Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

A3Q1

0 views
Skip to first unread message

Ravi

unread,
Mar 17, 2006, 2:49:29 AM3/17/06
to
Hello,

I'm a little confused about question 1. The question implies that we
should be able to prove that the lamport timestamp on event e is smaller
than the lamport timestamp on event e' even if zero messages occur. How
is this possible? Say event e occurs on computer X, and event e' occurs
on computer Y. Furthermore, say absolutely no messages have passed
between X and Y. Say 10 events took place before event e happened on X
(there L(e) = 11) and say 5 events took place before event e' happened on
Y (so L(e') = 6). Since no messages pass between the two computers, isn't
such a case possible? Thanks in advance for your help,

Ravi

Khuzaima Daudjee

unread,
Mar 17, 2006, 11:10:52 AM3/17/06
to
You have given a scenario where the events belong to different
processes. The question, however, does not specify which event belongs
to which process.

Khuzaima

Khuzaima Daudjee

unread,
Mar 19, 2006, 12:03:28 PM3/19/06
to Kang Chu, j2x...@uwaterloo.ca, h8c...@uwaterloo.ca
Please see my post dated Mar 19 at around 12 noon.

Khuzaima

Kang Chu wrote:

> So what does the question mean? Are the events from the same process
> or different process?
>
> kang

Asem Rustum

unread,
Mar 22, 2006, 1:30:21 AM3/22/06
to
Hello,

Shouldn't Question 1 say:
e --> e' => C(e) < C(e')

instead of:
e --> e' => C(e) --> C(e').

In other words, use the less-than sign <, instead of an arrow -->,
following the convention used in the textbook p. 254 for Lamport's logical
clocks.

Thanks,
Asem

Zaid Masud

unread,
Mar 22, 2006, 3:30:27 AM3/22/06
to
I agree that the question seems to be using totally incorrect notation for
comparing C(e) and C(e'). But they have clarified in another post that for
the clock, "->" is the same as "<".

Jin Xiao

unread,
Mar 22, 2006, 12:37:57 PM3/22/06
to
Yes, my earlier post have cleared that "-->" for the logical clock means
"<".

cheers,
Jin

Filip Lazar

unread,
Mar 25, 2006, 6:03:31 PM3/25/06
to Jin Xiao
So which notation should I use in my final answer? I think i've figured it
out, (hopefully), but I would like to ensure that whatever notation i use
is considered correct. Are they interchangeable? Also, for the purpose
of the final exam, should
there by questions where we would need to make use of this notation, which
should we use?

Thanks,

Filip

Jin Xiao

unread,
Mar 25, 2006, 6:09:14 PM3/25/06
to Filip Lazar
Hi Filip,

You may use either notation, they are both acceptable for this assignment.
In final exam, both notations are acceptable too.

Cheers,
Jin

Ravi

unread,
Mar 26, 2006, 6:03:42 PM3/26/06
to
Can we assume e and e' are the events in which messages were actually
sent/received?

Thanks,
Ravi

zaid....@gmail.com

unread,
Mar 26, 2006, 7:49:43 PM3/26/06
to
If e and e' are in the same processes then messages don't have to be
sent/received.

Jin Xiao

unread,
Mar 26, 2006, 11:24:39 PM3/26/06
to Ravi
Hi Ravi,

You do not need to assume e and e' as the send and receive event to make a
valid induction. They could be any events.

cheers,
Jin

0 new messages