COMMENT ⊗   VALID 00005 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	∂08-Sep-82  1528	lark@Diablo (SuNet)  	atsign functions  
C00004 00003	∂09-Sep-82  1224	<CSD.DIETTERICH at SU-SCORE> 	"Do before 5:00 today" and other tidbits
C00007 00004	∂ 9 Sep 1982 1338-PDT	<CSD.DIETTERICH at SU-SCORE>	comments on REPN
C00010 00005	∂TO CSD.DIETTERICH@SCORE 16:27 9-Sep
C00013 ENDMK
C⊗;
∂08-Sep-82  1528	lark@Diablo (SuNet)  	atsign functions  

Here is a partial list of functions that I have compiled.  It is probably
not complete.

 mergecar	@mergecar
 nunion	@nunion
 merge	@merge
 memq	@memq
 nunionq	@nunionq
 nmerge	@nmerge
 nmergecar	@nmergecar
 unionq	@unionq
 unite	@unite
 nunite	@nunite
 union	@union
 mem	@mem
 assq	@assq
 assoc	@assoc
 ass	@ass
 memassoc	@memassoc
 memassq	@memassq
 memass	@memass
 rplaca	@rplaca
 rplacd	@rplacd
 last	@last
 intersection	@intersection
 intersectionq	@intersectionq
 intersect	@intersect
 nintersection	@nintersection
 nintersectionq	@nintersectionq
 nintersect	@nintersect
 synonym	@synonym
 error	@error
 mapc	@mapc
 mapl	@mapl
 mapcar	@mapcar
 maplist	@maplist
 mapcan	@mapcan
 mapcon	@mapcon
 do	@do
 do*	@do*
 prog	@prog
 prog*	@prog*
 get	@get
 getl	@getl
 putprop	@putprop
 defprop	@defprop
 remprop	@remprop
 plist	@plist
 pop	@pop
 push	@push
 if	@if
 when	@when
 unless	@unless
 caseq	@caseq
 pairp	@pairp
 listp	@listp
 symbol	@symbol
 make-hash-table	@make-hash-table
 gethash	@gethash
 puthash	@puthash
 remhash	@remhash
 catch	@catch
 throw	@throw

∂09-Sep-82  1224	<CSD.DIETTERICH at SU-SCORE> 	"Do before 5:00 today" and other tidbits
To: MRS-documenters: ;

Hello everyone,

A couple of things:

1. Please make a list of all of the dictionary terms that you think
you are responsible for.  Put them in a msg to me, and also bring them
to this afternoon's meeting.

2.  Please make a list of terms that you think you want someone else
to do.  Russ has already started a file called "orphans" in
<csd.mrs.manual> of such terms.  Feel free at edit this file and claim
some of these terms as your very own or add terms that you have found
at your doorstep.

3.  Keep the following points in mind when writing entries--especially
the description part:

  a.  Use @T around anything that is a "computer word" or computer
      i/o.

  b.  Capitalize NIL.

  c.  When referring to the arguments passed to a function, use the
      phrase from the Usage entry, surrounded by <>'s.  This usually
      makes it easier to write and less ambiguous.

  d.  In writing your "brief", try to make the phrase fit into the
      following sentence:  "In order to <brief>, I should type
      <name>."  For example, "In order to get or set the cannonical
      coreferential symbol I should use coref."  This makes the
      <brief> more like "the goal that this satisfies" rather than a
      summary of "what" the command does.

  e.  In general, try to give more "why's" in addition to all of the
      "what's".  For instance, tell why $truepn treats all free
      variables as existentials.  Especially important for obscure or
      bizarre "what's".


Thanks,

-Tom
-------

∂ 9 Sep 1982 1338-PDT	<CSD.DIETTERICH at SU-SCORE>	comments on REPN
To: csd.greiner at SU-SCORE
cc: CSD.DIETTERICH at SU-SCORE

Russ,

I have the following comments in reading your REPN file.

1.  Your "brief" statements are not brief enough.  For example, I
would have said the brief for PDT would be
"discrimination net representation for a particular theory"

or maybe just

"discrimination net representation"

If there is no other discrimination net that we are trying to
discriminate this one from.

2.  In general, we aren't capitalizing function names like REPN.
How do you feel about this.  I know in the Interlisp environment, this
is important...

3. We are putting @T's around any computer-ese, including the function
name in the Usage box.

4. What's a "tertiary proposition" and a "tertiary relation"? (see
your PPT entries).  Do you mean "ternary"?

5. What is FIing (see diagnositics for REPN).?

6. In general, this stuff is very dense.  The section needs a general
introduction describing what's "going on".  I didn't even try to read
all of the stuff at the end of the file.  We need some discussion of
the relative trade-offs of the different representations.  I guess a
lot of that belongs in the HOW-TO part of the manual.

--Tom
-------
-------

∂TO CSD.DIETTERICH@SCORE 16:27 9-Sep
Tom -
0) Thanks first for your suggestions wrt my REPN section.  I'll do what I can
to "operationalize" them.

1) Thanks also for creating that orphan file, and publicizing(sp) it.
Perhaps there should be a general Q/A file as well, containing suggestions
such as the ones you made in your most recent general message,
questions like the one which follows (and, I guess, this point as well...,)
and the notes now cluttering your SKELETON file.
Another entry would be a description of the desired contents of the history slot:
many people still seem to think it refers to this dictionary entry, rather than
to the term itself.
...or do these things belong in <CSD.MRS>MAIL.TXT?
...but enough file proliferation schemes for now...

2) What should go in the SEE-ALSO slot:
the minimal set of *immediate* pointers, or everything which may be relevant?
Eg, I want everyone reading about PDT-STASH to know about PDT-LOOKUPN.
Is it sufficient for PDT-STASH to point to PDT,
exploiting the fact that PDT does SEE-ALSO this PDT-LOOKUPN?
Or should I put that PDT-LOOKUPN value in PDT-STASH's SEE-ALSO slot?

3) My terms  (tentative -- subject to Andy's comments):

LISP procedures:
  PDT-STASH PDT-UNSTASH PDT-LOOKUPN PDT-LOOKUPVALN
  PPT-STASH PPT-UNSTASH PPT-LOOKUPN PPT-LOOKUPVALN
  PP-STASH  PP-UNSTASH  PP-LOOKUPN  PP-LOOKUPVALN
  PM-STASH  PM-UNSTASH  PM-LOOKUPN  PM-LOOKUPVALN
MRS Constants (Representations):
  PDT, PPT, PP, PM
MRS Relations:
  REPN, USE-FN, DOMAIN, OPDOMAIN, ARITY

General terms/ideas:
  Representation  		<-- I'd rather MRG did this.
	[This is an Overview - what is "representation", and how does MRS deal
	 with it.]
  Representation Naming Strategy
  Discrimination Net
  Property List
  Indexing Strategy

---
Russ