This paper deals with the problem of scheduling mobile Internet traffic over a single cell wireless downlink data channel. On the one hand, we compare the performance of different size-based or/and channel-aware scheduling disciplines in mean delay terms, concluding that an approach that combines both size- and channel-awareness is the best alternative, which guarantees acceptable uplink overhead due to channel information reports. On the other hand, we study the impact of different scheduling algorithms on user's satisfaction for web browsing service. As concluded, in order to achieve a trade-off between maximizing overall mean subjective quality and reducing uplink signaling overhead due to channel quality reports, a scheduling policy that benefits from opportunistic gains in combination with a size-aware tie-breaking rule gives the best results.