home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.parallel
- Path: sparky!uunet!gatech!hubcap!fpst
- From: Simon.Moore@cl.cam.ac.uk (Simon Moore)
- Subject: Re: Synchonizing watches
- Message-ID: <1992Sep4.121436.18178@hubcap.clemson.edu>
- Apparently-To: comp-parallel@uknet
- Sender: Piete.Brooks@cl.cam.ac.uk (Piete Brooks)
- Reply-To: Simon.Moore@cl.cam.ac.uk (Simon Moore)
- Organization: U of Cambridge Computer Lab, UK
- References: <1992Sep3.123226.14205@hubcap.clemson.edu> <1992Sep3.201927.25562@hubcap.clemson.edu>
- Date: Fri, 4 Sep 92 10:26:52 GMT
- Approved: parallel@hubcap.clemson.edu
- Lines: 34
-
- In article <1992Sep3.201927.25562@hubcap.clemson.edu>, leo@nas.nasa.gov
- (Leo Dagum) writes:
- |> In article <1992Sep3.123226.14205@hubcap.clemson.edu>
- steve@hubcap.clemson.edu ("Steve" Stevenson) writes:
- |> >[I saw this on sci.logic. I thought it was a nice problem for this group.
- |> > steve]
- |> >>From: dougs@tvnews.tv.tek.com (Doug Stevens)
- |> >>Message-ID: <1992Sep1.234452.26917@tvnews.tv.tek.com>
- |> >
- |> >[....]
- |>
- |> Unless I'm missing something, this is fairly straightforward.
- |>
- |> 1) A zeroes watch, calls B and hangs up.
- |> 2) B waits till phone ends ringing then calls A and hangs up.
- |> 3) B calls A again and hangs up.
- |> 4) A times the first call back, and now knows (AB + BA), when he gets
- |> the second call back, he now knows BA as well (and therefore AB).
- |> 5) A sets his watch back by AB, and calls B
- |> 6) B picks up the phone and zeroes his watch, they now have (approximately)
- |> synchronized watches.
- |> 7) Bob's your uncle.
-
- I have a problem with this - B could place the second call (number 3) at
- any time and therefore A is unable to determine the delay BA. For example,
- if B places the second call to A 1 second after the first, then all A will
- see is two calls 1 second apart.
-
- Bob isn't my uncle?
-
- - Simon
-
- [Simon Moore][University of Cambridge, Computer Laboratory][swm11@uk.ac.cam.cl]
-
-