Serial interfaces (was Re: Any former Psion 5 owners out there?)

On 7/20/10, Chris Elmquist <chrise at pobox.com> wrote: I mildly disturbed that Amazon decided to share with me that customers who bought that, also bought the OXO Pastry Scraper. I do think that providing moderate amounts of power along _a_ serial interface is a handy thing. There's a pseudo-standard to do that for embedded devices with real serial ports over pin 9 of an AT-pinned DE-9 (normally RI). I have more than one peripheral where that is either always so, or optionally so (my 1999 Matrix Orbital intelligent serial LCD has a solder pad to optionally short pin 9 to Vcc, but don't do that and use the wrong cable!) If one were to define a new standard connector so that there weren't issues with re-using a pin on a common connector that means something else to different devices (power over pin 9 in particular), between that and some sort of device ident protocol (HEREIS anyone?), at one point in the past it might have been possible to give many of the benefits of USB to an RS-232 serial port (max transfer rate difference would still remain, naturally). OTOH, Commodore tried to change connectors (by swapping the genders compared to an IBM-compatible machine) of the serial and parallel ports on the Amiga 1000, and added in power to peripherals at the same time. It didn't catch on - the A500 and A2000 came with "PC compatible" serial and parallel port pinouts. There have been various peripheral hacks over the years to steal power for devices using one connector for data (serial, parallel) and one for power (AT and PS/2 connectors, ADB, etc) I do think that USB started out right by requiring power on the primary cable, but looking at "powered USB" (with even more cable-end varieties), they clearly weren't forward thinking enough. -ethan