klely using this to help the user to understand
the concepts.
... now of course I could be entirely wrong about all this, in which case
I'm sure someone will let me know.
Cheers, Bob.
begin: vcard
fn: Bob Orchard
n: Orchard;Bob
org:National Research
A very minor error in the message of a ReteException in the Fact class
...
final int findSlot(String slotname) throws ReteException
{
// try to find this slotname in the deftemplate
int i;
for (i=RU.FIRST_SLOT; i m_deft.size();
i+=RU.DT_SLOT_SIZE)
if (m_deft.get(i +
Take a look at the code in MultiFinctions.java and you'll see that to
do the subsetp test the Values of the list are converted to strings
with ToString and the type is appended to this string. These strings
are used to compare and determine 'equality'. This probably works
most of the time but
Following this a bit further, it seems that even:
(eq (new java.lang.StringBuffer "a") (new java.lang.StringBuffer "a"))
does not give the expected result. A quick look at the code in Value.java
suggests that it does use equals but I might be overlooking some details.
Bob.
Bob Johnson wrote:
in the development process?
The quick answer is yes ... the logistics etc may be a bit more complicated.
For now I'd like to get a beta version ready so that I can see what the
integration issues really are.
Bob.
Regards
S.T.
Bob Orchard wrote:
Hopefully a beta for FuzzyJess in April/May
I think perhaps even with the simple 1 pattern rule the overhead of putting the
rule on the agenda and firing the rule would make a big difference if the
test were done on the RHS, especially if there were a lot of 'something'
facts. Of course it can potentially get much worse if there are a lot
The programme manner_data.c did not work properly with VC++ 6.0 under
Windows NT without a couple of minor mods (use the constant RAND_MAX
in a few places so that the generating random numbers between 0.0 and 1.0
are correct).
See modifed program attached.
Ernest Friedman-Hill wrote:
Hi
I'm working on an addition to Jess to deal with Fuzzy reasoning ... hope for a beta
by mid june.
Bob.
Saiid Paryavi wrote:
Has anyone tried to mimic fuzzy behavior in the current version
of Jess? If so, could you pass on some ideas.
Thanks
Dr. Saiid Paryavi
[EMAIL PROTECTED]
As it is implemented, insert$ can only insert lists into a list. So you need
to write:
(insert$ $?l 1 (create$ 362032))
This may be a bug since the docs suggest that single values can be
in the items to be inserted in one place ... single-or-multifield-expression
but just below it states that
Consider the following simple program ...
(defrule r1
(f a)
=
(printout t OK crlf)
)
Then the following actions are taken with Jess50a5
Jess (batch e:/alexis.eller/resetBug.clp)
TRUE
Jess (watch all)
TRUE
Jess (assert (f a))
== f-0 (f a)
== Activation: r1 : f-0
Fact-0
Jess (reset)
Hello Ernest,
I'm making good progress toward a FuzzyJess ... the Fuzzy part is ready it's
just the Jess part that needs to be integrated.
Here is my problem ... please feel free to ask for clarifications,
make comments, make suggestions, be critical in any way ...
Criteria (conflicting of
Bob,
In CLIPS you could use the 'progn' construct to wrap several
expressions:
(progn
(assert (one 1)) (assert (two 2))
)
In Jess perhaps the simplest thing to do
is define a function (with no args perhaps) and execute it.
(deffunction x ()
(assert (one 1)) (assert (two 2))
)
(x)
For those who are interested in the Fuzzy Toolkit and FuzzyJava, I have posted a
slightly modified version (1.0 beta 2) that fixes a minor bug -- see
http://ai.iit.nrc.ca/IR_public/fuzzy/fuzzyJavaToolkit.html
or
http://ai.iit.nrc.ca/IR_public/fuzzy
Also I added another example, Fuzzy
See below for my thoughts on your questions ...
RAY X wrote:
Hi,
I have the following questions or doubts:
1) JESS reaches the goal with deductive reasoning, that is fires rules
sequentially.
If the data permits to fire many rules, select one rules
for firing and other rules are stacked
for
Sorry to use this user-group for this message but I think there are some
users of FuzzyCLIPS who read these messages and I wanted to get the
message out with as wide distribution as possible.
A new version of FuzzyCLIPS, 6.04B, has been released. It fixes a problem
introduced in after the first
Please replace 6.02 with 6.04 eveywhere in my last message.
It was very late and I didn't notice the mistake.
Thanks, Bob.
-
To unsubscribe, send the words 'unsubscribe jess-users [EMAIL PROTECTED]'
in the BODY of a message to
Rod Mancisidor reported a problem with FuzzyJess that showed
different behaviour for a pair of rules that if modified slightly gave
incorrect
results. The modifications should have resulted in equivalent rules with
the same output but did not. A temporary fix (shown below) requires a change
to
We've used A*E quite extensively. It is good but has become
quite expensive since we initially started using it. I believe it
now costs in the order of 10's of thousands of dollars plus royalties
for distribution. We haven't had a new release sent to us for well over a
year (perhaps 2) even
For thos who use the FuzzyJ Toolkit and the FuzzyJess extension
please note that there is a new version (1.1) available that is
required if you plan to move to Jess 6.0a5 or later.
If you encounter any problems pleae let me know.
Bob Orchard
NRC
[EMAIL PROTECTED
Check out the java runtime options to control the size of the VM heap.
These are -ms and -mx options (or -Xms and -Xmx in Java 2).
- Original Message -
From: [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Friday, June 29, 2001 11:24 AM
Subject: JESS: Manners test error
Hello,
I have
Send me an email at orcha...@rogers.com and I'll see what I can do.
Bob Orchard.
On 2012-10-07, at 5:55 AM, dselva80 wrote:
Hi,
Can I find Bob Orchard's fuzzy Jess toolkit for download anywhere? (academic
license) All links to the toolkit seem to be broken since Bob retired, and I
don't
progress has been made recently).
Bob Orchard
On 2013-03-09, at 8:13 AM, Horacio Paggi wrote:
Dear Sirs:
In the docs attached to the FuzzyJess Toolkit it is mentioned the use of
certainty factors in the rules' specifications. As long as I understand, to
allow the processing of the rules
22 matches
Mail list logo
