I agree that the proliferation of Printer protocols has been BAD, but how about if
each printer manufacturer on top of that used their own cabling system, electrical
levels and so on? Hardly better, would you say, no matter how well it is documented.
The Linux community, and in fact also the Windows users (my sister for instance), is
suffering from the hardware driver problem. The idea is good, but the effect is
disasterous.
And it is my belief that the Printers will do away with "a protocol per model" and
settle around a few concrete protocols.
Now, I am not the poor Open-Source coder who need to implement 7945 Printer drivers,
573 video drivers, 2645 modem drivers, and so on. But I bet they are not as good as
they would be if they were fewer of them!! Just look at the manufacturers own
drivers. Work well on Win95, quite well on NT, not so well on OS/2 and so on. Not
enough time available to produce good ones. OTOH, my sister could fix the bugs
herself (since only she and 4 others are using that video card), but she can't even
spell "C"!!
Niclas
Greg Olszewski wrote:
> On Tue, Jan 04, 2000 at 01:04:38PM +0800, Niclas Hedhman wrote:
> >
> > The standard external modem uses a set of STANDARDS to perform its tasks. RS-232
> > and Hayes commands, basically. (I acknowledge that there unfortunately are
> > variations to the Hayes command set). The WinModem specification removes the
> > RS-232 and Hayes command set away from the modem and into a software API
> > instead, and allow the Modem manufacturer to implement the communication between
> > the modem and the computer in any way he chooses.
>
> The same is true of ethernet cards, video cards, ide controllers, SCSI
> cards....
>
> > However, there are more subtle undertones in the whole WinModem thing. Next
> > there are WinPrinters, WinMouse, WinKeyboard and WinMonitor, where the
> > manufacturers just provide the software API, removing the Plug/Play between
> > many
>
> I follow you with keyboards and monitors, but you lose me with mice
> and printers. If you run xf86config, or look at the options in gpm.conf
> you'll notice that there are between 10 and 30 different *standards* for
> mice talking to computers.
>
> Printers are another big headache. I can't name a *standard* for speaking
> to them, but maybe postscript? Linux deals with this by using ghostscript,
> software postscript -> proprietary printer language, in software. Basically
> the same concept as winmodems.
>
> The difference I see here is accessability to the communication protocol.
> I think the real problem with software modems, is that the manufacturers
> haven't(won't?) release the information about how they work to open-source
> developers. This is a problem, where the manufacturers need to give, but
> not really a reason to call software modems junk.
>
> my cent
>
> greg
>
> --
> this is not here
