Conference noted::hackers_v1

    i just always assumed that it came from fenceposts which divide
    property.  is the fencepost the end of your property or the
    beginning of you neighbors property?  just like in a loop counter
    where you aren't sure whether you want to stop just "before" the end
    condition or just "at" the end condition.
    
    i could be wrong though.
    

    Could be its from the common mistake of getting one too
many posts when building a closed fence or not getting enough
when building a open fence.

    I could be wrong though.
    
..Dan

From "The Hacker's Dictionary" Steele, Woods,Finkel etc.:

FENCEPOST ERROR noun.

An "off by one" error: the discrete equivalent of a boundary condition.

This problem is often exhibited in programs containing iterative loops: 
something will be done one time too few or too many. The term comes from the 
following problem: "If you build a fence 100 feet long with posts 10 feet 
apart, how many posts do you need?" (Either 9 or 11 is a better answer than the
obvious 10)



Graham Pye.

PS Maybe the other derivation is that you feel like hitting your head against a 
fence post once you spot the error!

If two neighbors were quibbling about exactly where their land boundary
lay, they might argue that a fencepost is a slight distance the wrong
way, into one person's property or the other.

Although such a slight amount, it might make a BIG difference to
quarreling neighbors.

Similarly, a range check in a computer program that is supposed to
check for I<1000 and erroneously checks for I<=1000 instead, can make
a BIG difference.

For example, suppose the program is finding the minimum value from a table.
If it uses the I<=1000, it will check one element too many, and that
"non-element" may be 0, which might be less than any actual element seen,
so program prints out 0.

/Eric

    Eric,  "you used to have a truck like that"??
    
    	That reminds me of an old joke about two land owners in
    Texas.  You're not letting your sense of humor show, are you??
    
    -mike

    
    In the UK we talk about telegraph poles rather than fence poles
    but the Hackers dictionary explaination is the one that is correct
    (i.e. if you count spaces between then that is one less than the number
    of poles)
    
    However I have never heard the phrase used in the context of loop
    counters being one out though I shall now!
    
    -Mike-
    

    In the dark ages, the monks of the Venerable Bede had an expression
    'fens post error' (a Vulgate latin term from the latin: 'fenus post
    errorem', meaning 'profit after error'). This was applied to the
    situation where someone made a mistake in their accounts, and somehow
    the maker of the mistake came out showing a profit (I have heard
    rumours that this sometimes happens to people when they do their
    expences). Of course the most common such mistake was one that could
    reasonably be seen to be the sort that anyone could make - eg be off by
    one in a complex calculation. [It should be remebered that calculations
    with only roman numerals were quite difficult, especially
    multiplication and division]. See also 'A Canticle For Leibowitz, pub.
    Harper and Row for the Perennial Library, 1986; P. 321 (paperback ed)
    for further detail. 

(hope this helps)
Pete

    	I like the explanation in .8 best. Can we request that DEBUG
    and SDA only display in Roman  numerals, to discourage amateur hackers?
    Maybe we could get all of this MCS/French/Norwegian business scrapped
    if we just standardised on *the* international language - Latin.

    .8 sounds like folk etymology to me, especially since it comes from an
    after-the-holocaust novel. This genre often creates amusing folk
    etymologies to be witty. Do you have a more traditional reference?
    
    	Jon

    re .9
    
    TOPS-20 has a monitor call ("system service" to VAXoids) that
    will output a number in most popular radices.  There was once a
    real SPR from a customer which requested Roman Numeral output as
    an option.

    re .11:
    
    And we are going to implement said feature as soon as we implement
    fire and smoke :-) :-) :-)
    
    Mike

    I thought we had a product or two that resembled that remark. 
    
    :-)
    
    - M

    Doesn't DSR do Roman Numerals?
    
    Someone could always bug the developers for an FAO argument for
    them. (%RN?)
    
    /pjh

TeX does Roman Numerals.  I think nroff (Unix formatter) does, too.

There was a request for DECtalk to convert them, but we were able
to squirm out of it.

M.

    	For sure VAX Lisp does too.  Both flavors of Roman Numerals.
    
    -mike z

    	You mean there isn't OTS$CVT_L_TR ?
    
    			;-}

I don't really want to admit it, but VAX DOCUMENT (or whatever we end up calling
 it) can do Roman Numerals too.
  <_Jym_>

Re .11:

>There was once a real SPR from a customer which requested Roman Numeral
>output as an option.

If I recall correctly, calling the author of that SPR a "customer" really
strains the term to the limit.
				/AHM