[I know almost nothing about coding. My experience consists solely of some self taught LUA string manipulation and perl regex. You might understand why I am not just coding it myself.]
What I want to do is enable anyone with an internet connection and a few thousand dollars to get a mathematics degree that is as prestigious as one from MIT/Cambridge/Oxford/Harvard etc. by lowering the cost of exams because I have become convinced that in the age of the internet exams are all that any institutions have to offer because information can be distributed more effectively and efficiently via the internet than via in person lectures or physical text books.
TL,DR: Why do we need MIT while we have Khanacademy? If it isn’t MIT’s job to distribute the information but only to accredit it, why not render institutionalized accreditation obsolete and be done with MIT altogether?
I have an idea for a piece of software that might contribute to the freedom of accreditation but it is almost certainly beyond my capacity to code, let alone in my lifetime. I am nonetheless reasonably sure it is a good idea. I selected this forum because my interests seemed aligned with that of the forum. Any advice appreciated, here is the idea:
The program has 4 essential specifications:
it is open source and hopefully written in a relatively legible language like python to ease distributed source checking.
The program can take a set of mathematics question&answer pairs (an exam) and return an “specification”, a description of what the exam was testing the testee on.
The program can take a specification and generate an math exam to test the testee on the mathematical knowledge demarcated by the specification.
The program must be able to mark it’s own exams.
The reasons that these specifications are essential:
If it is not open source, the full impact of the software will be delayed by some years via patent. If it is not legible, it will be adopted more slowly and trusted less readily.
If the program could ONLY take a specification as input and generate an exam as output then it might still be cheap and useful but would not fully usurp the “trust-prestigious-instituions-know-what-they-are-talking-about” system because who is going to trust some open source exam board over MIT!?
If on the other hand the program could also take an exam or set of exams and generate a specification, then it could become the objective metric by which other exams are judged and everyone would know implicitly that the generated exams were legit because MIT etc would be using it to rate their own exams.
If it isn’t able to mark it’s own exams, we must pay a bunch of busibodies to mark the papers manually and this is an inexcusable cost given how relatively trivial self-marking is compared to the rest of the program.
I beg you to please throw together some hashed up attempt on github just to make the conjuction of these ideas unpatentable.