The traditional web caching is currently limited to static documents only. A page generated on the fly from a server side script may have different contents on different accesses and hence cannot be cached. A number of proposals for attacking the problem have emerged based on the observation that different instances of a dynamic document are usually quite similar in most cases, i.e. they have a lot of common HTML code. In this paper, we first review these related techniques and show their inadequacy for practical use. We then present a general and fully automatic technique called Vcache based on the decomposition of dynamic documents into a hierarchy of templates and bindings. The technique is designed keeping in mind languages like Perl and C etc that generate the documents using low-level print like statements. These languages together, account for the largest number of dynamic documents on the web.