It's odd, unless the refresh URL points to a file (or the file includes a CSS doc) that is only intermittently there. What you *might* do is have the script that generates the refresh tag verify that the file at the end of the URL in the tag is there. This is imperfect for a number of reasons (only easy test is the -e filetest operator, which you can only use if the URL is on the a filesystem accessible to the machine the script's running on) , and if the file is only intermittently there, it might be gone by the time the refresh takes effect.

but you might cut down on the number of 404s this way.

Philosophy can be made out of anything. Or less -- Jerry A. Fodor