If Riku is up to what I think, it's very slick and there's a technique that makes it all work: Let's say you have a page that displays text from a database. The content doesn't change very often, but the content in the DB is maintained by a content manager through an admin screen (e.g. "help" text, "about" text, etc.). Your objective is to present content without having to hit the database each time the page is served.
Try this: the link to this page is to an ASP page. The ASP page uses the filesystem object to check for the existance of an HTML page with the same name as the asp page. If the corresponding HTML page exists, the script redirects (or even better uses response.transfer) to the HTML page.
If the corresponding HTML page does not exist, then build the needed HTML in a string (see articles about using many response.write versus one large one) and use the filesystem object to create the desired HTML file. Write the string to this file, and after closing it, do a redirect (or transfer) to it.
In your admin screen, when someone changes the content for this (or other pages), use the filesystem object to delete the HTML file derived from the content (this is really only practical for a very small number of pages). This effectively forces the HTML to be rebuilt, on demand, whenever the content changes.
Think of it as "Just-in-time HTML"