Satgen313   Millennial Software Problems by GM4IHJ           25 March 95
BID of this msg is SGEN313 Use this BID if you retransmit msg on packet

It may seem a trifle early to think about the next thousand years. But the
new millennia starts in just over 4 years , and it calls for serious prior
study by satellite software writers and users, and sat information
providers such as NASA.

    A typical software problem will arise where the program asks the user
to input the date as day/month/year, where the year number is presently
given as 95, ie not the full number 1995. This minor difference can mean
that as Jan 1st 2000 dawns , the entry of either 00 or 2000 will have
disastrous effects. This problem occurs because the shorthand usage of 95
for 1995 is usually offset inside the program with a line which say
something like :-
   LET YEAR NUMBER = YEAR NUMBER + 1900
In fact the only answer which will work in this situation is to enter the
year number as 100 . Which is going to be too illogical for most folks to
swallow. So quite a few programs in wide circulation are going to have to
be amended some time before the millennia.

     Equally at risk are the Keplerian Elements, where even NASA use a
system which may produce errors in year 2000. At present Keplerian
elements give the Epoch time as say 95087.375 = year 1995 day 87, decimal
day .375 . The question is will NASA use the format  00001.5 for Epoch Jan
1st year 2000 at utc noon. If they do, direct input to several modern
programs could cause problems.

     So what is to be done ? Clearly, the earlier software is checked to
ensure it will survive in the 21st century, the better. No one should buy
software from now on which is not guaranteed to work after Jan 1 2000 .
Amsat organisations should be appraising their members of this problem and
should start telling members which software is 21st century proof and
which is not.

     HOWEVER it should be clear from the previous paragraphs that software
writers need to know what NASA will be doing with 21st century Keplerian
elements. Probably they will stick with just two digit year numbers, but
if they do a great deal of software will need amendment. A situation which
will cause a lot of problems, although strangely enough, as far as this
writer is aware, there has been no discussion anywhere about this topic.
Perhaps we should ask NASA now ? Then forewarned and forearmed we will be
able to give a very cheerful greeting as we see in this very special
millenial New Year 2000, secure in the knowledge, that when next we are
sober enough to look for RS12 , our software will not deceive us.

PS Next weeks satgen will discuss ways to check whether your software is
millenia proof. 73 de GM4IHJ @ GB7SAN or gm4ihj@delphi.com