On Tue, 30 Dec 1997, Bruce Momjian wrote:
> The following was sent to me. Does it fit our needs anywhere? Let's
> discuss it.
> >
> > I wrote an indexed file system some time ago in ANSI C. I've compiled
> > and used it on several platforms, but have never aspired to do very much
> > with it. I'm not really the marketing type, and I don't want to compete
> > with existing standards.
> >
> > I wonder if it could make any contribution to the PostgreSQL effort?
> > Here are the pluses and minuses:
[snip]
> > Maybe it could do strange sorts or handle BLOBs. If you think it could
> > make a contribution I'd be willing to learn and work on the appropriate
> > code. You're welcome to a copy of it or any additional information you
> > might want.
Bruce, while reading another thread about the tuple size, could this be
utilised in building some form of MEMO field, getting round the 8k limit?
We have the existing large objects that are ideal for binary data, but for
textual data this could be of use. From the client's point of view, this
is stored with the tuple, but in reality, the content is stored in a large
object.
Also, quite some time ago, someone did ask on how to do searches on large
objects (consisting of large unicode documents). The existing stuff
doesn't support this, but could it be done with this code?
Anyhow, just a thought.
--
Peter T Mount petermount@earthling.net or pmount@maidast.demon.co.uk
Main Homepage: http://www.demon.co.uk/finder
Work Homepage: http://www.maidstone.gov.uk Work EMail: peter@maidstone.gov.uk