) Let's return to the technical specifications of the Telnet Computer. What would it look like? There are a few possibilities: The first is a plain box with some sockets: A tv connector, a serial connector, a keyboard socket. The second is something that looks similar to those electronic organizers though perhaps a little bit bigger - a flip-top device with a rubberized keyboard and an LCD screen. This makes it a more self-contained device excellent for small business use.
What about software? In order to make a connection through someone like an ISP, you need to have running a TCP/IP stack, PPP, a VT-100 emulator, and the telnet program. All this can fit quite comfortably into 128K of ROM. With a good programmer, maybe into 32K. With a really good programmer, there's probably enough room to fit another bunch of general purpose utilities in as well. To give some perspective, remember machines likes the old TRS-80, the Microbee, or the BBC micro. The BBC fit an operating system, advanced BASIC, an Assembler, and EcoNet network transport all in it's tiny 32K of ROM. You probably can't even buy Static RAM chips as small as the 2K devices that were in those machines anymore. My pocket organizer has more memory. Of course, something that's missing from this scheme is a modem. (as it is from all the NC devices) But, as long as you are limiting yourself to the text speeds needed to support VT-100, you won't need more than 9600 baud. 1200 will do, being as it is, too fast to read. How much would building a 9600 baud modem cost these days? $50? And I have a funny feeling that shortly, modems won't even be necessary. Suburban nets are coming. Optus is putting cable straight in people's houses. Any of the big Telco's could supply low-speed serial connections if they saw a reason. But, OK, let's assume $50 for a 9600 baud modem. We've just doubled the total cost of the Telnet Computer, but it still means the total outlay to get on the internet is $100 in hardware. And the central idea is simple: Cheap, almost disposable computers, applicances really, which have the single purpose of getting you on the internet with a simple Telnet session. Which is more than I had, when I was a lad.
|
|
