jbwaters@bsu-cs.bsu.edu (J. Brian Waters) (06/02/89)
In article <18984@cup.portal.com>, FelineGrace@cup.portal.com (Dana B Bourgeois) writes: > I have this glimpse of the Amy hardware running flatout with a multi- > plexer of some sort running off it splitting up the bit stream say > eight ways to eight UARTs(?). Maybe Ser: would have a problem with it. > But couldn't a custom device driver handle it? OK, OK. Go ahead and > let me have it. *WHY* won't it work? It should work just fine. I have had this idea before, but I never got around to finding a source of information on serial multiplexers. A replacment serial.device should do the trick, and you would want the mulitplexer to not have any inband controls so 8 bit protocols would still work. Does anyone know where I can find out about multiplexers prices and protocols? This might make a good summer project, if a multiplexer would be cheaper then a multi-serial card. -- Brian Waters <backbone>!{iuvax|pur-ee}!bsu-cs!jbwaters
rokicki@polya.Stanford.EDU (Tomas G. Rokicki) (06/02/89)
> > I have this glimpse of the Amy hardware running flatout with a multi- > > plexer of some sort running off it splitting up the bit stream say > It should work just fine. I have had this idea before, This may work, indeed, but getting the software to fly would be a major pain. Also, don't forget that a serial line usually is bidirectional, and there is no way Amy's serial port will read at 1,000,000 baud. I will personally guaranetee you that, especially with the DSP available, the time investment would not be worth it. Indeed, there are a lot of very very excellent neat software projects, each of which isn't terribly difficult, that have been on my todo list for a long, long time. I've been just getting busier and busier, so I'll probably not get around to them. They would make much nicer summer products and have a much greater chance of being generally used. Among these are: - Make GCC fly on the Amiga. Already several people are doing this; ever considered contributing to the effort? - Making GhostScript fly on the Amiga. It already works, but if we add a nice screen driver, Preferences driver, and IFF output, we could have a hell of a combination. (Stephen Vermeulen has done much of this, but a lot remains to be done.) - Add Preferences and IFF support to PLPlot---no more having to write a new device for each printer; just go through Preferences! Also, make it accept commands, equations, and data in a script file, rather than requiring C compilation. (I'm almost done with this latter.) - Add an ARexx port to mg3---Mike Meyer is doing this, and will do a much better job than I could have. Or any popular utility. - Test, debug, and further develop DNET/RFS---this could be the hit of the year. Perhaps modify the link protocol so it will work over lines that aren't eight-bit transparent. - Attempt to hack VM support onto the Amiga. Perhaps not so easy, but we could all learn a lot from a few such attempts. - Write a simple, clean, structured drawing package. - Write a SCSI tape driver for backups. There are a lot of really neat things to be done. Most of the above (except VM and draw) would be (IMHO) easier than figuring out how to replace serial.device with a multiplexer, and would have far greater impact on the community as a whole. There is so much energy out there, and so much of it is being directed into duplicate efforts and dead ends. I wish there was some sort of a clearinghouse for such ideas where expertise could be exchanged and efforts unified somehow . . . -tom
wolff@cs.purdue.EDU (Robert M. Wolff) (06/02/89)
From article <9650@polya.Stanford.EDU>, by rokicki@polya.Stanford.EDU (Tomas G. Rokicki): > [....discussion of projects for Amy....] > - Test, debug, and further develop DNET/RFS---this could be the > hit of the year. Perhaps modify the link protocol so > it will work over lines that aren't eight-bit transparent. > [....] > There are a lot of really neat things to be done. Most of the above > (except VM and draw) would be (IMHO) easier than figuring out how to > replace serial.device with a multiplexer, and would have far greater > impact on the community as a whole. > > -tom Well, Tom, I personally would REALLY like to see dnet using flow control and 7-bit (flow control is more important to me personally here at Purdue with their really odd fibre optic network...) Anyway, between Matt's dnet, dnet-nfs (up-and-coming) and Software Distillery's NET:, we REALLY have some potential...I would like to see Matt's DNET working on the parallel port so I could have my serial link up at 9600 to campus and NET: between my amiga and another amiga so I can use the printer on the other amiga and his resources too... :-D I'd like to contribute some time and effort...I've got a 'bit' of time this summer in the evenings... bob -- -=-=-=- Insert Standard Disclaimer Here, Please... -=-=-=- /************* MS-Dos? Me? No, You must be mistaken! *************/
FelineGrace@cup.portal.com (Dana B Bourgeois) (06/07/89)
I retract the suggestion of multiplexing SER:, and I offer my help on one of the programming projects Tom mentioned. It'll be just a wee bit 'o help since I'm a 3 month veteran of C but I have some time and plenty of enthusiasm. Who needs another pair of hands? PS. Email me please. Dana