Extra vertical space between TABLEs when viewed in Safari

Hi Joe

I made a test page with your code & it renders without any visible gap using Safari 2.0.4 on my mac, a quick screenshot at
http://www.windywoo.f9.co.uk/tablepic.tiff





iMac G5 rev B (OSX 10.4.7) 2GB 250GB

I too was able to get the tables to be joined together with no gaps.

I have seen this before, especially when the cells contains some graphics. I cannot fully explain what the cause is, but one way I have been able to solve it in the past is through using one of the styles: margin, margin-right, margin-left, margin-top, or margin-bottom. Use a negative value such as -1 or -2 and see if that helps.

Example would be:<pre>
<TABLE cellpadding=0 cellspacing=0 border=0 bgcolor="blue">
<tr><td>content here...</td></tr>
</TABLE>

<TABLE cellpadding=0 cellspacing=0 border=0 bgcolor="blue" style="margin-top:-2;">
<tr><td>content here...</td></tr>
</TABLE>
</pre>

Figures... as soon as I clicked on the Post Message button I thought of a way to illustrate this better... Here is a new example. Notice the <BR /> between the tables. The larger negative adjustment closes the gap as if the <BR /> is not there.

New Example:<pre>
<TABLE cellpadding=0 cellspacing=0 border=0 bgcolor="blue">
<tr><td>content here...</td></tr>
</TABLE>
<br />
<TABLE cellpadding=0 cellspacing=0 border=0 bgcolor="blue" style="margin-top:-18;">
<tr><td>content here...</td></tr>
</TABLE>
</pre>

The attribute cols on the TABLE tag is valid (in some historical standard/browser), but not part of the W3.org's standards for HTML v4.1.

I did find that there was a proposal in a document dated May 1996 through w3c that did suggest its usage. It was in RFC 1942: http://www.ietf.org/rfc/rfc1942.txt

Apparently the proposal was discussing the pros and cons of a two pass versus a one pass algorithm for table rendering, and the use of rows would have been a "hint" to allow the one pass algorithm to be used.

Unfortunately, I have a feeling that if a programmer puts a value of 4 in there but then adds another column it caused too much trouble where if it would then have to fall back on the two pass algorithm, then it would have wasted too much time.

I can almost remember the days when you did have to worry about performance. It appears like performance issue now a days has the general trend of "lets just make it fast enough and then wait for the next generation of processors and hardware to be released". Could one almost say that the computer sector of the economy is partially fueled by laziness to write efficient code? 😉

JoeCuv, thanks for pulling out that little oddity. Interesting where computers and their languages have traveled in the last 10 years!