defrules in CLIPS

[Razan Abu-Khashabeh]

Hello,

I have problem in CLIPS rules: 

(defrule case-1
         (patient
            (age ?age)
        (MRN ?MRN)
        (first-name ?first-name)
        (surname ?surname)
        (gender ?gender))
    (physical-feature
            (HPO-ID ?HPO-ID)
        (physical-feature-description ?physical-feature-description))
    (mode-of-inheritance
            (HPO-ID HP:0000006)
        (mode-of-inheritance-description Autosomal-dominant-inheritance))
    (age-of-onset-classifier
            (childhood ?childhood))
=>
(assert (IBMFS-DB (cps-db))))

The error is: 

TMPLTDEF2] The single field slot 'cps-db' can only contain a single field value.

ERROR:
(defrule MAIN::case-1
   (patient (age ?age) (MRN ?MRN) (first-name ?first-name) (surname ?surname) (gender ?gender))
   (physical-feature (HPO-ID ?HPO-ID) (physical-feature-description ?physical-feature-description))
   (mode-of-inheritance (HPO-ID HP:0000006) (mode-of-inheritance-description Autosomal-dominant-inheritance))
   (age-of-onset-classifier (childhood ?childhood))
   =>
   (assert (IBMFS-DB (cps-db )

Any help is appreciated

Regards,

Razan

[CLIPS Support]

If you defined the cps-db slot in the corresponding IBMFS-DB deftemplate as (slot cps-db), then it's a single field slot and you have to provide a value if the slot is specified in an assert command. In the actions of your rules, the command (assert (IBMFS-DB (cps-db)))) specifies the cps-db slot, but does not follow the slot name with a value. If you want the slot to be able to contain any number of values, including none, you must define it as (multislot cps-db).

[Razan Abu-Khashabeh]

Thank you 

The deftemplate was (deftemplate IBMFS-DB (slot cps-db))

deffact: (deffacts IBMFS-DB (cps-db DBA))

and the defrule: (defrule case-1

         (patient
            (age ?age)
        (MRN ?MRN)
        (first-name ?first-name)
        (surname ?surname)
        (gender ?gender))
    (physical-feature
            (HPO-ID ?HPO-ID)
        (physical-feature-description ?physical-feature-description))
    (mode-of-inheritance
            (HPO-ID HP:0000006)
        (mode-of-inheritance-description Autosomal-dominant-inheritance))
    (age-of-onset-classifier
            (childhood ?childhood))
=>

(assert (IBMFS-DB (cps-db DBA))))

In defining the facts, should I write it as an assert command?

I tried it, but I face an error with the defrule syntax "[PRNTUTIL2] Syntax Error:  Check appropriate syntax for defrule.

ERROR:
(defrule MAIN::case-1
   (patient (age ?age) (MRN ?MRN) (first-name ?first-name) (surname ?surname) (gender ?gender))
   (physical-feature "

Any idea about this?

Regards,

Razan

[CLIPS Support]

One possible explanation is that you didn't define a physical-feature deftemplate:

         CLIPS (6.4 2/9/21)
CLIPS> (defrule case-1

   (physical-feature
      (HPO-ID ?HPO-ID)
      (physical-feature-description ?physical-feature-description))

    =>)

[PRNTUTIL2] Syntax Error:  Check appropriate syntax for defrule.

ERROR:
(defrule MAIN::case-1

   (physical-feature (
CLIPS> 

You've only provided code fragments in your questions, so I can only guess as to what you did wrong in your complete set of code.

[Razan Abu-Khashabeh]

Please see attached, the deftemplates, deffacts and deffrules.

Can you please check them and let me know.

Thanks,

Razan

[CLIPS Support]

If all of the deftemplates are defined before the rule, the rule loads without error.

CLIPS>
(deftemplate patient
   (slot age)
   (slot MRN)
   (slot date-of-birth)
   (slot first-name)
   (slot surname)
   (slot gender))
CLIPS>
(deftemplate physical-feature
   (slot phenotypic-abnormality)
   (slot HPO-ID)
   (slot clinical-investigation-ID)
   (slot physical-feature-description))
CLIPS>
(deftemplate mode-of-inheritance
   (slot pattern-of-inheritance)
   (slot HPO-ID)
   (slot mode-of-inheritance-description))
CLIPS>
(deftemplate age-of-onset-classifier
   (slot infantile)
   (slot childhood)
   (slot juvenile)
   (slot adult))
CLIPS>
(deftemplate IBMFS-DB (slot cps-db))
CLIPS>

(defrule case-1
   (patient
       (age ?age)
       (MRN ?MRN)
       (first-name ?first-name)
       (surname ?surname)
       (gender ?gender))
   (physical-feature
       (HPO-ID ?HPO-ID)
       (physical-feature-description ?physical-feature-description))
   (mode-of-inheritance
      (HPO-ID HP:0000006)
      (mode-of-inheritance-description Autosomal-dominant-inheritance))
   (age-of-onset-classifier
      (childhood ?childhood))
   =>
   (assert (IBMFS-DB (cps-db DBA))))

CLIPS> 

[Razan Abu-Khashabeh]

I tried it as you mentioned. It succeeded for the first rule. However, for the others, it gave the following error: [TMPLTDEF1] Invalid slot 'HPO-ID-is' not defined in corresponding deftemplate 'laboratoryscreening-and-hematologicalfinding' for defrule case-2 & case-4. [TMPLTDEF1] Invalid slot 'HPO-I-is' not defined in corresponding deftemplate 'physical-feature' for defrule case-5.

[CLIPS Support]

Every single time you reference a slot in a deftemplate pattern, it has to be defined in the corresponding deftemplate or you'll get an error message. If you searched for HPO-ID-is in your source files, you'd see that it's referenced by patterns but it's never defined in a deftemplate. The error message is literally telling you the place in your code to look. In your rules, you're matching the slot HPO-ID-is, but your deftemplate defines the slot as HPO-ID.

[Razan Abu-Khashabeh]

I correct it, but in your opinion, what is the error in the deftemplate patterns in this defrule:

(defrule case-2

          (patient
            (age ?age)
        (MRN ?MRN)
        (first-name ?first-name)
        (surname ?surname)
        (gender ?gender))
           (physical-feature
            (HPO-ID ?HPO-ID)
        (physical-feature-description ?physical-feature-description))

    (laboratoryscreening-and-hematologicalfinding
            (HPO-ID HP:0001903
                HP:0012410
            HP:0001896
            HP:0001972))
        (age-of-onset-classifier
            (infantile ?infantile))

    (mode-of-inheritance
                  (HPO-ID HP:0000006)
              (mode-of-inheritance-description Autosomal-dominant-inheritance))

=>
(assert(IBMFS-DB (cps-db DBA))))

[PRNTUTIL2] Syntax Error:  Check appropriate syntax for deftemplate patterns.

Thank you for your help. I am not from IT Background. Therefore, I am facing such problems. 

Is there a possibility for a ZOOM meeting in which we can communicate faster?

Regards,

[CLIPS Support]

Post your entire program. Do not post it as a pdf or image file that cannot be directly loaded into CLIPS. Post the actual ASCII text files that you are loading.

[CLIPS Support]

If you look at the error message, it shows you the point at which the syntax error occurs. This is similar to the first error you posted, only in this case instead of treating a single field slot as if it could contain no values, you're treating it as if it could contain two.

Defining defrule: case-2

[PRNTUTIL2] Syntax Error:  Check appropriate syntax for deftemplate patterns.

ERROR:
(defrule MAIN::case-2

   (patient (age ?age) (MRN ?MRN) (first-name ?first-name) (surname ?surname) (gender ?gender))
   (physical-feature (HPO-ID ?HPO-ID) (physical-feature-description ?physical-feature-description))
   (laboratoryscreening-and-hematologicalfinding (HPO-ID HP:0001903 HP:0012410

On Wednesday, January 18, 2023 at 4:35:24 PM UTC-6 rabukh...@gmail.com wrote:

[Razan Abu-Khashabeh]

Please see attached.

Regards,

Razan

[Razan Abu-Khashabeh]

Any suggestion for the deftemplate patterns that should be?

[Razan Abu-Khashabeh]

What is the correct way to write the deftemplate pattern "(laboratoryscreening-and-hematologicalfinding
            (HPO-ID HP:0001903)

                HP:0012410
            HP:0001896
            HP:0001972))"

in 

(defrule case-2
          (patient
            (age ?age)
        (MRN ?MRN)
        (first-name ?first-name)
        (surname ?surname)
        (gender ?gender))
           (physical-feature
            (HPO-ID ?HPO-ID)
        (physical-feature-description ?physical-feature-description))
    (laboratoryscreening-and-hematologicalfinding

            (HPO-ID HP:0001903)

                HP:0012410
            HP:0001896
            HP:0001972))

        (mode-of-inheritance
                  (HPO-ID HP:0000006)
              (mode-of-inheritance-description Autosomal-dominant-inheritance))
        (age-of-onset-classifier

            (infantile ?infantile))
=>
(assert(IBMFS-DB (cps-db DBA))))

?

[CLIPS Support]

Fixes for typos and cardinality issues with slot definitions are attached. The remaining errors, referencing variables that have never been assigned a value, are semantic.

[Razan Abu-Khashabeh]

Thanks, I added more rules and all the errors have been overcome. Attached, while running the program. In order to make sure that all the rules are going to be fired, all the deftemplate patterns should have facts related to, am I right?

[CLIPS Support]

Yes, you would need to test each rule with a facts matching the patterns in the rules.

[Razan Abu-Khashabeh]

May I have your opinion about my work?

Please see attached.

[CLIPS Support]

Your program executes 3780 rules, but does nothing as all of the facts asserted in the actions of your rules are already present because they are included in deffacts.

[Razan Abu-Khashabeh]

I removed them from the deffacts. However, I got the same result. See attached.

[Ronald Finkbine]

I have built a large Clips expert system to search graphs, recognizing software algorithms that have come from compilers, for my doctoral dissertation.

My project worked and worked well, So I feel qualified to issue an opinion about this overall project. 

Correction, my Version 2 worked well, Version 1 was built crappy and I finally realized it and took 6 days to start over and build in a designed way.

The author states he is not an IT guy. Well, you NEED one very much.

I had to build my expert system in very small pieces, the foundation pieces, then grouping sub-searches into larger ones.

I had 110 different test cases to verify small pieces of my matches.

When I saw something in this thread that said 3780 rules in system. I spit out my coffee.

You ask "What is wrong?"

Heck, can you show me one thing that is RIGHT? Verified CORRECT? Paper and pencil checkmarks and list of facts before/after a run of a few rules?

Incrementalism. Test-driven development.

And when you edit the code even one little character, you have invalidated ALL of the previous test cases executions. Regression testing is necessary.

You need to be able to run ALL the test cases in a batch environment from one command.

My 110 test cases took 45 minutes to run, I used to jog out to the railroad tracks and come back later during a regression run.

Save all your code, create a new EMPTY project, add back SMALL pieces with some test cases that you can see are accurate on like 2 sheets of paper. Not 3780.

Gary Riley, thanks for all your efforts over the years. It has been much appreciated.

Ronald Finkbine
ronaldf...@hotmail.com

---

From: clip...@googlegroups.com <clip...@googlegroups.com> on behalf of Razan Abu-Khashabeh <rabukh...@gmail.com>
Sent: Friday, January 27, 2023 5:50 PM
To: CLIPSESG <clip...@googlegroups.com>
Subject: Re: defrules in CLIPS

 

--
You received this message because you are subscribed to the Google Groups "CLIPSESG" group.
To post to this group, send email to CLIP...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/CLIPSESG?hl=en
 
--> IF YOU NO LONGER WANT TO RECEIVE EMAIL <--
Visit this group at http://groups.google.com/group/CLIPSESG?hl=en
Click on "Edit my membership" link.
Select the "No Email" radio button.
Click the "Save these settings" button.

--> IF YOU WANT TO UNSUBSCRIBE <--
Visit this group at http://groups.google.com/group/CLIPSESG?hl=en
Sign in
Click on "Edit my membership" link.
Click the "Unsubscribe" button.
Note: This appears to be the most reliable way to unsubscribe
 
Alternately, send email to CLIPSESG-u...@googlegroups.com. You will receive an email which you must respond to as well to unsubscribe. Clicking the link mentioned in the unsubscribe reply does not appear to work reliably.
---
You received this message because you are subscribed to the Google Groups "CLIPSESG" group.
To unsubscribe from this group and stop receiving emails from it, send an email to clipsesg+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/clipsesg/a127354e-329f-4e30-9953-15dab47b20c0n%40googlegroups.com.

[CLIPS Support]

You get a few facts asserted by your rules. The problem with your program appears to be that the facts and results need to be patient specific, but they're not, so you're getting thousands of invalid activations for your rules. 

         CLIPS (6.4 2/9/21)
CLIPS> (load "CLIPS Program.clp")
%%%%%%%%%%%$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$********************
TRUE
CLIPS> (reset)
CLIPS> (watch statistics)
CLIPS> (watch facts)
CLIPS> (run)
==> f-47    (IBMFS-DB (cps-db DBA))
==> f-48    (IBMFS-FA (cps-fa FA))
==> f-49    (IBMFS-SDS (cps-sds SDS))
3780 rules fired        Run time is 0.00253701210021973 seconds.
1489941.65210037 rules per second.
49 mean number of facts (49 maximum).
0 mean number of instances (0 maximum).
1890 mean number of activations (3780 maximum).
CLIPS> 

[Razan Abu-Khashabeh]

Thank you for your reply.

Can you please give me an example how the results and facts being as you mentioned?

[CLIPS Support]

It's your program; you should be telling us what it's supposed to be doing. You shouldn't start coding until you have a few examples of what the program should be doing and why. That includes examples of input and expected output and the explanation for how the output is derived. If you can't explain it to another person, then you're going to have problems coding it regardless of what programming language you're using. And if you're unfamiliar with a programming language, don't try to write the entire program in one fell swoop. Write one small part of it first and make sure that part works before adding more code. 

[Ronald Finkbine]

What is your most simple rule?

Any rule with just one template as a CE?

This would be the most simple TC001 (Test case 001).

Get one small example together that will fire and print "THIS RULE WORKS".

Then repeat. Add one more template and another rule and make that TC002.

Repeat.

And again.

You cannot just put all of this together and expect it to work at first.

It has to be engineered.

People who are good at this are called Knowledge Engineers.

Ronald Finkbine
ronaldf...@hotmail.com

---

From: clip...@googlegroups.com <clip...@googlegroups.com> on behalf of Razan Abu-Khashabeh <rabukh...@gmail.com>

Sent: Wednesday, February 1, 2023 5:31 PM

To: CLIPSESG <clip...@googlegroups.com>
Subject: Re: defrules in CLIPS

--

You received this message because you are subscribed to the Google Groups "CLIPSESG" group.
To post to this group, send email to CLIP...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/CLIPSESG?hl=en
 
--> IF YOU NO LONGER WANT TO RECEIVE EMAIL <--
Visit this group at http://groups.google.com/group/CLIPSESG?hl=en
Click on "Edit my membership" link.
Select the "No Email" radio button.
Click the "Save these settings" button.

--> IF YOU WANT TO UNSUBSCRIBE <--
Visit this group at http://groups.google.com/group/CLIPSESG?hl=en
Sign in
Click on "Edit my membership" link.
Click the "Unsubscribe" button.
Note: This appears to be the most reliable way to unsubscribe
 
Alternately, send email to CLIPSESG-u...@googlegroups.com. You will receive an email which you must respond to as well to unsubscribe. Clicking the link mentioned in the unsubscribe reply does not appear to work reliably.
---
You received this message because you are subscribed to the Google Groups "CLIPSESG" group.
To unsubscribe from this group and stop receiving emails from it, send an email to clipsesg+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/clipsesg/2d22dd34-b398-4c15-b92f-c0c643ee83d5n%40googlegroups.com.

[Razan Abu-Khashabeh]

What do you mean by patient specific?

This program is about diagnosing four disorders and the rules are about the approach(es) for diagnosing them. The facts the case when a patient has certain abnormalies associated with each of these disorders.

Regards,

[CLIPS Support]

Your program has over 3000 rule firings that mostly do nothing. The only way that seems plausible is if you've got data for thousands of patients mixed together. Your programming methodology makes it practically impossible to try to figure out what you're doing. I can tell that your program is doing something medical, but that's it.

[Razan Abu-Khashabeh]

My Program contains only the constructs. I have not loaded the patients data yet.

I have to load the patients data then check the activations. right?

[CLIPS Support]

Here's your 1200+ line program where I've cut out 95% of your code to illustrate the benefit of the advice you've been ignoring:

         CLIPS (6.4 2/9/21)

CLIPS>
(deftemplate patient
   (slot age)
   (slot MRN)
   (slot date-of-birth)
   (slot first-name)
   (slot surname)
   (slot gender))
CLIPS>

(deftemplate mode-of-inheritance
   (slot pattern-of-inheritance)
   (multislot HPO-ID)
   (multislot mode-of-inheritance-description))

CLIPS>
(deftemplate age-of-onset-classifier
   (slot infantile)
   (slot childhood)
   (slot juvenile)
   (slot adult))
CLIPS>

(deftemplate IBMFS-SDS
   (slot cps-sds))
CLIPS>
(deffacts CPS-4
   (patient
     (age 0)
     (MRN 236748)
     (first-name Rani)
     (surname Khateeb)
     (date-of-birth 09/10/2022)
     (gender male))
   (age-of-onset-classifier
      (infantile Infantile-onset))
   (mode-of-inheritance
      (HPO-ID HP:0001419)
      (mode-of-inheritance-description X-linked-recessive-inheritance)))
CLIPS>
(deffacts CPS-5
   (patient
      (age 9)
      (MRN 204561)
      (first-name Rani)
      (surname Khateeb)
      (date-of-birth 4/10/2014)
      (gender male))
   (age-of-onset-classifier
      (childhood Childhood-onset))
   (mode-of-inheritance
      (HPO-ID HP:0000007)
      (mode-of-inheritance-description Autosomal-recessive-inheritance)))
CLIPS>
(defrule case-7

   (patient
      (age ?age)
      (MRN ?MRN)
      (first-name ?first-name)
      (surname ?surname)
      (gender ?gender))

   (age-of-onset-classifier
      (childhood ?childhood))
   (mode-of-inheritance
      (HPO-ID HP:0000007)
      (mode-of-inheritance-description Autosomal-recessive-inheritance))
   =>
   (assert (IBMFS-SDS (cps-sds SDS))))
CLIPS> (reset)
CLIPS> (facts)
f-1     (patient (age 0) (MRN 236748) (date-of-birth 09/10/2022) (first-name Rani) (surname Khateeb) (gender male))
f-2     (age-of-onset-classifier (infantile Infantile-onset) (childhood nil) (juvenile nil) (adult nil))
f-3     (mode-of-inheritance (pattern-of-inheritance nil) (HPO-ID HP:0001419) (mode-of-inheritance-description X-linked-recessive-inheritance))
f-4     (patient (age 9) (MRN 204561) (date-of-birth 4/10/2014) (first-name Rani) (surname Khateeb) (gender male))
f-5     (age-of-onset-classifier (infantile nil) (childhood Childhood-onset) (juvenile nil) (adult nil))
f-6     (mode-of-inheritance (pattern-of-inheritance nil) (HPO-ID HP:0000007) (mode-of-inheritance-description Autosomal-recessive-inheritance))
For a total of 6 facts.
CLIPS> (agenda)
0      case-7: f-1,f-2,f-6
0      case-7: f-4,f-2,f-6
0      case-7: f-4,f-5,f-6
0      case-7: f-1,f-5,f-6
For a total of 4 activations.
CLIPS> 

You haven't provided any information about what you're expecting your program to do and why, again ignoring the advice you've been given, but my guess would be that you're expecting a single activation of case-7 for the patient with MRN 204561 because you're assuming that putting a group of facts in the same deffacts somehow forces the rule to keep those facts together when pattern-matching occurs, but that's clearly not what's happening when you look at the activations for the rule. The mode-of-inheritance with Autosomal-recessive-inheritance is being matched for both patients as well as the age-of-onset-classifier. If that's not what you want to happen and the MRN is unique for each patient, then what you need to do is add an MRN slot to the age-of-onset-classifier and mode-of-inheritance so you have a means of associating the right information with the right patient.

[Razan Abu-Khashabeh]

May I ask what is the difference between the fatcs "the initial knowledge" and the patients specific data?

[Razan Abu-Khashabeh]

The confusion I have now is How to assess the effectiveness of my expert system using the real data.

Is the data to be used instead of the facts?

[CLIPS Support]

What do you think is the data and what do you think is the facts? What is the extent of your knowledge? What documentation have you read? Which parts of the documentation do you understand?

[Razan Abu-Khashabeh]

My understanding is in order to develop Expert System the three constructs "deftemplates, facts and rules" should be loaded then the effectiveness of the expert system should be assessed by sufficient and representative data.

I understand that the three main components of the expert system are: Facts (Working"short" Memory), Rules (Productive "long term" Memory) and the Inference Engine. Should I understand that the initial facts are initial knowledge which then to be replaced by real facts "the real data"?

[CLIPS Support]

Read through the CLIPS User's Guide. That will familiarize you with basic terminology and explain how constructs such as deffacts work.

[Razan Abu-Khashabeh]

I did read the userguide. Also, I read EPERT SYSTEMS Principles and Programming. I developed the expert system based on my readings from both

[CLIPS Support]

OK, that would have been the appropriate response when I specifically asked you what documentation you read. When you're asking questions about brain surgery, it helps to let people know if you're a brain surgeon, another type of surgeon, a doctor, a nurse, a layman, or someone who has no knowledge of basic human anatomy. You're asking question about CLIPS as if your only exposure to it is a 15 minute powerpoint presentation. One would assume that if you'd read through the User's Guide and ES:P&P, you'd have a basic understanding of the difference between a slot and a multislot in a deftemplate and that the deffacts construct is just a series of facts that get asserted whenever a reset command is performed. It's still not clear to me whether you have any prior experience programming in another language or if you developed this program from scratch or modified one  you found online, and that's important information to know if you're asking questions about testing. No one is going to spend their time developing a personalized training course for you to learn how to program. Even if you're not a software professional, on most forums there's going to be an expectation that you ask questions in a semi-professional manner, and if you don't do that people are going to quickly lose patience with you. Asking vague unanswerable questions with no evidence of research is a good way to do that. If you don't know how to do software testing, don't go on a forum and ask how you do software testing. That's lazy. If you use a search engine, there's lots of resources online you can find that will explain the basics of software testing. You can then use that knowledge to ask specific directed questions about testing. 

Learning how to ask a good question is the single most important thing you can learn to become a good programmer because all the things needed for someone else to answer your question for you are the same things you need to answer it for yourself. Here's a good place to start: https://stackoverflow.com/help/how-to-ask     

[Ronald Finkbine]

You should be testing ONE RULE at a time, not dozens.

Multiple templates, and multiple facts, ok.

But not rules.

Ronald Finkbine
ronaldf...@hotmail.com

---

From: clip...@googlegroups.com <clip...@googlegroups.com> on behalf of Razan Abu-Khashabeh <rabukh...@gmail.com>

Sent: Sunday, February 26, 2023 11:04 AM

To: CLIPSESG <clip...@googlegroups.com>
Subject: Re: defrules in CLIPS

--

You received this message because you are subscribed to the Google Groups "CLIPSESG" group.
To post to this group, send email to CLIP...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/CLIPSESG?hl=en
 
--> IF YOU NO LONGER WANT TO RECEIVE EMAIL <--
Visit this group at http://groups.google.com/group/CLIPSESG?hl=en
Click on "Edit my membership" link.
Select the "No Email" radio button.
Click the "Save these settings" button.

--> IF YOU WANT TO UNSUBSCRIBE <--
Visit this group at http://groups.google.com/group/CLIPSESG?hl=en
Sign in
Click on "Edit my membership" link.
Click the "Unsubscribe" button.
Note: This appears to be the most reliable way to unsubscribe
 
Alternately, send email to CLIPSESG-u...@googlegroups.com. You will receive an email which you must respond to as well to unsubscribe. Clicking the link mentioned in the unsubscribe reply does not appear to work reliably.
---
You received this message because you are subscribed to the Google Groups "CLIPSESG" group.
To unsubscribe from this group and stop receiving emails from it, send an email to clipsesg+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/clipsesg/5cf6e324-ac41-4e3d-ad39-35b0dc5d7ce8n%40googlegroups.com.