Advanced Media Project
SAKE Research Track


Project Partners

Project Rationale

The aim of the Advanced Media project is to investigate new and promising technologies in the context of Media Producers (Television & Radio). In it we have identified 3 main research tracks:

Since PROG (48 person months) is mainly involved in the SAKE research track, we will limit the discourse on this page to that track.

The official website for Advanced Media is (soon to be) hosted at

Software Automation and Knowledge Engineering

(Overview presentation in pdf-format available)

The visionary context for this work is to provide a total Media experience in which people no longer want to "sit back, relax, and consume the media". This demand for active participation results in a move towards so-called interactive media. Interactive media are the integration of a traditional content component (audio, video, ...) with a behavioural component (software). Examples of such behavioural components are online gaming, quiz software, and virtual community worlds. All of course integrated with the content component (e.g. a detective TV show). The way in which this kind of software is (or should be) produced not only requires an advanced set of development tools but also a different view on development in general. This need is easily motivated by the following characteristics of interactive media software development :

It is clear that many of the above characteristics can also be found in other software development contexts (who wouldn't want to meet a strict deadline?). Yet in this case the characteristics are all on the extreme end of the spectrum and are rigid. Failing to meet one of them (even partial) results in a development fiasco.

Besides this there also is the extremely "creative" aspect of media development. This results in a situation in which the product-to-be-built is often emerging while it is being built. Hence the constant transformation step of producer-conceptualisation to programmer-conceptualisation forms a bottleneck. This implies that more power should go to the media producer, thus whenever possible bypassing the transformation step.

Concretely we advocate a situation in which the interactive media producer writes down a high level specification (using a domain specific language), pushes a button that automatically generates the corresponding software implementation(s) (using generative programming). All of this with special attention to the evolvability and reusability of the underlying generation system. Of course this story sounds as the holy grail of computer science -- that is why we limit ourselves to fairly strict (and small) domains such as quiz software for instance.

Just as with sake, we think this research should be served hot or used as an essential ingredient for good cooking. But beware, don't buy the regular stuff - go for our Daiginjo-shu variant, which is considered to be the pinnacle of the brewer's art and is generally light, yet complex and quite fragrant :-)

Keywords : Software Evolution, Automated Software, Domain Specific Languages, Generative Programming, Knowledge Based Software Engineering, Transformation Systems, Generators, Metaprogramming, Ontology, AOSD, Software Streaming, ...

Related Research @ PROG

The brains of the following persons are tapped to realize the SAKE vision:

(Contact Dirk Deridder for more information)