Using "shape" attribute withour "coords" in an "a" tag

[Albert Wiersch]

I know it is technically valid, but why would one use the "shape" attribute
with an "a" tag but not use the "coords" attribute too? Is there any logical
reason for doing this?

 

[Jukka K. Korpela]

I know it is technically valid,

Are you sure? Did you use a real validator?

but why would one use the "shape"
attribute with an "a" tag but not use the "coords" attribute too?

Beats me. Have you seen someone do that? The effect is undefined.
But maybe someone intends to set the coords attribute via client-side
scripting. That would normally be a poor idea, of course.

 

[Benjamin Niemann]

I know it is technically valid, but why would one use the "shape"
attribute with an "a" tag but not use the "coords" attribute too? Is there
any logical reason for doing this?

shape="default" specifies the entire region (See
<http://www.w3.org/TR/html4/struct/objects.html#adef-shape>), hence no need
for explicit coordinates.

 

[Albert Wiersch]

"Albert Wiersch" <[email protected]>
wrote:

Beats me. Have you seen someone do that? The effect is undefined.
But maybe someone intends to set the coords attribute via client-side
scripting. That would normally be a poor idea, of course.

Yes, it came up when someone checked the W3C page with CSE HTML Validator
and brought it to my attention. CSE HTML Validators shows this as an error
because it doesn't make logical sense. The page is:
http://www.w3.org/

The above page sets the "shape" attribute to "rect" but does not use the
"coords" attribute. They do this numerous times.

I was wondering why they would do that. It seems there is no logical reason
why and that the W3C should remove the "shape" attribute in this situation
to "clean up the page". Please correct me if there is a reason for doing
this that I am missing.

By the way, I would prefer not to get into the argument of the definition of
"HTML Validator" again.

 

[David Håsäther]

Yes, it came up when someone checked the W3C page with CSE HTML
Validator and brought it to my attention. CSE HTML Validators
shows this as an error because it doesn't make logical sense. The
page is: http://www.w3.org/

Well, it's not an error, as you will see if you try it in a validator
;-)

The default value for the shape attribute is "rect", so if you don't
specify it, it will be inferred, making

<a href="...">...</a>

and

<a href="..." shape="rect">...</a>

exactly the same.

As Benjamin Niemann said, shape="default" specifies the entire region,
so it can make sense to use shape but not coords.

The above page sets the "shape" attribute to "rect" but does not
use the "coords" attribute. They do this numerous times.
I was wondering why they would do that.

I have no idea. Maybe the tool they used inserts all default values.
Looks like they used HTML Tidy - maybe it has an option to do that.

 

[Albert Wiersch]

Well, it's not an error, as you will see if you try it in a validator
;-)

Yes, I know it is another issue that is technically valid but is not good
syntax/form.

As Benjamin Niemann said, shape="default" specifies the entire region,
so it can make sense to use shape but not coords.

Yes, but only when it is shape="default", not shape="rect".

I have no idea. Maybe the tool they used inserts all default values.
Looks like they used HTML Tidy - maybe it has an option to do that.

Maybe... but it seems like bad form that shouldn't be used.

 

[Neredbojias]

With neither quill nor qualm, Albert Wiersch quothed:

I know it is technically valid, but why would one use the "shape" attribute
with an "a" tag but not use the "coords" attribute too? Is there any logical
reason for doing this?

Maybe they're not quite sure about the shape of things to come.

 

[Jukka K. Korpela]

Yes, it came up when someone checked the W3C page with CSE HTML
Validator and brought it to my attention. CSE HTML Validators shows
this as an error because it doesn't make logical sense.

This just proves that your product is not a validator and that it spits out
messages that tell your opinions on various matters.

The above page sets the "shape" attribute to "rect" but does not use
the "coords" attribute. They do this numerous times.

Of course, the attribute shape="rect" has no effect if the <a> element is
not actually used in the context of an image map. It's simply futile, not
an error. Adding a coords attribute without changing anything else, as your
"validator" seems to be doing, would just add another futile attribute.

I was wondering why they would do that.

Probably because they use authoring software that likes to produce shape
attributes. But is that your problem? Your problem is that your product
reports an error when there is none, and you are even trying to make this a
case _for_ your product. And you _first_ made your product report something
as an error, and apparently only then did you start asking whether it is an
error.

By the way, I would prefer not to get into the argument of the
definition of "HTML Validator" again.

There was never an argument. People were just reminded of the fact that you
sell a product as a validator when it is not.

 

[Albert Wiersch]

"Albert Wiersch" <[email protected]>
wrote:

This just proves that your product is not a validator and that it spits
out
messages that tell your opinions on various matters.

Pretty good idea when a real validator doesn't catch so many things that
don't make sense, are futile, are against good form, or just plain wrong
and/or likely to cause problems with real-world browsers. Besdies, like I
said before, my opinions are mainly based on the W3C recommendations,
real-world experience, and user experience.

Of course, the attribute shape="rect" has no effect if the <a> element is
not actually used in the context of an image map. It's simply futile, not
an error. Adding a coords attribute without changing anything else, as
your
"validator" seems to be doing, would just add another futile attribute.

Yes, that makes sense. I can probably improve the messages.

Probably because they use authoring software that likes to produce shape
attributes. But is that your problem? Your problem is that your product
reports an error when there is none, and you are even trying to make this
a
case _for_ your product. And you _first_ made your product report
something
as an error, and apparently only then did you start asking whether it is
an
error.

I don't really care to make a case for my product (that's not why I asked),
but it should be clear how limited a real validator is and that HTML linters
and checkers can find many issues that real validators can't. This is just
one of the cases where a "non-real" validator found what you would call a
"futile" attribute. Obviously it is not good form to use futile attributes.
I don't think you'd disagree with that.

Anyone, it seems my original question was answered well enough. In the case
of the w3.org page, the "shape" attribute was technically valid to use but
"futile", so it was bad form. It was futile because there was no "coords"
attribute and because it wasn't used with an image map.

 

[Jukka K. Korpela]

but it should be clear how limited a real validator is

It performs a well-defined job, as opposite to looking here and there and
making proposals and issuing error messages based on someone's opinions.

and that HTML
linters and checkers can find many issues that real validators can't.

You mean like the non-issue that your phoney validator reports as an error?
It's sad that there are no good HTML linters (actually never were), but
it's probably too late now. That, however, is not a reason to pay for a
broken checker called, in an intentionally misleading way, a validator.

This is just one of the cases where a "non-real" validator found what
you would call a "futile" attribute. Obviously it is not good form to
use futile attributes.

That's your opinion. It is based on limited experience with the practice of
HTML authoring. And it is certainly objectively wrong to call a futile
attribute an error when it in fact conforms to any relevant specification.

I don't think you'd disagree with that.

Then you are wrong.

I see little reason to remove futile attributes, if a document contains
them as by-products of some HTML generator or as holdovers from some
previous version of the document where they were not futile, or maybe in
preparation for the next version that will drop the "f" from futility.

 

[Albert Wiersch]

It performs a well-defined job, as opposite to looking here and there and
making proposals and issuing error messages based on someone's opinions.

Which is fine if you want to limit yourself to what is "well-defined" and
ignore other potential issues like bgcolor="dfsdgfdsfgdfg".

You mean like the non-issue that your phoney validator reports as an
error?
It's sad that there are no good HTML linters (actually never were), but
it's probably too late now. That, however, is not a reason to pay for a
broken checker called, in an intentionally misleading way, a validator.

It seems like you've never used my validator, or don't understand it (or,
more likely, you can't get over the fact that it doesn't conform to your
chosen definition of validator). If it points out an issue that is not
really an issue, then feel free to bring it to my attention. I may change
it. You can also configure it as you want on a per-message basis.

That's your opinion. It is based on limited experience with the practice
of
HTML authoring. And it is certainly objectively wrong to call a futile
attribute an error when it in fact conforms to any relevant specification.

I don't limit the definition of "error" to what is only technically wrong.
Again, it just seems you have a problem with the name of the program and
"error" which means, according to the dictionary "a mistake" or "deviation
from what is correct", among other definitions.

I see little reason to remove futile attributes, if a document contains
them as by-products of some HTML generator or as holdovers from some
previous version of the document where they were not futile, or maybe in
preparation for the next version that will drop the "f" from futility.

Then you can disable the message in CSE HTML Validator and use all the
futile attributes you want in hopes that they become useful, or for whatever
other reason you want.

Anyhow, I'm not going to continue this argument. It's lost its usefulness to
the group.

 

[Richard Rundle]

Which is fine if you want to limit yourself to what is "well-defined" and
ignore other potential issues like bgcolor="dfsdgfdsfgdfg".


It seems like you've never used my validator, or don't understand it (or,
more likely, you can't get over the fact that it doesn't conform to your
chosen definition of validator). If it points out an issue that is not
really an issue, then feel free to bring it to my attention. I may change
it. You can also configure it as you want on a per-message basis.
specification.

I don't limit the definition of "error" to what is only technically wrong.
Again, it just seems you have a problem with the name of the program and
"error" which means, according to the dictionary "a mistake" or "deviation
from what is correct", among other definitions.


Then you can disable the message in CSE HTML Validator and use all the
futile attributes you want in hopes that they become useful, or for whatever
other reason you want.

Anyhow, I'm not going to continue this argument. It's lost its usefulness to
the group.

On the contrary. If it's not a validator, just don't call it a validator.
Not hard, is it ?

 

[Jukka K. Korpela]

Which is fine if you want to limit yourself to what is "well-defined"
and ignore other potential issues like bgcolor="dfsdgfdsfgdfg".

I'd rather have a validator that performs some useful checking objectively
than a phoney validator that cannot be trusted.

It seems like you've never used my validator,

As I have told previously, I have used it enough to see what it is, and you
yourself provide further evidence.

or don't understand it

Which one of us is knows HTML, seriously?

(or, more likely, you can't get over the fact that it doesn't conform
to your chosen definition of validator).

I wouldn't get over your calling a cow a horse, no matter how often you
claim it's just about my private definition of a horse.

If it points out an issue that
is not really an issue, then feel free to bring it to my attention. I
may change it.

How noble.

Now you have confirmed that the product you sell produces just messages
according to your taste.

I don't limit the definition of "error" to what is only technically
wrong.

Why does this remind me of the question how to define "sex"? You seem to
use the word "technically" just to call a cow a horse. After all, we should
not be limited to the technical definition of a horse, so if a cow sells
better as a horse...

Well, it's just part of your making your personal opinions the ultimate
criterion on HTML.

Anyhow, I'm not going to continue this argument.

You never answered why you keep calling a non-validator a validator, except
by admitting that it is not "technically" (i.e., in reality) a validator.

It's lost its usefulness to the group.

Au contraire, it is useful to people to know what your product really is.
We can also conjecture that it is pointless to try to teach you HTML by
answering your questions that you apparently ask in order to work on the
product just to add a few additional "error messages".

 

[Albert Wiersch]

"Albert Wiersch" <[email protected]>
wrote:

You never answered why you keep calling a non-validator a validator,
except
by admitting that it is not "technically" (i.e., in reality) a validator.

I've answered it many times. This is an example of why I'm not going to
continue to talk about it - it goes no where. See my previous messages for
why it is called a validator. Look up the word validate in the dictionary
(note that there is more than one definition).

 

[Neredbojias]

With neither quill nor qualm, Albert Wiersch quothed:

I've answered it many times. This is an example of why I'm not going to
continue to talk about it - it goes no where. See my previous messages for
why it is called a validator. Look up the word validate in the dictionary
(note that there is more than one definition).

Hey, I might be a smartass and make immature posts on purpose, but you 2
sound like little kids obliviously.