<div class="moz-cite-prefix">On 22/08/12 10:56, Kevin Grittner wrote:<br /></div><blockquote
cite="mid:5033CBE002000025000499BD@gw.wicourts.gov"type="cite"><pre wrap="">Josh Berkus <a
class="moz-txt-link-rfc2396E"href="mailto:josh@agliodbs.com"><josh@agliodbs.com></a> wrote:
</pre><blockquote type="cite"><pre wrap="">First, note the change in topic.
This whole discussion has gone rather far afield from Miroslav's
original submission, which was for temporal tables, which is NOT
the same thing as audit logs, although the use cases overlap
significantly.
</pre></blockquote><pre wrap="">
I don't think the concerns I raised about apparent order of
execution for serializable transactions apply to audit logs. If
we've moved entirely off the topic of the original subject, it is a
complete non-issue.
-Kevin
</pre></blockquote><p class="western" style="margin-bottom: 0cm">Hmm...<br /><br /> I was simply using an audit log
exampleas a more specific case to understand things!<p class="western" style="margin-bottom: 0cm">Right now, I am meant
tobe working on a project I'm way behind on! Hopefully later, I will have the time to read more carefully the
interestingposts following my mention of the audit log example and to reply as appropriate.<p class="western"
style="margin-bottom:0cm">About 10 years ago, I implemented some temporal features in a database to cope with insurance
quotesthat had to be valid for a specified number of days in the future that was invariant with respect to future
changesin premiums with effective dates within the period of validity of the quote. If anyone is interested, I'll see
ifI can find my notes and write it up (but in a different thread!).<p class="western" style="margin-bottom: 0cm"><br
/><pclass="western" style="margin-bottom: 0cm">Cheers,<br /> Gavin