Although several Digital Musical Instruments (DMIs) have been presented at NIME, very few of them remain accessible to the community. Rebuilding a DMI is often a necessary step to allow for performance with NIMEs. Rebuilding a DMI exactly similar to its original, however, might not be possible due to technology obsolescence, lack of documentation or other reasons. It might then be interesting to re-interpret a DMI and build an instrument inspired by the original one, creating novel performance opportunities. This paper presents the challenges and approaches involved in rebuilding and re-interpreting an existing DMI, The Sponge by Martin Marier. The rebuilt versions make use of newer/improved technology and customized design aspects like addition of vibrotactile feedback and implementation of different mapping strategies. It also discusses the implications of embedding sound synthesis within the DMI, by using the Prynth framework and further presents a comparison between this approach and the more traditional ground-up approach. As a result of the evaluation and comparison of the two rebuilt DMIs, we present a third version which combines the benefits and discuss performance issues with these devices.