Re: Oracle Innobase Purchase Impacts MySQL.

Thomas Kyte wrote:
> In article <1129018489.325794.222960_at_z14g2000cwz.googlegroups.com>, Niall
> Litchfield says...
> >
> >Jim Kennedy wrote:
> >
> >> > Multiversion read consistency is hardly new nor unique to Oracle.
> >> >
> >> >
> >> > --
> >> > With regards,
> >> >
> >> > Martijn Tonies
> >> > Database Workbench - tool for InterBase, Firebird, MySQL, Oracle & MS SQL
> >> > Server
> >> > Upscene Productions
> >> > http://www.upscene.com
> >> > Database development questions? Check the forum!
> >> > http://www.databasedevelopmentforum.com
> >> >
> >> >
> >>
> >> Lets see I think they have had it since version 3, which was probably in the
> >> 1970's. So they have had it for quite awhile. (Yes interbase claims they
> >> had it before Oracle and then claims Oracle had it starting with version 7.
> >> Which is just incorrect, they had it long before version 7.)
> >
> >per Tom Kyte - version 4 in 1984 was the first Oracle release that had
> >the feature.
> >http://tkyte.blogspot.com/2005/08/something-different-part-i-of-iii.html#c112584079541388964
> >has the reference.
> >
>
> Actually, in following up with Ken Jacobs yesterday - it turns out version 3 had
> multi-versioning for non-blocking reads.

Oddly enough, I recall doing an evaluation of O3 and not testing this - we just assumed we would have to be doing versioning/locking manually, as we did with the non-O way we eventually went. At the time, it wouldn't have occurred to me not to put it in the specs, no engine did such things automatically, it was just something you dealt with (more or less - it took some skill to minimize blockers) in a multiuser design. We were glad we didn't have to manually calculate hash values to find rows.

jg

--
@home.com is bogus
Gotta get that bankruptcy in as a labor negotiation device before the
rules change:
http://www.signonsandiego.com/uniontrib/20051011/news_1b11delphi.html