When building software we have learned to take into account a lot of non functionals ranging from security, scalability, testability, loosely-coupled and so on. What is missing is an aspect that handles complexity. A small company has notepad to manage contacts. A large company has an enterpise CRM system. I wonder if we could not create a repository which for starters creates replacements for basic operations: for example write-file-to-directory could check if there are more than 500 files already and if so automatically create a folder A-M and N-Z and split the files without user interaction. Nowadays these operations only look at existence of, writeable, rights and so on but not take into account complexity.
If a complete set of basic methods in all classes would take complexity as one of their design principles they would work slighly differently. But by then leveraging on this and building another layer / layers on top of this, it would create software systems that would auto-grow with users in terms of complexity. It would auto-add data elements or taxonomy when complexity rises.
My experience with reading history is that what you think as of a fact of life changes when you read the causes of that thing. This includes large topics like slavery, wars, injustice, revolutions and more. What if we would leverage on open historic material like delpher.nl (all newspapers since 1500) and transform this in cause-effect diagrams. So literally connecting everything to everything with arrows "this happened because of this and that happened because of that". That would transform history in a domino effect system and it would be more clear to students why what happened, just tracing back through the diagrams. It would also make clear that nothing stands on itself. In the end it might make history a harry seldom exact science when we could derive patterns by meta-analyzing the cause and effect diagrams and describing them on a higher level using knowledge from cognitive science, group dynamics and sociology.
Households have in general less IT knowledge. A device would be need per household contains enough storage and soem kind of generic API to interface with each household based on a standardized model. Maybe this can be fumbled in to the home-automation trend although that trend is highly focussed on automation and not core household processes such as "analyse my weekly spending on mcdonalds". Another angle could be to work together with NAS providers and inject this in this way in the market.
Every month has a name and every weekday has a name but hours and minutes do not have names. An idea would be to create a site with 1440 webpages and a forum on each page and propose a name for each minute by taking the most important 1440 humans until now. (since days and months also refer to probably once human beings (...)). This will be also a fun social experiment to see people of different cultures and political conviction debate over which important persons ever belong on the list of the most important 1440 people.
In e.g. blockchain you add a property with the hash of the parent. In general when comparing with a taxonomy when having multiple parents this becomes tedious. Would a hash over multiple hashes not be a better idea to uniquely identify a time-period instead? So not to single hashs on transaction but to hash multiple transactions as a "state". This is probably a bad description. But if you then find patterns in hashes e.g. the hash is the same because timeperiod deliver exactly the same state changes you can then store these accordingly.