> Um, I noticed the same thing as Brett did. I think it *is* related with the
> size of cells. I have quite large database (TRMASTER.ASC size is 1.7 MB),
> with 'S5' cell being the largest, and there are disappearing calls right
> from that cell. That doesn't wonder me, as Pascal allows just 64 kB large
> cells, so larger ones are obviously truncated.
I think Mirko is on the right track here.
The obvious thing in common with Chip (K7JA) and Jay (VY1JA) is the JA in their
callsign... which is often the largest cell in the database. The other thing
they have in common, is that they are well known, and you will notice when
they are missing.
There is something less obvious about the calls however.
Each of these calls resolve to the following cells:
K7 7J JA
and
VY Y1 1J JA
The program saves cells first from AA to // going through the letters
first, then numbers.
So - for each of the above calls, JA will be the first cell saved to an
ACSII file. When the other cells come up, the program will not save them
again. There is no need to save a call more than once in the ASCII file.
If the JA cell is overcapacity (total information saved > 64K), the calls
at the end of the JA list will be missing (as Mirko pointed out). These
would be the JA calls that appear towards the end of the TRMASTER.ASC file.
Calls that have been edited will appear at the end of the list and will be
the ones that fall out if the cell is over capacity.
So - if this theory is true, someone should be able to change K7JA or VY1JA
to something that has a cell that will show up before JA - and the call
will not be lost. Then, they can change it back and it should get lost.
Like A7JA or VE1JA.
That's my theory and I am sticking to it...
It isn't likely that I can do much about this in the program. Manually
checking each call to see if it has already been written would really slow
down the operation. I'll ponder it for awhile to see if anything comes to
mind.
In the mean time, I predict you will only notice this for people who are
famous enough for you to remember - that have JA in their call, and don't
have any letters in their call from A to I with a letter or number after it.
In Mirko's case - with many S50 to S59 calls - he has lots of opportunity.
S5#x Where # is any number and x is any letter would be first saved as
S5.
Furthurmore, S5#Tx to S5#Zx would also have S5 as its first cell.
Tree
PS: I have another theory... I call it my theory #2...
|