On Thu, 9 Sep 2010 Grant Ingersoll <[email protected]> wrote:
> On Sep 8, 2010, at 4:31 PM, Jake Mannix wrote:
> 
> > Hey all,
> >  One question I'm wrestling with is whether we should have
> > disk-backed (local or HDFS) vectors and matrices transparently
> > implement Vector and Matrix, or whether this will lead to confusion
> > and trick people into treating them in the same way they think of
> > in-memory versions (ie think they can do lots of random access
> > operations which just thrash and aren't performant).
> > 
> >  Any thoughts?
> 
> I'd say add the disk backed ones and we'll worry about education
> separately.

+1


> Perhaps it's possible for the vector to keep track of
> thrashing and spit out warnings. Either that or you override the
> random accessors on the file-based ones to throw exceptions so that
> it fails early for users.

The first options gives users more freedom about what to do with the
implementation - while the latter one probably saves us a few mails
about the implementation being sooooo slow... I'd be fine to go with
either option.

Isabel


Reply via email to