Re: A WTF in the Oracle reference...

On Thu, 19 Jan 2006 13:26:30 -0800, thomas.kyte wrote:

> Actually, the publishers/authors of most books are more than willing
> (in fact they are dying) to listen to constructive feedback on their
> work.
>
> Somethings that sound "not so smart" in books and quoted in forums:
>
> o are taken out of context (and in context are just fine)
>
> o are admitedly not explained well enough in the text (the author left
> out too much assumed knowledge, they are for the most part very much
> ready to revise or clarify)
>
> o are not so smart (and need correcting, we all make mistakes)
>
> If you were to have pointed this out to them - and they could not
> justify/clarify it - but just ignored it, that would be a different
> thing.

A book I'm reading right now makes the statement

"
Serious problems with rollback segments can sometimes cause snapshot too old errors. Snapshot too old errors occur when a rollback segment runs out-of-space and uncommitted transactions are overwritten.
"

With a big hmmmmm about the 'uncommitted transactions are overwritten', I wonder about categorizing this one.

Comments appreciated, before I contact the author.

-- 
Hans Forbrich                           
Canada-wide Oracle training and consulting
mailto: Fuzzy.GreyBeard_at_gmail.com   
*** Top posting [replies] guarantees I won't respond. ***