Talk:Thomas write rule

This article is rated Stub-class on Wikipedia's content assessment scale.
It is of interest to the following WikiProjects:

Computing

	This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing articles
???	This article has not yet received a rating on the project's importance scale.
	This article has been automatically rated by a bot or other tool as Stub-class because it uses a stub template. Please ensure the assessment is correct before removing the `\|auto=` parameter.

Hi,

In the middle of this entry, you stated that

"The Thomas Write Rule relies on the fact that T1's write on object A is never seen by any transaction and postulates that the schedule above is equivalent to the schedule below where T2 occurs strictly after T1, and that hence the write of T1 can be ignored:"

What doe you mean by "the write of T1 can be ignored"?

Thanks, Fong

Answering Fong: If you have a logical ordering (e.g., timestamps) that differs from the wall-clock ordering of two write operations, you want to enforce the logical ordering. Let's say we have 2 transactions, T1 and T2, both of which perform a single write to a single data value A. We'll assign T1 and T2 a timestamp when they start, so let's say that T1's timestamp comes first (denoted T(T1) < T(T2)). Given this (logical) ordering, we should see T2's write after both transactions finish.

If T1 is allowed to execute first, then T1's write will be overwritten by T2, and A ends as it should. If T2 executes first however (out of logical order), then we need to drop T1's write operation, since the final value of A has already been written. Xthemage (talk) 15:27, 23 April 2013 (UTC)[reply]

Hi, this article seems not so orphan as stated, at least this page make references to it:

Timestamp-based concurrency control

It seems an important concept though, even if not so long to say :)

(Maybe lost updates are goin' in some place in the fifth dimension, as do all our lost goods :-D )

[Sorry for not logging in, had to recover my userid] Bert--83.103.109.6 (talk) 13:45, 3 August 2009 (UTC)[reply]