∂26-MAY-73  1348		1,JMJ
 LARRY,
 MAYBE I'M SPELLING YOUR NAME WRONG. YOU DIDN'T EXIST AT PARC FOR
 THAT LAST MESSGE EITHER. IS IT MASINTER AT BOTH ISI AND PARC?
 ⎇
 
∂25-MAY-73  1429		206,MJC
 ILISP.MAC[UCI,WD] IS CURRENT HERE.  THE SEGMENT STUFF HAS BEEN 
 CHANGED FROM THE ORIGINAL.  
Net mail from site BBN rcvd at 23-MAY-73  2314 PDT
-------
Date: 24-MAY-73 0217-EDT
From: BROWN at BBN-TENEX
Re:   LOADER
cc:    MASINTER  at ISI
- - - -
LARRY,
IN REWRITING YOUR MACRO CODE, ****DON'T**** INDEX THRU S. PUTSYM, GETSYM,
*PUTSYM AND *GETSYM DO NOT HANDLE RELOCATION IN THIS MANNER.
DO PUTSYMS ON THE ATOMS YOU NEED TO REFERENCE. THEN LOAD YOUR
RELOC CODE. YOU SHOULD LOAD IT INTO BPS. ****ONCE YOU
HAVE LOADED THE CODE, YOU CANNOT EXPAND BPS******. THE ONLY
REASON OFR THE S INDEXING IS SO THAT BPS CAN BE EXPANDED. AT THIS
POINT I DO NOT KNOW WHAT IS IN THE LISP SYMBOL TABLE FILE
(LLSP.SYM).NOTE THAT THE RU OR ST COMMANDS *ALWAYS*CLOBBER
THE SYMBOL TABLE(POSSIBLY INCORRECTLY). ANYWAY, I WILL LET YOU KNOW MORE
AS SOON AS I CAN. AS THINGS LOOK AT THE MOMENT, THE LOADER
IS FOR ALL PURPOSES
THE SAME AS STANFORD'S. I'LL LET YOU KNOW ABSOLUTELY TOMORROW,
BUT PUTSYMS WILL DO WHAT THEY DID AT STANFORD.
JEFF
-------
Net mail from site BBN rcvd at 24-MAY-73  0017 PDT
-------
Date: 24-MAY-73 0320-EDT
From: BROWN at BBN-TENEX
Re:   MACRO CODE
cc:    MASINTER  at ISI
- - - -
LARRY,
OK THINGS STRAIGHT.PREVIOUS MESSAGE ESSENTIALLY CORRECT. LOADER
TE SAME AS STANFORD'S. THE SYMBOLS YOU CAN ACESS ARE THOSE
DECLARD INTERN (VIA  THE MKENT MACRO) WHICH DON'T GO ANYWHERE,
OF COURSE. AS FOR PUTSYM, WHAT I SAID HOLDS TRUE...
I'LL BE AROUND FOR A COUPLE MORE HOURS IF YOU HAVE ANY MORE TROUBLE.
(IT'S 12:20 PST)
JEFF
PS. FYI, WRITING A RELOC PUTSYM TO AVOID THE PROBLEM THAT IS
CAUSED BY THE ABOVE.
-------
Net mail from site BBN rcvd at 26-MAY-73  1339 PDT
Net mail from site BBN rcvd at 26-MAY-73  1346 PDT
-------
Date: 26-MAY-73 1648-EDT
From: BROWN at BBN-TENEX
Re:   LOADER
cc:    MASINTER  at PARC-MAXC,  MASINTER  at ISI
- - - -
LARRY,
IS ALL GOING WELL? I AM ADDING TWO NEW FUNCTIONS, *RPUTSYM
AND *RGETSYM. THESE WILL ENABLE THE USER TO  WRITE CODE
USING THE INDEX REGISTER. AVOIDS PROBLEM OF NOT BEEING ABLE
TO EXPAND BPS AFTER LOADING. IF YOU ARE INTERESTED, I CAN GIVE YOU THE
CODE (IT'S SHORT). ALSO, I HAVE HAD TROUBLE SENDING YOU MAIL
FROM BBN. YOU DON'T EXIST AT ISI. CAN YOU GIVE ME YOU #'S AGAIN?
                JEFF
-------
Net mail from site BBN rcvd at 28-MAY-73  1236 PDT
-------
Date: 28-MAY-73 1537-EDT
From: BROWN at BBN-TENEX
Re:   LOADER
cc:    MASINTER  at USC-ISI,  MASINTER  at PARC-MAXC
- - - -
LARRY,
RE MSG TODAY... THERE IS A MESSAGE AT SU-AI ON THIS SUBJECT.
YOUR METHOD USING THE DIFFERENCE BETWEEN THE SYMBOL AND REG S
WILL WORK **IF** YOU CHANGE ARITH.MAC SO THAT THE REFERNCES
TO  FLONUM AND FIXNUM ARE THRUOUGH THAT REGISTER (I AM MAKING THE
CHANGE CURRENTLY).
 IN REFERENCE TO PROBLEM 2 , YES REALLOCATION WIPES OUT THE
SYMBOL TABLES AND THE BUFFERS. I AM LOOKING AT CHANGING THINGS SO
THAT THE USER CAN REALLOCATE FROM INSIDE LISP AND PRESERVE
THESE THINGS (IT'S) IMPOSSIBLE TO USEE DDT TO DEBUG THE THING
IN MOST CASES (PUTTING THE TABLE ELSEWHERE HAS PROBLEMS ALSO).
YOU SHOULD NOTE THAT REALLOCATING ONLY SCREWS YOUR LOADED CODE
IF YOU ALLOCATE MORE **BPS**.
 I WANT TO THANK YOU. FOR A LONG PERIOD OF TIME, THE ONLY BUGS
OR PROTESTS THAT HAVE OCCURED HAVE BEEN DUE TO THE WORK OF THREE
PEOPLE. SENDING IT OUT HAS HELPED SHAKE DOWN THE SYSTEM AND YOU
HAVE CONRIBUTED THE LION'S SHARE OF PROBLEMS, BUGS ETC.
THOSE OF US WHO HAVE BEEN WORKING ON IT (PRIMARILY MYSELF
DARYLE LEWIS, WHO IS NOW AT BBN) HAVE GOTTEN INOT THE BAD HABIT
OF SYTEM'S PROGRAMMING FOR OUR SAKES. IT'S KIND OF NICE
TO HAVE SOME USER'S AROUND WHO WANT SOMETHING ELSE
AND KNOW WHAT THEY WANT. I'LL BE AROUND  HERE TILL THE 17 OF JUNE.
HOPE TO MEET YOU AT THE AI CONFERENCE, IF YOU'LL BE THERE. I
MAY STILL DECIDE NOT TO COME BACK HERE FOR GRAD SCHOLL NEXT
YEAR AND SEE IF I CAA FIND A JOB IN PALO-ALTO. LATER...
JEFF
-------