Re: Roadmap for the future.

[Nat Friedman <nat nat org>]

The only frontend that will generate wildcard clues, as far as I can
imagine, is the GUI.

Also, it's only (num clues) * (num clue types), it's not actually
combinatoric, is it?

Nat

On Thu, 2004-02-19 at 12:50, dave wrote:
> On Thu, 2004-02-19 at 15:37, Joe Shaw wrote:
> > One way to consider doing it is in the engine so that the backends don't
> > need to know about the special "wildcard" match type.
> > 
> > Backends already subscribe to certain match types, so you could
> > synthesize new clues based on the union of all the supported match types
> > and blast those out.
> > 
> > I haven't looked at it myself, but it feels conceptually cleaner to me.
> 
> That's probably true. The only problem is where a cluepacket contains
> multiple wildcard clues - this leads to combinatorial explosion of the
> number of cluepackets we need to generate. OTH this shouldn't happen
> (should it?).
> 
> Attached is a patch which does it by generating appropriate new
> cluepackets. It only allows for at most one wildcarded clue per packet.
> I can fix this if you think this is a problem.
> 
> dave
> 
>