CSS1TS Errata: First Draft

The DOCTYPE should be HTML 4.0 Strict instead of HTML 4.0 Transitional.

One suggestion: it would be easier to use if the output lines were numbered. This would make it much quicker to match the code with the corresponding output.

There are identical IDs in the same document, thanks to the table-testing replication. For example:

I think there is a HTML error in section 4.1.2 [1] in the CSS1 Test Suite (and possibly other sections as well) : The complete test is repeated twice on the page (once by itself, and once inside a table). The problem is that there is a paragraph with an id in the test, which means that there is two paragraphs with the same id (<P ID="ruled"> on lines 76 and 146).

Do borders apply to table cells and rows, let alone TABLE elements? If not, the border tests are in serious error.

Also, the inline border tests are hard to see since both the inline and block-level borders are black. I propose changing the block-level to gray, or something similarly contrasting with the inline border colors.

The previous page is not "sec03.htm", but "sec04.htm".

I would like to change the text to say something like:

"This sentence should be black, not red-- ID selectors cannot begin with digits in CSS1. Note that CSS2 doesn't have this restriction and this sentence should be red with CSS2-compliant browsers."

1. I think that "for the A element" is better than "for the A tag".
2. "It also should NOT turn orange" should read "turn maroon".

On the "2.4 first-letter" test page, the third example states "...should by 350% bigger than the rest of the paragraph...". This may be incorrect, since the specification implies that the percentage is taken from the normal font size (where 100% is the same size as the normal font). Suggested corrections (unless I am wrong):

"...350% THE FONT SIZE of the rest of the paragraph..."
"...250% bigger than the rest of the paragraph..."

It's late, and I may not be making sense. But I think this, despite being a common practised "feature" of our language, misleads those who intend to use the suite. When I first read it, I thought, "that W isn't 350% percent bigger, it's only about 250% percent bigger!".

I think that "The first line of this paragraph, and only ..." is better than "The first line of this sentence sould be ...", because the last two paragraph have more than one sentence.

1. I think that all "This sentence ..." should read "This paragraph".
2. In the paragraph id1, I think that "than the declarations which have ID selectors" is better than "given in a STYLE attribute".

1. In the sixth paragraph, I think that "the left margin" is better than "the margins that are set to ...".
2. In the seventh paragraph, I think that "the right margin" is better than "the margins that are set to ...".

In the fourth and fifth paragraph, "and the table in the second" sould read "in the table cell in the second".

The "fifty pixel square" is not intrinsically fifty pixels tall or wide-- the size is actually 15x15. Should we change the wording, or the image?

"SPANned word 'sentence' " sould read "SPANned word 'background' ".

This page does not test for the "alignment" of fixed backgrounds on child elements with the fixed background of ancestor elements (including BODY/HTML). Should this be added here, or should it wait for the CSS2 test suite?

This element is unstyled save for a background color of gray. However, it contains an inline element of class two which should result in -10px left padding (which should also be a light blue) only in the first line in which the inline box appears. http://www.w3.org/Style/CSS/Test/current/sec5509b.htm

As Tapio from Finland just pointed out to me and David, padding cannot take negative values! As far as I can tell, the other padding pages do not have this error. I do not know how this page has escaped the careful error pruning of a few months back...

In section 5.5.17 the style shows correctly but it is described incorrectly. Class 2 should be red (and shows that way) but the text says it should be purple.

I feel the sec5526.htm is not enough for 'clear' test. To check 'clear'ing tallest float element on one side, I think there must be some float images in various height on both sides.

The images run together, making it hard to see what's happening. Change image to circle or other visually discontinuous shape.

There is an example ("P.two") whose text states:

"This paragraph should have a solid gray background (or a white grid), because in CSS1, inherit is an invalid keyword, and in CSS2, it will cause the gray background (not the white grid) to be inherited."

The applicable rule is:

 P.two {background-color: inherit;}

This statement correct for the first occurrence of this statement, where the parent of the P is the BODY which has background-color: #CCCCCC, and therefore is inherited to the P in CSS2.

However, the second occurrence of this statement is in a P inside a TD, and the TD has no background-color specified (therefore gets the initial value of "transparent"), and therefore the P's inherit value inherits "transparent" in CSS2, and therefore the white grid of BODY shows through.

A slight rewording of the descriptive text would fix "P.two":

"This paragraph should have a solid gray background (or a white grid), because in CSS1, inherit is an invalid keyword, and in CSS2, it will cause the gray background (not the white grid) to be inherited in the first part of the document, and a transparent background (thus allowing the white grid background of BODY to shine through) to be inherited in the second part of the document when inside the TD."