Whilst I agree with your point that digital is by nature ephemerous, relying on ever changing layers of software and hardware, I feel you rule out its feasability a bit hastily.
1./ What if we adjust the scope of the challenge? If we identify what exactly in a project is to be “kept for eternity”, perhaps a feasible way towards a solution could build up. Perhaps a useful first step would be to define constraints for the digital design gallery. “Looking for inspiration” sounds a bit vague and an even doubtful purpose, as you mentioned in a previous post how inspiration is overstated, what one really needs is to (paraphrasing here) get into the process. Perhaps the starting point (a proper brief) is needed before ruling it out entirely: objectives, user scenarios, resources, etc.
2./ I agree with you that freezing an entire digital system (software + hardware) is hardly possible, because the context justifying the form will have changed. Yet it doesn’t imply that we shouldn’t properly record, say, that interesting combination of alternative colours offered to colorblind folks, or that ingenious solution for guiding the user on a lengthy profile form, etc. There are so many sub-systems composing an entire digital project. Many of these sit inside a whole system that is not so remarkable in itself: only parts of it are worth archiving as “great design artefacts”.
I think littlebigdetails.com is really onto something in term of content and ubiquitous capturing format (I don’t see Animated Gifs become unsupported any time soon, nor the hardware/software necessary to display them). LBD does lack a proper “library-like” interface, adapted to the purpose of a library: perhaps an interface that illustrates the fact that “the medium is indeed the message”. Perhaps it should have no visual interface, controled by voice “ok library: show me …”. That would make sense, given the current trend towards conversational interactions. That may mean the design of the library should focus essentially on its content architecture, externally accessible via an API. Then pluggable interfaces, that could be outsourced, for any designer to play with this content. That would be very Marshall McLuhan-esque, and a statement to the power of design: to turn data into information, information into content and content into stories.
littlebigdetails.com’s content also lacks how those bits of UI affect real users: perhaps that would be better captured by UX testing recording excerpts. Your Digital design library could take that into account, if it is in its requirements.
So, there are solution depending on what you want to document: a clever solution to a tricky problem (it should explicitate how it was deemed “clever”), an elegant solution (it should then document why it was deemed “elegant”), perhaps also failed attempts and unsolved problems, why not. So much is to be learned from failed attempts, if a collection of learning material is in the requirements of your library.
Like you, I see many, many ways in which this endeavour could fail. Perhaps too many to even try. But by toning down your expectation and framing correctly what you really need out of this library, perhaps a more modest, but elegant solution could be worked out.
°-*
