Re: Frames still broke in NSN 3?

Stephanos Piperoglou wrote:
> Both of you are getting this entirely wrong. As a matter of fact,
NS is the
> most correct in this respect. The problem is not border, padding
and spacing
> but *WHITESPACE* which is collapsed into a single space character
between
> the images.

"I see," said the blind carpenter, as he picked up his hammer and
saw.

Thanks for making this clear. Closing tags were suggested to me
earlier, but I had adopted the style of putting a newline before them
when they close a container and so they had no effect.

The source of the problem would have been apparent had it not been
for the space added between ROWS. Why is an equal amount of space
being added in the vertical dimension?

Try this and come up with a logical explanation of NSN's result:

<TABLE CELLPADDING=0 CELLSPACING=0 BORDER=0>
<TR>
<TD><IMG SRC="image.gif" BORDER=0></TD>
<TD><IMG SRC="image.gif" BORDER=0></TD>
<TR>
<TD><IMG SRC="image.gif" BORDER=0>
<TD><IMG SRC="image.gif" BORDER=0></TD>
</TABLE>

There's no </TD> on cell R2C1, so the two columns have a space
between. In the first column the images butt vertically, but the
image in cell R2C2 is dropped about 1/2 the intercolumn space.

If I add </TD> to cell R2C1 and remove it from cell R2C2, the images
butt horizontally and the images in C2 butt vertically. But the image
in R2C1 is now dropped by the same amount as the image in R2C2 was in
the previous example.

The logic eludes me.

The only reason for collapsing whitespace is to allow structured
markup. In this regard, MSIE's practice of dropping trailing space
entirely makes sense. A helluva lot more sense that whatever NS is
doing, anyway.

David Perrell

Received on Wednesday, 28 August 1996 09:38:58 UTC