Re: attribute forbidden but needed?

2012-03-26 19:54, Wilfried Hanisch wrote:

> Am Montag 26 März 2012 17:57:47 schrieb Andreas Prilop:
>> On Sun, 25 Mar 2012, Wilfried Hanisch wrote:
>>> <frameset frameborder="0" cols="280,*,40"  rows="*">
>>> what is wrong?
>> Frames are wrong.
>> Don't bother with validating frames.
>> Get rid of frames!
> This is not the answer I have expected.

Nobody expects the Spanish inquisition!

It was the most the useful answer, though. To maintain a site, you don't 
really need to validate it. To rewrite it, you really don't need frames. 
If you validate, just accept the error messages, after analyzing that 
they do not reveal any *real* errors, and go forward.

 > My question was more factual then religous.

What was factual? It might be regarded as a fact that the validator 
reports an error when it detects a reportable markup error. What was 
your question about this? The markup is wrong, as judged on the basis of 
the document type definition against which it was validated.

> I did no read before that frames are forbidden.

You may have missed the news from the late 1990s that they are 
hopelessly outdates and were never a good idea. I might slightly 
disagree on this, saying that one site out of 10,000 could make 
meaningful use of frames, but I'm in a minority.

> I have adopted the site recently and before I redesign it I want to make it conform to W3C

That would be rather pointless, would it not? Painting a house just 
before tearing it down.

> Does anyone other be able to give a factual and kind answer?

Another factual and kind answer is that if you really want to keep using 
frames, for now, and use a validator, you could use the frameborder=0 
attribute on the <frame> elements, where it validates in HTML 4.01 Frameset.

The cost is that this method of removing borders works less often than 
the frameset="0" attribute on the <frameset> tag, as mentioned at
But that's fairly old information, and on modern browsers, there's 
probably not much difference.

On the other hand, the only thing that you would achieve is that an 
error message does not appear in the validator's report. And the only 
reasons why this would matter is that someone (who does not understand 
the issues) pays you for making pages valid and that you expect the 
frameset page to contain other errors, so that removing one of the 
messages makes it marginally easier to focus on the others.


Received on Monday, 26 March 2012 17:29:23 UTC