First design is an ADC going directly into a FX2 or FX3 USB interface. Similar to the FX2-based logic analyzers supported by Sigrok, the record length is only limited by how much RAM your PC has. Or write a driver for GNUradio and have nearly limitless ways to do real time processing.
There are already ADC+FX2 combos, for example the
Hantek 6022BL or
Hantek 6022BE. USB HS has too little bandwidth for a serious streaming scope, so USB SS it would need to be. However, the dumb people at Cypress apparently thought "Oh hey, the FX2 was such a major success, let's make everything 10x as complicated. Customers will love it!", which resulted in the FX3 now requiring an RTOS (ThreadX) with a proprietary license that does not allow redistribution of the code. That means a completely open source FX3 firmware isn't going to happen and I don't even know if it's allowed to redistribute the library binary. Way to go, Cypress!
Either way, I agree with you. I really, really,
really want an open source USB3 streaming scope. Once someone makes it, though, it will be cloned 1000000x in China and sold for lowest prices, so there's no money to be made for the original developer(s). Not a strong motivator for anyone, I figure.
Second design uses a FPGA and RAM like a traditional digital scope does, but instead of a built in display, it uses an ESP8266 or similar to connect to a smartphone or tablet. Basically like the Mooshimeter but it's a scope.
Aside from perfect galvanic isolation, I never understood the point of wireless scopes. Why artificially limit your I/O bandwidth and add tons of latency for no other reason than "it can be done"? For examining mains circuits, those things could however be literal life savers indeed.