Re: Select count(*), the sequel

> Even if somebody had a
> great idea that would make things smaller without any other penalty,
> which I'm not sure I believe either.

I'd say that the only things likely to bring an improvement significant
enough to warrant the (quite large) hassle of implementation would be :

- read-only / archive tables (get rid of row header overhead)
- in-page compression using per-column delta storage for instance (no
random access penalty, but hard to implement, maybe easier for read-only
tables)
- dumb LZO-style compression (license problems, needs parallel
decompressor, random access penalty, hard to implement too)