>>> Peter Eisentraut <peter_e@gmx.net> wrote:
> Kevin Grittner wrote:
>>> "is a natural consequence of the fact" --- There is nothing
>>> natural about any of this. Why is it a consequence and how?
>>
>> How could you possibly get any of those phenomena if there are no
>> concurrent transactions?
>
> I see what you mean now, but you could write out that logic in more
> detail.
Those weren't my words; I was quoting the SQL spec. It came about
half a page after they had defined serializable transactions by saying
that they must produce the same effect as if they had been run (in
some order) one at a time. The spec then defined "phenomena that can
occur during the execution of concurrent SQL-transactions" and gave a
table of which phenomena could occur at which transaction isolation
level. Immediately after the table was the note about "natural
consequence".
Since these phenomena are defined in terms of the visibility of the
effects of concurrent transactions, and serializable transactions must
have the same effect as if they were run one at a time, a natural
consequence is that none of the effects can occur. What gaps in the
logic to you see?
-Kevin