使用表格布局的工具可能会因为创建布局所需的大量代码而变得异常沉重。SAP的Netweaver Portal默认使用TABLE来布局页面。

在我目前的工作中，生产SAP门户有一个主页，它的HTML超过60K，有7个表那么深，在页面中有3次。再加上Javascript，误用了16个iframe，其中有类似的表格问题，CSS过重等，页面重量超过5MB。

花点时间降低页面重量，这样你就可以利用带宽与用户进行互动，这是值得的。

内容和布局之间的分离也使它更容易为您的网站生成打印机友好的布局或不同的皮肤(样式)，而不必创建不同的html文件。有些浏览器(如Firefox)甚至支持从视图菜单中选择样式表。

而且我确实认为保持无表格布局更容易。你不需要担心行span, colspan等等。您只需创建一些容器div并将内容放置在需要的位置。也就是说，我认为它也更有可读性(<div id="sidebar"> vs <tr><td>…</td><td>…<td>sidebar</td></tr>)。

这只是一个你必须学会的小“技巧”(一旦你掌握了这个技巧，我认为它会更容易，更有意义)。

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.

我想使用表格元素进行布局确实与表格数据没有什么关系。那又怎样?我的老板在乎吗?我的用户关心吗?

谷歌和其他自动化系统确实关心这些问题，而且它们在许多情况下同样重要。语义代码对于非智能系统来说更容易解析和处理。

如果你在这方面支持表格角度，找一个有表格的网站，然后给自己买一个屏幕阅读器——关掉屏幕阅读器，关掉你的显示器。

然后尝试一个不错的语义正确的div布局网站。

你会发现其中的不同。

如果表格中的数据是表格而不是为了布局页面，那么表格并不是邪恶的。

我曾经了解到，一个表是立即加载的，换句话说，当连接很慢的时候，表所在的空间保持空白，直到整个表被加载，另一方面，一个div加载从上到下的速度与数据到达的速度一样快，不管它是否已经完成。