D-Book Connected TV Standards from the Digital Television Group
In 2013, id3as, my company, was commissioned by Arqiva (UK TV infrastructure providers) to build an encoding and distribution workflow to underpin the UK hybridization of the national digital broadcast television to include OTT services as “red button” features on the services.
The digital television group that oversees the main DTT service called “Freeview” publish the technical model for the service as a standard called the “D-Book” The D-Book is only available to DTG members.
As we came into the project, there was a sense of achievement internally at Arqiva that 16% of the target D-Book compliant Set-Top Boxes and smart TVs were already able to receive the new standard of TV programming and delivery. However, all their attempts to increase this were resulting in a two-way finger-pointing exercise between vendors, with each saying that they were implemented correctly and that any changes needed to be carried out by the other party. This endemic intransigence is all too common. It is typically a result of the fact that primary developers are rarely a resource afforded by integrators farther up the tree (the ones that typically own the project/client), so there is rarely a sense that “going back to the code to make it work” is feasible or affordable. In turn this means that the badly made core tools are extended by shims/workarounds and hacks at the integration points, rather than fixing the problem at core.
Once id3as got involved, we (owning the encoding platform) took a pragmatic view, looked at the actual packet capture of the sessions between our delivery points and the individual devices, and adjusted the core system to be able to accommodate all the different platforms. We managed this in a matter of two weeks, and took Arqiva from 16% to 96% penetration (the remaining 4% had several unsoldered chips on their circuit boards!). The essence of our approach was to throw the D-Book out of the window and look at what the system actually did in practice, and pragmatically just make it work with no regard to the standard.
The interesting thing is that our resulting platform is arguably one the most universally compliant with D-Book standards, but we simply didn't follow the standard ourselves. So it would be difficult for us to badge the system as standard compliant.
There are obviously many examples like this, and they highlight the complexity faced by operators and developers as they attempt to harmonize standards to produce cost controls, while also competing with each other on features and function.
-  http://dtg.org.uk/publications/dbook.html