Now if the original URL (containing the <ISINDEX> tag) is just a HTML document, then it will not know what to do with such a query. Although the documentation does not say so, the only practical way of using <ISINDEX> queries is as follows: the original URL must be a program. If the program has no query then it outputs the required HTML containing the <ISINDEX> statement. If the program has a query then it knows that it must be the result of its earlier <ISINDEX> emanation, and so can process the result accordingly.
All this convoluted process makes it not seem worth while. Perhaps it might be useful if you know that a certain browser does not have forms capability.