> I think you're describing something like a hybrid social network/semantic web with versioning, knowledge provenance, and asynchronous peer review.
Great breakdown, thanks.
> One serious problem for any such system is ontology selection: how is one to represent the entire body of scientific knowledge under a single type system?
I think you can do it through market forces and forking. Similar to Linux distributions you could have "science" distributions. As for the type system and unifying language, I think you can do a thing that can start from just a dot and no dot and build up characters and numbers and words and types etc, with no extra parts. So perhaps if you had something like that, where simplicity could be rigorously defined, you could get consensus on base level types. If people had strong differences, you could go off and fork a new distribution. I'd imagine you'd have a few distros emerge with decent gravity.
Pull request could be really neat. I can imagine you'd have the speed of light defined somewhere in a science distro, and there would be data from reproducible experiments that people have done. Perhaps someone comes up with an ingenuous at home experiment that is just a few steps and sends a pull request that would add that, and perhaps prune some more complex experiment.
> offloads the hard, fuzzy work to the...brain
Yes, exactly. I'd love it if it was computable (of course, this isn't an original idea--Wolfram Alpha is trying to pull it off). If I could "go to definition" of any scientific conclusion (not only to definitions, but to real data, and simple, repeatable experiments). If you kept "going to definition" all roads eventually would lead back to 0 and 1.
> One could try to compare the shape of various data under different encodings, for instance (some sort of topological analysis?) to identify similar structure?
I'm giving it a go with Trees. I think it will work, but still might be a few years before I know for sure.
> There are already a few entries in the social network/resource-sharing platform space. Have a look at Open Science Foundation. Academia and ResearchGate are similar, but without the materials-and-data-sharing.
I like those, especially OSF. Definitely a lot of activity in the space (I invested in some new ones as well). I haven't seen one yet that does the "science monorepo" thing, but hoping someone takes the lead there (and I'll do my best to support it with hopefully useful underlying tech and research).
Great breakdown, thanks.
> One serious problem for any such system is ontology selection: how is one to represent the entire body of scientific knowledge under a single type system?
I think you can do it through market forces and forking. Similar to Linux distributions you could have "science" distributions. As for the type system and unifying language, I think you can do a thing that can start from just a dot and no dot and build up characters and numbers and words and types etc, with no extra parts. So perhaps if you had something like that, where simplicity could be rigorously defined, you could get consensus on base level types. If people had strong differences, you could go off and fork a new distribution. I'd imagine you'd have a few distros emerge with decent gravity.
Pull request could be really neat. I can imagine you'd have the speed of light defined somewhere in a science distro, and there would be data from reproducible experiments that people have done. Perhaps someone comes up with an ingenuous at home experiment that is just a few steps and sends a pull request that would add that, and perhaps prune some more complex experiment.
> offloads the hard, fuzzy work to the...brain
Yes, exactly. I'd love it if it was computable (of course, this isn't an original idea--Wolfram Alpha is trying to pull it off). If I could "go to definition" of any scientific conclusion (not only to definitions, but to real data, and simple, repeatable experiments). If you kept "going to definition" all roads eventually would lead back to 0 and 1.
> One could try to compare the shape of various data under different encodings, for instance (some sort of topological analysis?) to identify similar structure?
I'm giving it a go with Trees. I think it will work, but still might be a few years before I know for sure.
> There are already a few entries in the social network/resource-sharing platform space. Have a look at Open Science Foundation. Academia and ResearchGate are similar, but without the materials-and-data-sharing.
I like those, especially OSF. Definitely a lot of activity in the space (I invested in some new ones as well). I haven't seen one yet that does the "science monorepo" thing, but hoping someone takes the lead there (and I'll do my best to support it with hopefully useful underlying tech and research).