I had an MS Word 97 document which contained a huge table, so I figured I'd
better split it into four smaller files. I did so, assigning to each of
them about a quarter of the rows in the table. To my surprise, each of the
four "smaller" files was nearly the same size as the original "big" one.

So I conducted an experiment. I created a new document containing a single
line (not a table):
   new -
I closed that and examined its size in Windows Explorer and in DOS, then
reopened the file, added the sizes to the first line and started another,
so the document now contained:
   new - 19K - 19456
   open1 -

What? 19K just for starters? I repeated the experiment, extending the file
line by line. To make a long story short, successive Windows Explorer
readings were:
    19K - 29K - 28K - 28K - 28K - 30K - 28K - 39K - 37K - 37K - 39K - 41K -
43K -
    46K - 48K - 51K - 19K - 29K - 28K - 28K - 28K - 33K - 32K - 34K.

Notice the recycling of 19K after 51K, and the two runs of three 28K's.

By MS Word's own "statistics" count, the file contained 471 characters
(with blanks) after the last line was added.

The corresponding DOS readings more or less backed up Explorer's story,
though there were anomalies, such as the pair 51K - 52224.

An analogous experiment with Notepad held the Explorer count to 1K all the
way through, with a DOS count that crept up in an expected way to 398
characters after the same number of lines.

Any ideas what is going on, and how it can be avoided?

-- Dick H. Fredericksen, Tucson, AZ

                                  -----
                PCSOFT mailing list -  http://nospin.com
         Bob Wright:[log in to unmask] - Drew Dunn:[log in to unmask]