根据508法规(对于视障屏幕阅读器)，表格应该只用于保存数据，而不是用于布局，因为它会导致屏幕阅读器崩溃。至少别人是这么告诉我的。

如果您为每个div分配名称，您也可以使用CSS将它们一起蒙皮。只是让他们坐成你想要的样子有点麻烦。

I have found that even with the best planning divs come up short in several respects. For instance. there is no way with divs to have a bottom bar that always sits at the bottom of the browser, even when the rest of the content does not go to the bottom of the browser. Also, you cannot elegantly do anything better than three columns, and you cannot have columns that grow and shrink according the the width of their content. In the end, we try to use divs first. However, we will not limit our html designs based on some religious content vs layout ideal.

当然，这篇文章有点小题大做，争论似乎很简单，很容易反驳。

网页是Web开发人员的领域，如果他们说div和CSS比表格更好，那对我来说已经足够好了。

如果布局是由服务器应用程序生成的表实现的，那么新的布局意味着对应用程序的更改，应用程序的重新构建和重新部署，而不仅仅是对css文件的更改。

另外,可访问性。用于布局的表格会使网站无法访问，所以不要使用它们。这是显而易见的，更不用说违法了。

I still don't quite understand how divs / CSS make it easier to change a page design when you consider the amount of testing to ensure the changes work on all browsers, especially with all the hacks and so on. Its a hugely frustrating and tedious process which wastes large amounts of time and money. Thankfully the 508 legislation only applies to the USA (land of the free - yeah right) and so being as I am based in the UK, I can develop web sites in whatever style I choose. Contrary to popular (US) belief, legislation made in Washington doesn't apply to the rest of the world - thank goodness for that. It must have been a good day in the world of web design the day the legislation came into force. I think I'm becoming increasingly cynical as I get older with 25 years in the IT industry but I feel sure this kind of legislation is just to protect jobs. In reality anyone can knock together a reasonable web page with a couple of tables. It takes a lot more effort and knowledge to do this with DIVs / CSS. In my experience it can take hours and hours Googling to find solutions to quite simple problems and reading incomprehensible articles in forums full of idealistic zealots all argueing about the 'right' way to do things. You can't just dip your toe in the water and get things to work properly in every case. It also seems to me that the lack of a definitive guide to using DIVS / CSS "out of the box", that applies to all situations, working on browsers, and written using 'normal' language with no geek speak, also smells of a bit of protectionism. I'm an application developer and I would say it takes almost twice as long to figure out layout problems and test against all browsers than it does to create the basic application, design and implement business objects, and create the database back end. My time = money, both for me and my customers alike so I am sorry if I don't reject all the pro DIV / CSS arguments in favour of cutting costs and providing value for money for my customers. Maybe its just the way that developers minds work, but it seems to me far easier to change a complex table structure than it is to modify DIVs / CSS. Thankfully it now appears that a solution to these issues is now available - its called WPF.

In the past, screen readers and other accessibility software had a difficult time handling tables in an efficient fashion. To some extent, this became handled in screen readers by the reader switching between a "table" mode and a "layout" mode based on what it saw inside the table. This was often wrong, and so the users had to manually switch the mode when navigating through tables. In any case, the large, often highly nested tables were, and to a large extent, are still very difficult to navigate through using a screen reader.

The same is true when divs or other block-level elements are used to recreate tables and are highly nested. The purpose of divs is to be used as a fomating and layout element, and as such, are intended used to hold similar information, and lay it out on the screen for visual users. When a screen reader encounters a page, it often ignores any layout information, both CSS based, as well as html attribute based(This isn't true for all screen readers, but for the most popular ones, like JAWS, Windows Eyes, and Orca for Linux it is).

为此，表格式数据，也就是逻辑上有意义的在二维或多维维度中排序的数据，具有某种标题，最好放在表中，并使用div来管理页面上内容的布局。(另一种思考“表格数据”的方式是尝试以图表形式绘制它……如果你不能，它可能不是最好的表示在一个表中)

Finally, with a table-based layout, in order to achieve a fine-grained control of the position of elements on the page, highly nested tables are often used. This has two effects: 1.) Increased code size for each page - Since navigation and common structure is often done with the tables, the same code is sent over the network for each request, whereas a div/css based layout pulls the css file over once, and then uses less wordy divs. 2.) Highly nested tables take much longer for the client's browser to render, leading to slightly slower load times.

在这两种情况下，“最后一英里”带宽的增加，以及更快的个人电脑缓解了这些因素，但它们仍然是许多网站存在的问题。

With all of this in mind, as others have said, tables are easier, because they are more grid-oriented, allowing for less thought. If the site in question is not expected to be around long, or will not be maintained, it might make sense to do what is easiest, because it might be the most cost effective. However, if the anticipated userbase might include a substantial portion of handicapped individuals, or if the site will be maintained by others for a long time, spending the time up front to do things in a concise, accessible way may payoff more in the end.

下面是来自最近项目的一段html:

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
    <title>{DYNAMIC(TITLE)}</title>
    <meta http-equiv="content-type" content="text/html;charset=utf-8" />
    <meta http-equiv="Content-Style-Type" content="text/css" />
    <link rel="stylesheet" type="text/css" href="./styles/base.css" />
</head>
<body>
    <div id="header">
        <h1><!-- Page title --></h1>
        <ol id="navigation">
            <!-- Navigation items -->
        </ol>
        <div class="clearfix"></div>
    </div>
    <div id="sidebar">
        <!-- Sidebar content -->
    </div>
    <!-- Page content -->
    <p id="footer"><!-- Footer content --></p>
</body>
</html>

这是与基于表格的布局相同的代码。

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
    <title>{DYNAMIC(TITLE)}</title>
    <meta http-equiv="content-type" content="text/html;charset=utf-8" />
    <meta http-equiv="Content-Style-Type" content="text/css" />
    <link rel="stylesheet" type="text/css" href="./styles/base.css" />
</head>
<body>
    <table cellspacing="0">
        <tr>
            <td><!-- Page Title --></td>
            <td>
                <table>
                    <tr>
                        <td>Navitem</td>
                        <td>Navitem</td>
                    </tr>
                </table>
            </td>
        </tr>
    </table>

    <table>
        <tr>
            <td><!-- Page content --></td>
            <td><!-- Sidebar content --></td>
        </tr>
        <tr>
            <td colspan="2">Footer</td>
        </tr>
    </table>
</body>
</html>

我在基于表格的布局中看到的唯一干净的地方是我对缩进的过度热情。我确信内容部分将有另外两个嵌入式表。

另一件需要考虑的事情是:文件大小。我发现基于表格的布局通常是CSS布局的两倍大。在我们的高速宽带上，这不是一个大问题，但在那些拨号调制解调器上。