most of the people do not do much with e-tag http response headers, and it’s probably ok, unless one is really trying to get the most out of client-side caching.
this has been written for apache 1.3.x, but is still relevant for apache 2.x:
An ETag is an HTTP response header returned by an HTTP /1.1 compliant Web server such as Apache 1.3x. By default, Apache calculates an ETag for a requested file using a combination of the file’s location in the file system (I-Node number on Unix systems), its modification time, and its size. [..]
Because the ETag is calculated using the file’s I-Node, and an I-Node is machine-specific, administrators of Web server farms will experience unexpected requests if the ETag differs from machine to machine.
To work around this issue, use the FileETag directive to configure your Apache server to use only the file modification time and file size when calculating the ETag.
The following example configures Apache to only use the modification time (MTime) and size (Size) when calculating the ETag for any file contained in the /usr/local/httpd/htdocs directory or a subdirectory.
<Directory /usr/local/httpd/htdocs>
FileETag MTime Size
</Directory>
in other words, if you are running multiple web servers that serve static content for which apache provides e-tag response headers, it might make sense to use the directive above to make sure that e-tag header values are not different from machine to machine for the same content. i confirm that this is the case in unix, but i have not verified this on windows.