I spend a lot of time thinking about this. What knowledge would we wish we had saved if civilization fell?  How much could we preserve if we collected all and only the most important knowledge and made long-lasting copies distributed throughout the world? What should we load onto this ark, how much room can we afford, how many copies can we make, how long can they be preserved?

Some of the top things I think need to be collected and preserved:

Agriculture and animal husbandry, with particular attention to soil and water management, appropriate technology from digging sticks to combine harvesters, as well as storage, milling and other processing. This is the most important subject – without food, nothing else can happen.

Languages – a picture dictionary for the base vocabulary  in the top languages, plus a comprehensive dictionary with definitions that bottom out in words from the base vocabulary, and self-playing recordings of the phonetic alphabet (string phonograph recordings with a built-in player), plus translation lists for a few thousand words in each of the top thousand languages, including all the dead languages which are known.

Mathematics – built from pictures on up in a self-contained manner, focusing first on concrete and applied math as illustrations of more general principles. “Dictionaries” for translating between different mathematical dialects.

Physics – integrated with the mathematics curriculum, presented in a uniform mathematical dialect. Emphasis on EM and materials science rather than cutting-edge theory.

Chemistry – beyond the basics into heavy chemical engineering, mapping out the dependencies, alternate paths, and pitfalls of synthesizing and using the most economically important compounds. All the data on the principal properties of compounds that we have, similar to an expanded version of the CRC handbook

Geography- as much of it as will fit, with a particular emphasis on locations of ores, winds and currents, and biological resources.

Geology – emphasizing economically important ores and other resources, field identification and labwork, but also stratigraphy and both uniform processes and catastrophic events.

Astronomy – orbital elements of the top few thousand objects in the solar system with special attention to the long-term, data on the compositions and forms of the planets,  the proper motions and spectra of the nearest and brightest few thousand stars with additional data on nearby galaxies, quasars, pulsars and variable stars, information on the structure of the Milky Way and galactic clusters.

Logistics and project management- crucial, but I don’t know enough about it to say anything. Accounting, economics, operations research

Manufacturing processes – starting from rocks, ending with microchips and jet engines. This will take a large fraction of the whole library. Getting some of the practical information will be difficult, much is trade secrets and implicit knowledge.  Detailed designs and processes for making crucial machines at a variety of technological levels are needed. Knowing what has been tried and found not to work will also be valuable.

History, anthropology, archaeology – an amount similar to or greater than that on languages.

Biology and medicine – too much in these areas to include everything. Most detailed biochemistry that does not have current applications will have to be left out. Emphasis on species identification, ranges and behaviors, microbiology, disease diagnosis, pharmaceuticals, surgical procedures, other treatments, both human and veterinary.

Culture – plays, novels, tales, myths, legends…

This is just scratching the surface, a sketch of an outline of the table of contents. Such a project will need thousands of people working for decades, maybe even centuries. Perhaps Wikipedia could serve as a nucleus for this more focused, detailed, permanent and expensive work.
Though it needs so much time, effort and cost, I think it will be the best thing we can do for the future.

http://www.econlib.org/library/Essays/rdPncl1.html
The proliferation of TED conferences is leading to videos like that.

http://immaculatetelegraphy.tumblr.com/

you need to know everything there is to knowbout oil drilling if you claim to know how to make from scratch something that contains plastic

That depends on the plastic. Celluloid or bakelite could be made without oil.

Really, if you wanted toast in a post-apocalyptic world, you’d really just be better off with an old-fashioned, pre-electric toaster, which held bread in a metal grille so you could toast it over the fire

Pure decadence. A toasting fork is more than sufficient. You can even make your own from a stick.

Pretty sure the object here was to do only the toaster from scratch.

On that basis, I reiterate that many could do a much better job…

Cheers for the tip, I’ll look up David Gingery : )

Similarly, one could describe the process to go from mined ore to a toaster that has some hope of actually looking good and working properly. It would take a person more than 9 months to master the necessary skill set. More like 9 years.

I’d be interested to see this attempted by someone with some skillz.

building a bandsaw out of wood

Just as long as you never ask, in that perfectly Murphy-tempting way Jeremy Clarkson has, what could possibly go wrong… ;)