I've done a ton of research on this and have asked many people with help and still no success. Here are the details...
I'm involved in developing a website that pulls data from various data files, combines them in a temp .csv file, and then is graphed using a popular graphing library: dygraphs. The bulk of the website is written in PHP. The parameters that determine the data that is graphed are stored in the users session, the .csv is named after the users session and available for download, and then the .csv file is written in a script that passes it to the dygraphs object. And we've found, even with the no-cache headers sent:
header("Cache-Control: no-cache, must-revalidate");
header("Expires: Sat, 26 Jul 1997 05:00:00 GMT");
Many users experience in the middle of a session, (if enough different graphs are generated) the page displaying an older, static rendering of the page (data they had graphed earlier in the session) as if it were cached and loaded instead of getting a new request. It only gets weirder though:
I've checked using developer tools in both Firefox and Chrome and both browsers are receiving the no-cache headers just fine; Even when the problem occurs if you view the page source, the source is the correct content (a table/legend is also dynamically created using php, the source shows the correct table, but what is rendered is older content); the page begins to render correctly until the graph is about to be display, and then shows the older content; the older content displays as if it were a completely static overlay--the cached graph does not have the same dynamic features (roll over data point display, zoom and pan, etc.) And it is as if the correct page were somewhere beneath it (the download button for the csv file moves depending on how large the table is. The older, static page does nothing if you click the download .csv button, but if you can manage to find the one in the page beneath it you can click and still download the .csv. The data in the .csv is correct)
It is one of the strangest things I've seen in development thus far. Some other relevant facts are that all the problems I've personally experience occurred while I was using Chrome. Non of these symptoms have been reported by Firefox users. IE users have had the same problems (IE users are forced to use chrome frame).
I'm at my wits end at this point. We've sent the php headers; we've tried setting the cache profile for php on IIS as "DisableCache" (or whatever); we've tried sending a random query string to the results page; we've tried all the appropriate meta tags--all with no success.