janssen@titan.sw.mcc.com (Bill Janssen) (11/22/88)
There was a posting last week that made a good point, and I'd like to follow up on it. The author posed the rhetorical question, What kind of network functionality do you want? Mr. Wexler seems to be proposing something like the ARPANET, in which packet-switching between TIPS is used to give virtual circuits. Others might be happy with a national mail forwarding network, like UUCP. The UUCP approach seems to be already here, in that a person may buy mail accounts on CompuServe and such, then run mail software on a PC to access the mailbox and transmit their mail. (Actually, I don't know of any software that does this, because I'm PC-illiterate, but there must be *something*. Anyone know what it is?). For those who don't like the restrictions/costs of CompuServe, FidoNet seems a reasonable alternative. But how do we get a commercially available InterNet? First of all, someone must take the responsibility for running it. I don't care if it's the US Postal Service, BBN, or the FSF. It could be different someones in different areas. (There seems to be a lot of historical precedents for this kind of start.) They have to construct TIP's, make them available from personal computers, and make the hookup attractive. They have to link the TIP's together via long-distance hookups. And they have to make all of this economically attractive. I wonder if it can be done. After all, the most used and asked-for service is E-mail, followed closely by anonymous FTP. Something like CompuServe provides both without having the hassle of supporting virtual circuits, if we allow downloading from a bulletin board to compete with anon FTP. What's the incremental payback of providing cross-country virtual circuits? What's the incremental cost? I think if anyone is going to build the commercial ARPANET, they are going to have to overcome some serious marketing problems. Any suggestions? How would *you* sell virtual circuits? Bill