[ColorForth] Why I'm Here

[Mark Sandford <pagercam@xxxxxxxxx>]

Being indepentant is great if you can get everything you want, but
if you want to print your code, attach a keyboard, attach a mouse
etc .... USB is at least a single port that allows one to get access to
a wide variety of standard peripherals that should support almost
everyones interest.  Serial or parallel ports are much easier to design

but the number of supported devices will be very limited and as the 
no-legacy PC movement continues the old style interfaces will go away
and you won't be able to find them, this day isn't that far off.  The 
no-legacy movement of a couple of years ago failed but as now almost
everything is available with USB now, the old interfaces of serial,
parallel, Floppy and ISA slots will disappear.  

USB isn't an easy standard to get working and this will be a major
downside, USB slave devices aren't too bad to program but masters 
require a significant effort to support a wide range of devices. 
Luckily Linux has been working on this for the past couple of years and
is fairly stable.  This is a pretty big chuck of code that would need
to be ported to Forth and goes against many of the recent comments of
I'll build a system to meet my needs and my needs alone.  I understand
this idea, as you often end up will significant overhead (baggage) that
isn't needed and contributes to system bloat and speed (lack of), but
for a single interface that can get to printers, modems, storage
devices(HDD,FDD, Flash cards), cameras, keyboards, MP3 players, the
effort seems worth while and those who want to run bare bones can.  

Another thread mentioned dropping TCP/IP support again I understand the
argument that this is a lot of work but again having this one protocol
gives access to so many other devices that it makes sense to put effort
into this one area.  One possible solution is to have a 10/100baseT
connection that can make requests of a PC/Mac/workstation and have the
drivers available on that machine using Windows/Linux/Unix/MacOS, i.e.
you want to write to a floppy send a ethernet/TCP/IP message and have a
small peice of code that receives this data on the PC that writes the
data to the floppy, no drivers required, but it does limit the overall
usage of the x25 to being a slave to the PC/Mac/Workstation when access
to a peripheral is required.  

If you are nonstandard you have to create everything yourself and the 
effort going into Linux should be a sign that creating device drivers
is a long and hard process.  If you only have intererst in a narrowly
scoped embedded system maybe you can get away with that but do not
expect help from others as they will be working on thier narrowly
scoped system that isn't compatible with yours.  Forth history seems to
be littered with many incompatable systems with limited or no
documenation, if there isn't some form of standard platform the history
will continue.  Computers have gone from multi-million dollar room
sized machines, to cheap small machines because they are generally
available and useful to multiple people, high volume => low cost.  The
x25 chip is only $2 if the volume is high, if there isn't support for a
wide variety of apps/devices the volume will be low, the cost will be
high and the concept fails.

The first thing that I learned as an engineer was don't invent when you
can buy, its always cheaper in the long run, so if interfaces or
protocols exist use them.  Interfaces or protocols that have wide
support are that much worth while, (Metcalf's Law: the value of the
network goes up with the cube of the number of nodes).  TCP/IP and USB
are perfect examples of a little concetrated effort paying off hugely. 
As stated above you can either take a stand-alone slant and use USB or
TCP/IP/ethernet for a network connectivity, but it seems to me that at
least one of the two should be considered a requirement.   


--- "Arthur W. Green" <goshawk@xxxxxxx> wrote:
> 
> > 
> > Will I abandon the PC when I get a 25x with USB peripherals?
> > 
> Along these lines...
> 
> Can anyone tell me if there is any particular reason why we couldn't
> just ditch
> most of the USB standard and just use a single port on a bus (if
> multiple ports
> exist per bus) as a high-speed serial port and forget all about
> device sharing
> and the like on a typical PC?
> 
> I figured at least for real-time applications, this would certainly
> cut down on
> granularity and the like...
> 
> If not, is there any particular reason why this isn't a good idea for
> USB
> capability on the 25x?
> 
> Best regards,
> -- Art
> 
> PS: I thought I sent this before, but I didn't see it in my inbox
> this morning.
> Sorry if this is a repeat.
> ------------------------
> 
> To Unsubscribe from this list, send mail to
> Mdaemon@xxxxxxxxxxxxxxxxxx with:
> unsubscribe ColorForth
> as the first and only line within the message body
> Problems   -   List-Admin@xxxxxxxxxxxxxxxxxx
> Main ColorForth site   -   http://www.colorforth.com
> 

__________________________________________________
Do You Yahoo!?
Buy the perfect holiday gifts at Yahoo! Shopping.
http://shopping.yahoo.com
------------------------

To Unsubscribe from this list, send mail to Mdaemon@xxxxxxxxxxxxxxxxxx with:
unsubscribe ColorForth
as the first and only line within the message body
Problems   -   List-Admin@xxxxxxxxxxxxxxxxxx
Main ColorForth site   -   http://www.colorforth.com