Errata

The authors recommend not using the <!DOCTYPE> tag for HTML documents. This is
a bad idea for several reasons. First, it will trigger "quirks" or "quirky"
mode in IE for Mac, Netscape 6, and all browsers built with Mozilla. See
http://www.hut.fi/u/hsivonen/doctype.html for details. Additionally, documents
without this tag will not pass through the W3C validator. Unvalidated
documents have a high probability of containing the "tag soup" that has
plagued the web. See http://web.oreilly.com/news/csstop10_0500.html for
details. Last, and maybe least, documents without <!DOCTYPE> won't let iCab
http://www.icab.de/smile.html smile!

precedes their HTML documents with the SGML doctype command. Because of the
confusion of versions and standards, we don't recommend that you include the
prefix with your HTML documents either."

Further, on page 50, section 3.6.1.3, 2nd paragraph, the author states,
"Serious authors should instead use an SGML <!doctype> tag at the beginning of
their documents, like this:"

I think the authors are a little confused...I am.

The example shows February 22, 1998 at 2:26 p.m. GMT. The equivalent Eastern
Standard Time would be 9:26 a.m. (the sun rises earlier in Greenwich than New
York) which should be coded 1998-02-22T09:26-05:00 rather than the 7:26 p.m.
example in the book.

The id attribute has a number of restrictions on the characters that can go into it.
For example it may not contain white space. The authors claim that it is any "quote-
enclosed string".

It's not clear whether "Wild Wild Web" was meant to be humorous or whether it
was a typographical error and should have been "World Wild Web".

To be compliant, with HTML 4 and XHTML standards, place
the <a> tag inside other markup tags, not the opposite.

I've created several XTHML pages that have, nested inside, <i></i> or <b></b>
tag sets and have successfully validated these pages using the STRICT DTD. I
just looked at the W3C specifications for HTML 4.01 and XTHML 1.0 and could
find no reference to a prohibiton on including other markup tags within the
<a></a> tag set. Of course, nesting of links within links is prohibited and a
well-formed document will be properly nested (i.e., without overlap).

The coordinates given for the rectangle shape ("75,75,99,99") in both <map> examples
only cover the lowest-rightmost sixteenth of a 100 x 100 image (not the entire lower-
right quarter).

In the second example, the author asserts that clicking anywhere in the lower-left
quarter will display the "link4.html" document, when in fact nothing would happen
(since the example does not define an <area> with the coordinates of that quarter,
i.e. "0,50,50,99").

You suggest in this section that the link information may be used by browsers
to provide "next, prev and parent links". How can this work if there is only
one title attribute?

The attributes listed for the <meta> tag includes "HTTP_EQUIV" (with an underbar). I believe this should be "HTTP-EQUIV" (with a d
ash or hyphen).

Both Netscape and Explorer honor the compact attribute (since version 3). If
left out, the <dt> and <dd> text is always placed on two lines. With the
compact attribute the text is placed on a single line only if the <dt> text
fits in the space used for indention, which is for Explorer very small.

There is no mention of the !important label or what effect user's settings of
browser preferences have on author's style settings? Documentation in IE5 and
Netscape 6 both say the user can override the browser's default settings and
the autor's style sheet settings. I gather "!important" is not supported but
is mentioned by W3C.

... the URL is relative to the immediate style sheet's
URL, the one in which it is declared.

On tests I've done for the background property, this is true for IE5.5 and for
Netscape 6 (PR3). It is not so, however, for Netscape 4.73 where the URL is
interpreted as being relative to the html file being browsed.

Is this a bug in this particular version of Netscape, or is this so for all
version 4 browsers?

The author says that the preceeding rule for the <h1> tag will display text "in the
boldest 24-point sans-serif font available."
There is nothing in the cooresponding rule to explicitly indicate bolding; is a value
missing (e.g. "900")?

You say the padding property is supported in Netscape and not in IE. I find
the opposite to be true, at least where I'm using it: a nested table, to
create a hanging indent.

Your demonstration of the hanging indent does not work inside tables in either
browser, but using {padding-left: .17in; text-indent: -.16in} does work in IE.
Using that combination in Netscape, the text-indent has an effect, but the
padding-left doesn't.

The first sentence of the third paragraph states:

"Although deprecated, the <span> tag also supports the standard tag attributes."

The span tag is not deprecated. Perhaps you meant this:

"The <span> tag also supports the deprecated standard tag attributes."

"(Use checkboxes for on/off, yes/no types of form controls.)"

Don't you mean radio buttons?

The last sentence of the first paragraph reads (referring to radio button elements):
"If no element in the group is checked, the browser automatically checks the first
element in the group."

This was true under HTML 2.0, but not anymore. According to the HTML 4.01
specification, "[i]f no radio button in a set sharing the same control name is
initially "on", user agent behavior for choosing which control is initially "on" is
undefined."

(http://www.w3.org/TR/978059600026401/interact/forms.html#radio)

IE 5.5, for one, leaves the field blank if no radio button is selected when a form is
submitted.

The headers attribute is a list of ids (IDREFS in xml). The authors do not state
that it is a white-space separated list, as opposed to, say, comma separated.

IDREFS is also not defined sufficiently to answer this question in the XML chapter.

should

<frame src="pref.html" id="view_frame">

actually be

<frame src="pref.html" name="view_frame">

?

A _parent target does not cause a document to be loaded into a frameset,
unless that frameset as a whole is defined within a document that is itself a
frame within a frameset. In that case, the new document is loaded into the
frame in which the frameset-defining document was loaded. If a _parent target
is used otherwise, it works just like _top.

In the example stated at the top of p.409 and shown in Figure 11-9, a
reference to _parent would act the same as a reference to _top. It works when
the three-column frameset is defined within a seperate document, which in its
turn is loaded in a row of the "top-level" frameset. (The obvious solution
would be to have named framesets, but apparently nobody ever considered this.)
Tested in MSIE 5.5 and below; NS 4.7 and below.

416 Section 12.2., 2nd paragraph, next to last line:

("helper")

consistency of quotation marks with p. 417, Section 12.2.1. The <object> Tag,
last line of the page:

(helper)

The "Top of the Tips" should be to validate your HTML and XHTML code (all
your code all the time, not just when you notice a problem!), as Eric Meyer
explains in his CSS tips http://web.oreilly.com/news/csstop10_0500.html

Invalid HTML is the #1 problem with websites that don't work, and it's far
easier and more effective than trying to test your web pages with all
versions of all browsers.

It is not clear where the "compliance document" is. If you follow the link as
suggested, there is no indication (or link) to find the "compliance document".
If you follow the CSS quick reference link in reinterates, the text in the
book links you back to the original page, catalog/9780596000264/.

The conformance column could replace most !!! by the appropriate entity,
e.g. &fnof; for &#131; etc. upto &Yuml; for &#159; For european users
adding &#128; and &euro; is interesting. For more test results compare
http://frank.ellermann.bei.t-online.de/ibm850.htm (with a link to a
site showing named entities for anything in the "Windows Glyph Set 4",
IIRC (e.g. &house; corresponding to the old 0x7F glyph on PCs).