Tim Rowledge wrote:

> Andreas Raab wrote:
>
> >
> > * Decreased SmallInteger range.
> Not neccesarily. One could use
> *1 -> SmallInt
> 10 -> other immed, other bits provide details (unless only immed chars
> is used)
> 00 -> oop

That does not work sorry, since the 10 combo is used as a sentry value by
the GC to detect where the object header starts.

> No loss of precision caused.
>
> > In short, I vote against this proposal because I have yet to see a
> > compelling reason for the advantage of whatever the new immediate class
> > would be compared to the combination of added complexity and possible
> > negative side effects, in particular those related to VM and plugin
> > interface modifications.
> I agree that there is not yet a compelling reason. That's why it is on
> the list to consider on what might be the last practical opportunity to
> give it due attention.
>
> Other opinions? HMM, do you to present the case _for_ twobit tags?

Of course :-) In another mail I wrote some things about it. However, I've
been a happy Squeaker with 1 tag bit for a long time, and other VM changes
such as block closures etc. have much higher priority for me, too.

Cheers,
Hans-Martin