Thanks. After I replied I realised that defining an "ad-hoc" template was
just as possible. Doh!
Thanks again,
Simon
On Tue, 18 May 2004 15:02:45 -0700 (PDT), <[EMAIL PROTECTED]> wrote:
I think Simon Harris wrote:
Is that for performans reasons or primarily understandability?
I'm building an app
I think Simon Harris wrote:
> Is that for performans reasons or primarily understandability?
>
> I'm building an app where the Java code constructs a very simple domain
> model that is asserted as a set of facts. The end user is then able to
> write ad-hoc rules, infering relationships that th
Is that for performans reasons or primarily understandability?
I'm building an app where the Java code constructs a very simple domain
model that is asserted as a set of facts. The end user is then able to
write ad-hoc rules, infering relationships that the app could never know
in advance.
O
I think Simon Harris wrote:
> As a general rule (pardon the pun), what would you consider "better" form:
>
Always prefer unordered facts (those for which you defune an explicit
deftemplate) over ordered ones. They will always be much more
efficient, and code that uses them is generally easier to
As a general rule (pardon the pun), what would you consider "better" form:
(deftemplate bundle
(slot name))
(deftemplate report
(slot name)
(slot bundle))
(bind ?b (assert (bundle (name "b"
(bind ?r (assert (report (name "r") (bundle ?b
OR
(bind ?b (assert (bundle "b")))
(bind ?r (assert (r