NavList:

   

Reply

> The application of the Hollerith tabulating machine to Brown's tables of
> the moon
> 



That article says, "A Hollerith installation was used in H.M. Nautical
Almanac Office for seven months in 1929; actually punching was started
six months before the arrival of the sorter and tabulator, as it was
necessary to punch 20 000 000 holes in half a million cards. The work
described on long and short period nutation, and on double entry tables,
as well as that of most of the single entry tables, was carried to the
year 2000. The greater part of the cost was incurred in doing the first
ten years, which would have sufficed for immediate needs. But to
continue for the next 55 years with a trained and organised staff added
very little to the cost, and was certainly more economical than
re-training and re-organising ten years later. Moreover, there is little
likelihood of Brown's Tables being superseded before the end of the
century; any acquisition to our knowledge of the Moon during the next
seven decades is almost certain to be expressed in the form of
correction to Brown's Tables, not in the form of new tables."

Comrie wasn't far from the truth, though the Brown tables were overtaken
by progress sooner than he anticipated. According to Standish et al in
the Explanatory Supplement to the Astronomical Almanac (1992), prior to
1984 "the lunar ephemeris was calculated directly from E.W. Brown's
algorithm instead of from his Tables of Motion of the Moon (1919). To
obtain a strictly gravitational ephemeris expressed in the measure of
time defined by Newcomb's Tables of the Sun, the fundamental arguments
of Brown's tables were amended by removing the empirical term and by
applying to the Moon's mean longitude the correction [formula omitted].
In addition, this ephemeris was based on the IAU (1964) System of
Astronomical Constants, and was further improved in its precision by
transformation corrections. The expressions for the mean longitude of
the Moon and its perigee were adjusted to remove the implicit partial
correction for aberration."

The article says a Hollerith card is 7 3/4 by 3 1/4 inches. Possibly the
length is a mistake, since the "IBM card" (as it was often called) was 7
3/8 inches long. US Government checks are still that size.

The 45° cut at one corner makes it easy to verify all cards in a stack
are oriented identically.

When I arrived in SAC in the early 80s the B-52 bombing / navigation
system included a small digital computer which was obviously a later
addition to a basically analog system. It had a magnetic core memory
which occasionally got corrupted and had to be reloaded from a stack of
about 65 IBM cards. You inserted a card in a slot, turned a handle 90°
to read the card, turned the handle back to release the card, pulled it
out, and on to the next card. It was a tedious and unfriendly process.
If any card was not in numerical sequence the device did not give you an
opportunity to fix the mistake. The load aborted and you had to go back
to Card 1. Grrr.

--
Paul Hirose
sofajpl.com

   

Reply