Xref: utzoo comp.windows.misc:830 comp.sys.next:1028 comp.sys.mac:24424
Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!rutgers!joyce!ames!think!barmar
From: barmar@think.COM (Barry Margolin)
Newsgroups: comp.windows.misc,comp.sys.next,comp.sys.mac
Subject: Re: wireless keyboard (was Re: replacing the desktop metaphor)
Keywords: desktop metaphor, graphical interfaces, computing environments
Message-ID: <34795@think.UUCP>
Date: 27 Dec 88 20:47:09 GMT
References: <4362@pitt.UUCP> <257@gloom.UUCP> <82702@sun.uucp> <8939@ut-emx.UUCP> <34173@think.UUCP> <83075@sun.uucp>
Sender: news@think.UUCP
Reply-To: barmar@kulla.think.com.UUCP (Barry Margolin)
Organization: Thinking Machines Corporation, Cambridge MA, USA
Lines: 51

In article <83075@sun.uucp> landman@sun.UUCP (Howard A. Landman) writes:
>In article <34173@think.UUCP> barmar@kulla.think.com.UUCP (Barry Margolin) writes:
>>Wireless communication is pretty noisy and error prone.  It is well
>>suited to low-bandwidth applications such as telegraphy, or less
>>error-sensitive applications such as voice.  For applications such as
>>terminal I/O integrity is important, so you would need an
>>error-detecting protocol between the PC and the device.

>Just because IBM's implementation failed to address the important issues
>and was unreliable, is no reason to punt on the concept.  

I never said the concept was worthless, it just has some problems, and
the one large-scale implementation of it didn't work too well.

>							   We already know
>of systems that face the same kinds of difficulties and work fine.
>Ethernet is one example, and even the old ALOHA packet-radio network
>worked well enough.  Implementing this at keyboard bandwidths should be
>childsplay.

A big difference between wireless networks and a wireless keyboard is
power levels.  Packet radio and microwave ethernets can use higher
power, so small disturbances aren't as likely to disrupt
communications.

>If someone walking between keyboard and computer interrupts communication,
>the keyboard should retry until the data is received.  Of course, to do that
>it has to *know* whether the data was received or not!

That's what I meant when I said that an error-detecting protocol is
needed.  The protocol would also need to be able to recognize and
discard duplicates, or else you'll get duplicate keystrokes when the
acknowledgement is lost.  To do this generally requires a processor
and memory, which would make the keyboard noticeably more expensive.
I'm currently skeptical that the benefits of a wireless keyboard (I
can't think of any myself -- why does the keyboard ever need to be
more than a foot or two from the monitor?) are worth the cost of
materials and development (not large, but not trivial, either).

By the way, the original article I was responding to also suggested a
wireless display (this would answer my question about why the keyboard
should be far from the monitor -- the keyboard/monitor combination
might want to be far from the CPU).  In this case, the bandwidth
requirements are much higher.


Barry Margolin
Thinking Machines Corp.

barmar@think.com
{uunet,harvard}!think!barmar