Dear Mr. Jens Niemeyer
I have some ideas for repository systems. It should be based on biological principles, as DNA replication, or on the principle of internet that it is stored in multiple locations. The practical realisation of such principle would be to store random parts of knowledge on a part of disk on every sold computer, or still better on a new, non active disk on a computer. When a disaster will happen, some of those computers will survive and they can behave like servers for new internet. Of course, this needs a calculation how much computers we need for such storage. Probably it is not enough space, but after catastrophe some knowledge will remain. Even, every computer drive or flash drive can be filled with such data and after being filled with data the data can be partially deleted. It is not necessary that the sold drives are empty.
Such approach will also enable some browsing on computer, although web will not be accessible.
Other possibility is to built one center on a moon, which will survive even in a case on nuclear war on earth. But this is not so easy feasible than the above model.
The next big question is organization of such knowledge. One help is to make knowledge to the final form from working form. For instance, when the theory of everything (TOE) will be found, it can be (maybe) written on a t-shirt. Thus much knowledge, how to obtain TOE, and many speculations will not be so necessary. The hierarchy and organization of all such knowledge will also be beneficial as new knowledge. On the other side, how to present knowledge is also a not finished story. For instance, I claim that special relativity is not yet presented clearly enough.
These ideas are obtained in one hour and if thorough reflection will not completely delete them, it will be a success.
Best regards
Janko Kokosar