The name is rather unfortunate, but the project is quite interesting. I had no idea Vuze, which used to be Azureus, had re-packaged itself as an abstracted DHT provider.
I'm disappointed the author chose to tightly couple their implementation to Vuze. What if an operator wanted to use the system on a private network of machines? Why not implement the functionality using a Kademlia library instead?
Finally, I'm extremely disappointed at the lack of openness in this paper, and indeed in all papers on Systems We Make. The lack of reproducibility in computational sciences is a current problem and I'd like to see more researchers attempt to be more open about sharing not only their code but also their setup. i.e.
"An article about computational science in a scientific publication is not the scholarship itself, it is merely advertising of the scholarship. The actual scholarship is the complete software development environment and the complete set of instructions which generated the figures." --D. Donoho
Although I agree that a research paper, if it included all the minutae required to reproduce the results on a new system, would be far to verbose, it's time researcers woke up and realised what the "scientific" part of "scientific research" means. In my mind, the bare minimum of detail would be:
- all the source code and build environment tools on a easy-to-access repository, e.g. GitHub.
- all raw data numbers available in CSV files.
- all source code used to generate pictorial graphs as the source code to R scripts, or your favourite stats package.
This isn't rocket science. If you need more help figuring this out consider reading:
