During the past few months I've put a lot of effort and time into the Scoundrel Project. I had an idea and I wanted to see it actually implemented. While p2p clients, networks and startups were popping out of the ground, they were not addressing the issue I felt most important - the automation of the whole process.

The idea is in itself not too original - "enter extended query-information, do some searching, some semi-intelligent matching and then download the stuff". The originality I believe lies in the use of what I call a linkless index. It contains information about what should ideally be out there rather than information about what is actually out there.

Let's look at the implications of this:

• When listing what should ideally be out there, it is possible to have a comprehensive, cross-linked, easy-to-browse interface to the "database". What you might call "the amazon experience" - almost everything is there, easy to browse and nicely cross-linked. Not a worry in the world about what might be available on some obscure file-sharing networks. Also:
  • The tedious part of the whole process, the searching and downloading, is automatable.
  • Even slow networks seem "snappy" since the user is browsing the (fast) linkless index and does not have to be there when the perhaps slow process of searching and downloading takes place.
    
    
• When listing what is actually out there, you have (at least) the following problems:
  • Redundancy - almost the same file being shared on many computers at the same time. This leads to a problem of selection - which one of the files should you get? (and the problem of screening hundreds of files)
  • Lack of comprehensiveness. Since you only see what is out there, perhaps what you are looking for is not there. It might be in an hour, tomorrow or whenever, but not now. You need to re-search.
  • No cross-links. You need to know exactly what you are looking for (more or less) or else you find nothing.
  • No automation. You have to manually enter all the search-phrases, manually select which files to download and manually retry other files if a download fails.

Scoundrel has demonstrated the feasibility and usability of the "linkless index"-based filesharing in the case of music-sharing over the Napster network using Amazon as the linkless index. However, the concept has much wider usages. It can be used for any type of filesharing where it is possible to tie meta-information to a specific resource (=file). In the case of music-sharing, this was quite easy since the names of music-files are usually sufficient to determine their content. However, it should be possible to do similar kinds of linking between other types of content and meta-information.

The power of the linkless index lies in the lack of links - the index does not require maintenance in the same sense as a linked index (it of course require maintenance, but not of the actual linking part). The "links" are discovered dynamically and without the need for the user's intervention.

Well, so much for what scoundrel has and has not done. As of today, March 1st, 2001, I will no longer be able to continue development on Scoundrel. I'll be disappearing from the face of the earth and will not be reachable. I will not go into the reasons behind this.

Scoundrel will continue to be distributed on this site until OpenAve decides against it.

I would very much like to see the development of Scoundrel continued and if you would like to take on the task, download the source and start a new project, either here, on sourceforge or wherever. Please post a notice about it on infoanarchy.org if you do.

The code is not heavily commented. The basic design of the program is however quite intuitive and easy, so you shouldn't have much of a problem getting started. Get ModelMaker and generate some overview UML if you want a kickstart.

See you when you get there.