I'd say that the best is probably the book "Lisp in Small Pieces".
Be aware that it's some 500 pages and makes several interpreters then a couple of compilers as I recall. If you read it though, you'll come out the other end with a pretty good understanding of how lisp implementations actually work.
It takes a very thoughtful approach to introducing an increasingly complex Scheme implementation. I doesn't shy away from the complexity that many LISP implementation tutorials try to push under the rug.
oh. the measure is how much code was used to implement it. at about 200 lines of javascript (to parse the text, convert it to json and render it as static HTML) it's at least a thousand times fewer lines of code as the early scala code that implemented twitter.
[edit - though it looks like handlebars (that I use for generating HTML from templates) has about 30k lines in files that end in .js, so maybe twitter is 30m lines of code?]
[second edit - hmm... maybe i'll change the name to "mini-blog" to signal it's a blog that's 25% the performance of a "mainframe-blog" but 10% the cost. though "mini-blog" might be more appropriate: halfway between a "normal" blog entry and a "micro-blog."]
https://github.com/kanaka/mal
Discussed here, a few times:
https://news.ycombinator.com/from?site=github.com/kanaka