tables for layout bugs question

[Gary Peek]

Can anyone tell us the browsers/versions that exhibit errors
when tables are nested too deeply? And how many levels of
nesting produces errors?

(not a tables vs CSS question)

[Spartanicus]

Gary Peek <my********@mycompanyname.comwrote:

>Can anyone tell us the browsers/versions that exhibit errors
when tables are nested too deeply? And how many levels of
nesting produces errors?

You are confused, using tables for layout causes problems, not errors.

>(not a tables vs CSS question)

But it is.

--
Spartanicus

[Zilbandy]

On Wed, 01 Nov 2006 14:58:05 +0000, Spartanicus
<in*****@invalid.invalidwrote:

>>Can anyone tell us the browsers/versions that exhibit errors
when tables are nested too deeply? And how many levels of
nesting produces errors?

You are confused, using tables for layout causes problems, not errors.

>>(not a tables vs CSS question)

But it is.

So, what you're saying is you don't have an answer to his question. :/

--
Zilbandy - Tucson, Arizona USA <zi*@zilbandyREMOVETHIS.com>
Dead Suburban's Home Page: http://zilbandy.com/suburb/
PGP Public Key: http://zilbandy.com/pgpkey.htm
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~

[Spartanicus]

Zilbandy <zi*@zilbandyREMOVETHIS.comwrote:

>>>Can anyone tell us the browsers/versions that exhibit errors
when tables are nested too deeply? And how many levels of
nesting produces errors?

You are confused, using tables for layout causes problems, not errors.

>>>(not a tables vs CSS question)

But it is.

So, what you're saying is you don't have an answer to his question. :/

What you are demonstrating is that you've not understood neither the
question, the answer, or usenet. Oh well.

--
Spartanicus

[Gary Peek]

Spartanicus wrote:

You are confused, using tables for layout causes problems, not errors.

Very, well, can anyone tell us the browsers/versions that exhibit
problems when tables are nested too deeply? And how many levels of
nesting produces these problems?

[Spartanicus]

Gary Peek <my********@mycompanyname.comwrote:

>You are confused, using tables for layout causes problems, not errors.

Very, well, can anyone tell us the browsers/versions that exhibit
problems when tables are nested too deeply? And how many levels of
nesting produces these problems?

Using a table for layout purposes causes problems in all browsers,
nesting tables makes it worse.

Search the archives of this group and it's stylesheet sister group to
find out what problems (it's been discussed at length at least a few
hundred times before).

--
Spartanicus

[VK]

Very, well, can anyone tell us the browsers/versions that exhibit

problems when tables are nested too deeply? And how many levels of
nesting produces these problems?

Are you having some problems, expecting some problems or just curious?
One day many years ago - being bored and in a bad mood - I managed to
hang up NN4.5 on 8th level. But as I forced a Flash movie in the
top-right cell of each table that experiment was not clear. :-)

Purely theoretically I wouldn't expect problems at least until 32-bit
overflow which means 4,294,967,295 nested tables. But the chances are
high that long before that it will trig UA's AID (Anti-Idiot Defence
:-). With the modern DOM methods you can easily conduct an "on denial"
test yourselve by dynamically generating and inserting tables onto your
page.

[Gary Peek]

VK wrote:

Are you having some problems, expecting some problems or just curious?
One day many years ago - being bored and in a bad mood - I managed to
hang up NN4.5 on 8th level.

Somewhat out of curiosity, but I was playing around with a
nested table layout, example at:
http://www.garypeek.com/web/tablemadness.htm

I had heard that some older browsers choked on things like this,
but what wondering how deep it had to be to cause this.

I'm not planning on using something this deep, but I don't
have any heartburn using something less complicated.

Do you think I'll piss off anyone if I ask how far down the
text needs to be before the search engines won't find it? :)

[Ben C]

On 2006-11-01, VK <sc**********@yahoo.comwrote:

>Very, well, can anyone tell us the browsers/versions that exhibit
problems when tables are nested too deeply? And how many levels of
nesting produces these problems?

Are you having some problems, expecting some problems or just curious?
One day many years ago - being bored and in a bad mood - I managed to
hang up NN4.5 on 8th level. But as I forced a Flash movie in the
top-right cell of each table that experiment was not clear. :-)

Purely theoretically I wouldn't expect problems at least until 32-bit
overflow which means 4,294,967,295 nested tables.

You could get problems much sooner than that if the table formatter
implementation is recursive (and doesn't have an AID). Eventually you
will run out of stack space.

But the chances are high that long before that it will trig UA's AID
(Anti-Idiot Defence
:-). With the modern DOM methods you can easily conduct an "on denial"
test yourselve by dynamically generating and inserting tables onto your
page.

This would be a good test to try, since a stack overflow is a security
vulnerability.

[VK]

Somewhat out of curiosity, but I was playing around with a

nested table layout, example at:
http://www.garypeek.com/web/tablemadness.htm

It's appropriately named "tablemadness" so I hope you didn't take it as
a design idea of any kind ;-)

I had heard that some older browsers choked on things like this,
but what wondering how deep it had to be to cause this.

Well, at the time of Browser Wars the speed and the reliability in
rendering table cells was one of major promo points of both Netscape
and Microsoft. There were special table test suites consisting of
hundreds colored cells in multi-nested tables and both periodically
organized a la Formula-1 races :-) by watching how quickly their
browser will fill the page if the competitor chocks on it or behind.
The results (if beneficial) were proudly announced on the company site.
Crazy times :-) Anyway I would say that the hardest way to nock out any
of post-war browsers is by using tables. All lesser-effective
algorithms were sorted out and eliminated long time ago.

I'm not planning on using something this deep, but I don't
have any heartburn using something less complicated.

If you decide for some reason to use table layouts, your heartburn has
to start automatically on the third level and should not be remedied by
any pils but only by re-thinking your layout. :-) :-|
The only exception is if you are a QA tester in a browser making
company.

Do you think I'll piss off anyone if I ask how far down the
text needs to be before the search engines won't find it? :)

As down as it's needed to find a long enough continuous text. How long
- that is engine specific, I don't know.

[Zilbandy]

On Wed, 01 Nov 2006 15:33:34 +0000, Spartanicus
<in*****@invalid.invalidwrote:

>>>You are confused, using tables for layout causes problems, not errors.

(not a tables vs CSS question)

But it is.

So, what you're saying is you don't have an answer to his question. :/

What you are demonstrating is that you've not understood neither the
question, the answer, or usenet. Oh well.

About his question: His question dealt with nesting tables and not
with CSS.

About your answer: Tables can cause problems . . . ok, I'll agree.
However, tables, whether politically correct in today's preferred web
page layout, are still a part of many web pages. Your reply to his
question, although technically correct, was totally useless.

About my knowledge of Usenet: I'm an average PC user who has been
using Usenet since about 1995. Whether I understand ALL there is to
know about Usenet is not the issue; to that, the answer is no.
However, I'm not a total "newbie" to its concept or its use. Oh, by
the way, Usenet should be spelled starting with a capital letter, so I
guess you're not quite the total "know it all" you believe you are. Oh
well.

--
Zilbandy - Tucson, Arizona USA <zi*@zilbandyREMOVETHIS.com>
Dead Suburban's Home Page: http://zilbandy.com/suburb/
PGP Public Key: http://zilbandy.com/pgpkey.htm
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~

[Spartanicus]

Zilbandy <zi*@zilbandyREMOVETHIS.comwrote:

>>What you are demonstrating is that you've not understood neither the
question, the answer, or usenet. Oh well.

About his question: His question dealt with nesting tables and not
with CSS.

Take the false assumption out of the question and the real question the
OP asked was about HTML table vs so called "CSS layouts", this does
however require a capacity to see beyond the question as phrased.

>About your answer: Tables can cause problems . . . ok, I'll agree.
However, tables, whether politically correct in today's preferred web
page layout, are still a part of many web pages. Your reply to his
question, although technically correct, was totally useless.

Answering a question that is based on a false premise not only doesn't
help anyone, it creates confusion. The useful thing to do is to point
out the erroneous assumption. This is good communication practice, also
on usenet, especially in this hierarchy.

--
Spartanicus

[Pete Verdon]

Spartanicus wrote:

Zilbandy <zi*@zilbandyREMOVETHIS.comwrote:

>Spartanicus wrote:

>>What you are demonstrating is that you've not understood neither the
question, the answer, or usenet. Oh well.

>About his question: His question dealt with nesting tables and not
with CSS.

Take the false assumption out of the question and the real question the
OP asked was about HTML table vs so called "CSS layouts",

Actually, he not only didn't ask about CSS, he explicitly stated that he
*wasn't* asking about 'HTML table vs so called "CSS layouts"':

: Can anyone tell us the browsers/versions that exhibit errors when
: tables are nested too deeply? And how many levels of nesting produces
: errors?
:
: (not a tables vs CSS question)

Compare the last line with the last of your lines quoted above.

Pete

[Zilbandy]

On Thu, 02 Nov 2006 17:20:08 +0000, Spartanicus
<in*****@invalid.invalidwrote:

>>About his question: His question dealt with nesting tables and not
with CSS.

Take the false assumption out of the question and the real question the
OP asked was about HTML table vs so called "CSS layouts", this does
however require a capacity to see beyond the question as phrased.

So, if I ask how deep I can nest tables, your answer will be to use
Cascading Style Sheets? Maybe the misunderstanding has to do with
which side of the pond we live on. :) I'm done with this topic.

--
Zilbandy - Tucson, Arizona USA <zi*@zilbandyREMOVETHIS.com>
Dead Suburban's Home Page: http://zilbandy.com/suburb/
PGP Public Key: http://zilbandy.com/pgpkey.htm
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ~~~~~~~~~

[Spartanicus]

Zilbandy <zi*@zilbandyREMOVETHIS.comwrote:

>>Take the false assumption out of the question and the real question the
OP asked was about HTML table vs so called "CSS layouts", this does
however require a capacity to see beyond the question as phrased.

So, if I ask how deep I can nest tables, your answer will be to use
Cascading Style Sheets?

I would answer that it is rarely appropriate to nest tables if used for
their intended purpose, that the question suggests that tables are being
used inappropriately, that this causes problems and that the real
solution is to stop doing that.

If you asked how to shoot yourself through the foot, I'd tell you that
doing so is not smart, not explain how to load the gun. This is a time
honoured tradition by the knowledgable regulars in this hierarchy.

--
Spartanicus

[Chris Morris]

Zilbandy <zi*@zilbandyREMOVETHIS.comwrites:

On Thu, 02 Nov 2006 17:20:08 +0000, Spartanicus

Take the false assumption out of the question and the real question the
OP asked was about HTML table vs so called "CSS layouts", this does
however require a capacity to see beyond the question as phrased.

So, if I ask how deep I can nest tables, your answer will be to use
Cascading Style Sheets? Maybe the misunderstanding has to do with
which side of the pond we live on. :) I'm done with this topic.

I've seen table layout (as in the process of laying out a table, not
the use of a table or tables to layout a page) have (minor but
visually ugly) bugs in all modern graphical browsers, for a
sufficiently complex data table [1] with no nesting at all and
entirely valid HTML code. If there are bugs at nesting depth 0, then
it obviously follows that there are also bugs at all greater nesting
depths.

All complex tables are going to have layout bugs, at the
extremes. Nesting makes it more likely that the complexity will pass
the 'critical point'. So, for example, the table generated by the
pseudocode:
for i in [1..10000] {
print ("<table><tr><td>");
}
print ("Content");
for i in [1..10000] {
print ("</td></tr></table>");
}
is likely to have bugs (despite all individual tables involved being simple)

Conversely, a sufficiently complex table can have bugs with no nesting
at all.

So your question is unanswerable. (In practice, there's no need to
nest tables at all in the vast majority of situations, and no need to
nest them more than one level deep in the extremely small minority
remaining, so that's not a major problem)

[1] colspans on almost every cell in most rows, and two columns per
month over a 10+ year period.
For the extremely small 'intranet-like' audience who were interested
in the data, it was the best way to present it. Of course, I'd
recommend for general web use breaking the table down into several
two smaller ones - maybe one per year.

--
Chris