That's a pretty plant, but it isnt a good example of the start of a new landrace - because it is the result of 1:1 pollination (dont mean to be snarky - I may have misunderstood what you are saying!). It has the narrowest genetic roots. In a big enough open pollinated population (ideally working with at least hundreds of plants), natural mutation will ensure that diversity increases over time, whilst plants continue to adapt to local conditions, and valued characteristics are preserved by culling off-types. Point being it needs a broad genetic base to remain both healthy and sufficiently true to type to be recognisable as the same variety over time.
You're spot on about the importance of working with varieties over many generations. And for a variety to be sustainable it needs enough genetic breadth to keep it healthy. That's not going to work with a variety that is repeatedly put through single parent bottlenecks, especially if repeatedly backcrossed.
I'm also in favour of developing lines that are kept separate from the mainstream genetic pools - and starting from landraces is clearly the way to go (rather than crossing and recrossing the latest cookie clone). Even small scale cupboard growers can contribute positively to this by ensuring that when they grow landraces or old heirloom varieties, they do open pollinations of those lines first to back up the genetics in seed form rather than just hybridizing and losing the original. And growing multiple inbred lines from a landrace still leaves the possibility that those lines could at some stage be recombined.
i would, it love it.