HTML

You can suggest this presentation in a style sheet. First, specify colors for normal links, like this:

a:link {color: blue; background: white}

a:visited {color: purple; background: white}

a:active {color: red; background: white}

Next, identify the links that you want to have different colors. You can use the CLASS attribute in your HTML, like this:

<a class="example1" href="[URL]">[link text]</a>

Then, in your style sheet, use a selector for links with this CLASS attribute, like this:

a.example1:link {color: yellow; background: black}

a.example1:visited {color: white; background: black}

a.example1:active {color: red; background: black}

Alternatively, you can identify an element that contains the links that you want to have different colors, like this:

<div class="example2">...

<a href="[URL]">[link text]</a>...

<a href="[URL]">[link text]</a>...

<a href="[URL]">[link text]</a>...

</div>

Then, in your style sheet, use a selector for links in this containing element, like this:

.example2 a:link {color: yellow; background: black}

.example2 a:visited {color: white; background: black}

.example2 a:active {color: red; background: black}

The special case of the less-than ('<'), greater-than ('>'), and ampersand ('&') characters. In general, the safest way to write HTML is in US-ASCII (ANSI X3.4, a 7-bit code), expressing characters from the upper half of the 8-bit code by using HTML entities.

Working with 8-bit characters can also be successful in many practical situations: Unix and MS-Windows (using Latin-1), and also Macs (with some reservations).

Latin-1 (ISO-8859-1) is intended for English, French, German, Spanish, Portuguese, and various other western European languages. (It is inadequate for many languages of central and eastern Europe and elsewhere, let alone for languages not written in the Roman alphabet.) On the Web, these are the only characters reliably supported. In particular, characters 128 through 159 as used in MS-Windows are not part of the ISO-8859-1 code set and will not be displayed as Windows users expect. These characters include the em dash, en dash, curly quotes, bullet, and trademark symbol; neither the actual character (the single byte) nor its &#nnn; decimal equivalent is correct in HTML. Also, ISO-8859-1 does not include the Euro currency character. (See the last paragraph of this answer for more about such characters.)

On platforms whose own character code isn't ISO-8859-1, such as MS-DOS and Mac OS, there may be problems: you have to use text transfer methods that convert between the platform's own code and ISO-8859-1 (e.g., Fetch for the Mac), or convert separately (e.g., GNU recode). Using 7-bit ASCII with entities avoids those problems, but this FAQ is too small to cover other possibilities in detail.

If you run a web server (httpd) on a platform whose own character code isn't ISO-8859-1, such as a Mac or an IBM mainframe, then it's the job of the server to convert text documents into ISO-8859-1 code when sending them to the network.

If you want to use characters not in ISO-8859-1, you must use HTML 4 or XHTML rather than HTML 3.2, choose an appropriate alternative character set (and for certain character sets, choose the encoding system too), and use one method or other of specifying this.

On current browsers, the entire table must be downloaded and the dimensions of everything in the table must to be known before the table can be rendered. That can delay the rendering of your content, especially if your table contains images without HEIGHT or WIDTH attributes.

If any of your table's content is too wide for the available display area, then the table stretches to accomodate the oversized content. The rest of the content then adjusts to fit the oversized table rather than fitting the available display area. This can force your readers to scroll horizontally to read your content, or can cause printed versions to be cropped.

For readers whose displays are narrower than the author anticipated, fixed-width tables cause the same problems as other oversized tables. For readers whose displays are wider than the author anticipated, fixed-width tables cause extremely wide margins, wasting much of the display area. For readers who need larger fonts, fixed-width tables can cause the content to be displayed in short choppy lines of only a few words each.

Many browsers are especially sensitive to invalid syntax when tables are involved. Correct syntax is especially critical. Even with correct syntax, nested tables may not display correctly in older versions of Netscape Navigator.

Some browsers ignore tables, or can be configured to ignore tables. These browsers will ignore any layout you've created with tables. Also, search engines ignore tables. Some search engines use the text at the beginning of a document to summarize it when it appears in search results, and some index only the first n bytes of a document. When tables are used for layout, the beginning of a document often contains many navigation links that appear before than actual content.

Many versions of Navigator have problems linking to named anchors when they are inside a table that uses the ALIGN attribute. These browsers seem to associate the named anchor with the top of the table, rather than with the content of the anchor. You can avoid this problem by not using the ALIGN attribute on your tables.

If you use tables for layout, you can still minimize the related problems with careful markup. Avoid placing wide images, PRE elements with long lines, long URLs, or other wide content inside tables. Rather than a single full-page layout table, use several independent tables. For example, you could use a table to lay out a navigation bar at the top/bottom of the page, and leave the main content completely outside any layout tables.

If your images are inside a table, be sure to set the BORDER, CELLSPACING, and CELLPADDING attributes to 0.

Extra space between images is often created by whitespace around the <IMG> tag in the markup. It is safe to use newlines inside a tag (between attributes), but not between two tags. For example, replace this:

<td ...>

<img src=... alt=...>

<img src=... alt=...>

</td>

with this:

<td ...><img src=... alt=...><img src=... alt=...></td>

According to the latest specifications, the two should be equivalent. However, common browsers do not comply with the specifications in this situation.

Finally, extra space between images can appear in documents that trigger the "standards" rendering mode of Gecko-based browsers like Mozilla and Firefox.

The URL structure defines a hierarchy (or relationship) that is similar to the hierarchy of subdirectories (or folders) in the filesystems used by most computer operating systems. The segments of a URL are separated by slash characters ("/"). When navigating the URL hierarchy, the final segment of the URL (i.e., everything after the final slash) is similar to a file in a filesystem. The other segments of the URL are similar to the subdirectories and folders in a filesystem.

A relative URL omits some of the information needed to locate the referenced document. The omitted information is assumed to be the same as for the base document that contains the relative URL. This reduces the length of the URLs needed to refer to related documents, and allows document trees to be accessed via multiple access schemes (e.g., "file", "http", and "ftp") or to be moved without changing any of the embedded URLs in those documents.

Before the browser can use a relative URL, it must resolve the relative URL to produce an absolute URL. If the relative URL begins with a double slash (e.g., //www.yoursite.com/faq/html/), then it will inherit only the base URL's scheme. If the relative URL begins with a single slash (e.g., /faq/html/), then it will inherit the base URL's scheme and network location.

If the relative URL does not begin with a slash (e.g., all.html , ./all.html or ../html/), then it has a relative path and is resolved as follows.

1. The browser strips everything after the last slash in the base document's URL and appends the relative URL to the result.

2. Each "." segment is deleted (e.g., ./all.html is the same as all.html, and ./ refers to the current "directory" level in the URL hierarchy).

3. Each ".." segment moves up one level in the URL hierarchy; the ".." segment is removed, along with the segment that precedes it (e.g., foo/../all.html is the same as all.html, and ../ refers to the parent "directory" level in the URL hierarchy).

Some examples may help make this clear. If the base document is <URL:http://www.yoursite.com/faq/html/basics.html>, then

all.html and ./all.html

refer to <URL:http://www.yoursite.com/faq/html/all.html>

./

refers to <URL:http://www.yoursite.com/faq/html/>

../

refers to <URL:http://www.yoursite.com/faq/>

../cgifaq.html

refers to <URL:http://www.yoursite.com/faq/cgifaq.html>

../../reference/

refers to <URL:http://www.yoursite.com/reference/>

Please note that the browser resolves relative URLs, not the server. The server sees only the resulting absolute URL. Also, relative URLs navigate the URL hierarchy. The relationship (if any) between the URL hierarchy and the server's filesystem hierarchy is irrelevant.

It is never wrong to quote attribute values, and many people recommend quoting all attribute values even when the quotation marks are technically optional. XHTML 1.0 requires all attribute values to be quoted. Like previous HTML specifications, HTML 4 allows attribute values to remain unquoted in many circumstances (e.g., when the value contains only letters and digits).

Be careful when your attribute value includes double quotes, for instance when you want ALT text like "the "King of Comedy" takes a bow" for an image. Humans can parse that to know where the quoted material ends, but browsers can't. You have to code the attribute value specially so that the first interior quote doesn't terminate the value prematurely. There are two main techniques:

* Escape any quotes inside the value with " so you don't terminate the value prematurely: ALT="the "King of Comedy" takes a bow".

* Use single quotes to enclose the attribute value: ALT='the "King of Comedy" takes a bow'.

Both these methods are correct according to the specification and are supported by current browsers, but both were poorly supported in some earlier browsers. The only truly safe advice is to rewrite the text so that the attribute value need not contain quotes, or to change the interior double quotes to single quotes, like this: ALT="the 'King of Comedy' takes a bow".

1. Bulleted Lists: <ul> begins a bulleted, indented list. Each item in the list is then prefaced with the <li> tag. It is not necessary to insert a break at the end of each line -- the <li> tag automatically creates a new line.

* with <li type=disc>

* with <li type=square>

* with <li type=circle>

2. Numbered Lists: <ol> begins a numbered, indented list. Each item in the list is then prefaced with the <li> tag. You need to close the list with the </ol> tag. Note: You can expand the <ol> to specify the TYPE of numbering:

<ol> 1 (decimal numbers: 1, 2, 3, 4, 5, ...)

<ol type="a"> a (lowercase alphabetic: a, b, c, d, e, ...)

<ol type="A"> A (uppercase alphabetic: A, B, C, D, E, ...)

<ol type="i"> i (lowercase Roman numerals: i, ii, iii, iv, v, ...)

<ol type="I"> I (uppercase Roman numerals: I, II, III, IV, V, ...)

If you want others to view your web page with specific colors, the most appropriate way is to suggest the colors with a style sheet. Cascading Style Sheets use the color and background-color properties to specify text and background colors. To avoid conflicts between the reader's default colors and those suggested by the author, these two properties should always be used together.

With HTML, you can suggest colors with the TEXT, LINK, VLINK (visited link), ALINK (active link), and BGCOLOR (background color) attributes of the BODY element.

Note that these attributes are deprecated by HTML 4. Also, if one of these attributes is used, then all of them should be used to ensure that the reader's default colors do not interfere with those suggested by the author. Here is an example:

<body bgcolor="#ffffff" text="#000000" link="#0000ff" vlink="#800080" alink="#000080">

Authors should not rely on the specified colors since browsers allow their users to override document-specified colors.

No. The server-side (e.g., CGI) program that processes the form submission must handle duplicate submissions gracefully.

You could generate the form with a server-side (e.g., CGI) program that adds a hidden field with a unique session ID. Then the server-side program that processes the form submission can check the session ID against a list of previously used session IDs. If the session ID has already been used, then an appropriate action can be taken (e.g., reject the submission, or update the previously submitted data).

Ultimately, your server-side program must be smart enough to handle resubmitted data. But you can avoid getting resubmitted data by not expiring the confirmation page from form submissions. Since you want to expire pages quickly when they have transient data, you might want to avoid putting transient data on the confirmation page. You could provide a link to a database query that returns transient data though.

Have the server-side (e.g., CGI) program that processes the form submission send an error message if the field is not filled in properly. Ideally, this error message should include a copy of the original form with the original (incomplete or incorrect) data filled in as the default values for the form fields. The Perl CGI.pm module provides helpful mechanisms for returning partially completed forms to the user.

In addition, you could use JavaScript in the form's ONSUBMIT attribute to check the form data. If JavaScript support is enabled, then the ONSUBMIT event handler can inform the user of the problem and return false to prevent the form from being submitted.

Note that the server-side program should not rely upon the checking done by the client-side script.

This is unfortunately not possible. When you navigate through a site using frames, the URL will not change as the documents in the individual frames change. This means that there is no way to indicate the combination of documents that make up the current state of the frameset.

The author can provide multiple frameset documents, one for each combination of frame content. These frameset documents can be generated automatically, perhaps being created on the fly by a CGI program. Rather than linking to individual content documents, the author can link to these separate frameset documents using TARGET="_top". Thus, the URL of the current frameset document will always specify the combination of frames being displayed, which allows links, bookmarks, etc. to function normally.