In order to make the Semantic Web real we need the infrastructure to store, query and update information adhering to the RDF paradigm. Such infrastructure can be developed from scratch or benefit from developments and experiences made in other science \& technology realms such as within the database domain. For querying RDF data the World Wide Web Consortium released a Working Draft for the SPARQL query language. A large portion of the Web is meanwhile driven by server-side Web applications. PHP is the scripting language most widely used for Web applications. In this paper we present our PHP implementation of the SPARQL standard directly interacting with an underlying database system. The approach is based on the rationale of pushing as much work into the RDBMS as possible in order to profit most from the query optimization techniques developed for relational database systems. The article includes an evaluation of the performance and standard compliance, surveys limitations we discovered when using SPARQL for the implementation of real-life Semantic Web applications and suggests extensions to the SPARQL standard in order to solve these obstacles.