# (j3.2006) Retro cool to go with your Fortran compiler

Kurt W Hirchert hirchert
Thu Dec 16 22:13:54 EST 2010

```The 12-inch diameter circular slide rule was a bit more manageable and
was nominally good for 5 digits.  It had a conventional circular slide
rule scale around the outside (a little less than a meter long), but the
real accuracy came from a much longer scale that spiraled inwards.  It
could be read to 5 digits, but you had to know which winding of the
spiral to read, so in practice, you often need to compute things twice,
once using the outer scale to get the first three digits and then a
second time using the spiral scale to get the additional digits by
looking a the winding that included the three digit value.  [I didn't
have one of these, but someone in my high school did, so I got a chance
to see how it worked.]

-Kurt

P.S. Personally, I always felt that the typical straight slide rule was
good for 3 digits on the right end and 4 on the left, but I admit that
depended on being able to visually interpolate the last digit between
two marked position.  If I think back to how well I did such
interpolation, perhaps a claim of 2.7 to 3.7 significant digits would be
closer to the truth.  Yes, that was and is sufficient for a wide range
of computations relating to physical measurements or physical
properties.  (It may be great that your calculator can display 10 digits
of precision, but if your computation is based on measurement that are
only good for 3 digits, most of those computed digits are meaningless.)

On 12/15/2010 2:14 PM, Jerry Wagener wrote:
> Ahhh... that's great, Van (pretty tough to carry around on your belt, though :).    -Jerry
>
> On Dec 15, 2010, at 2:09 PM, Van Snyder wrote:
>
>> On Wed, 2010-12-15 at 04:58 -0800, Jerry Wagener wrote:
>>> alas, one can compute with it to only about two significant figures
>>> (in some cases you can interpolate close to three significant figures)
>>> - guess that was good enough for much early engineering work.
>> There was an observatory in Germany that had a 20-meter slide rule with
>> a magnifying cursor.  It was reputedly good for six digits.
>>
>>
>> _______________________________________________
>> J3 mailing list
>> J3 at j3-fortran.org
>> http://j3-fortran.org/mailman/listinfo/j3
>
> _______________________________________________
> J3 mailing list
> J3 at j3-fortran.org
> http://j3-fortran.org/mailman/listinfo/j3
>

```