Strongly agree that wordlists are useless (beyond the first few hundred that you are going to see constantly anyway) and making connections is key to retaining vocabulary.
Another tip is to pull vocabulary lists from reading material, and then to review them you just re-read that text until it becomes easy.
I think wordlists and learning tons of vocabulary through deliberate study are generally underrated. However, I do agree that once you've got a good foundation, pulling words from texts is a great idea.
I think the ideal thing to do is to get a list of words from a text before you read it, then prime yourself on them, and then really solidify them by reading the text. I think it makes the text a lot more enjoyable.
I actually just put together a prototype of a tool for doing this and wanted to polish it more before posting it, but I'd love any feedback / ideas from people who are into language learning here: https://langtools.curtis.io/
You can test it out with a guest account: [email protected], password "guest". Or just sign up, I don't validate your email or anything.
It's pretty rough but I think the basic idea of uploading a text, and seeing which lemmas are unknown to you so that you can pre-learn them is a good one.
I wanted to remove the login (or set up a dedicated demo page) when I planned on posting it in a week or so but this seemed like an opportune time to drop a link.
I assumed someone just changed the guest password, but it appears the database got in some strange locked state over night. It's sqlite since I want this to be as self-hostable as possible. Not sure if I somehow created a deadlock.
It should be fixed now. I'd never seen this lextutor site, I find it very interesting, it looks like its they've got very similar ideas, I found this on the research page: "One ineresting use of VP (I believe the main one for users of online VP) is to evaluate the suitability of reading texts for various levels of learners.".
That's basically what I'm doing at the moment. I'd also like to add an estimate of how many of the unknown lemma's i'm showing that you'd need to learn to get your comprehensions to 98% (or something like that) so that you'd know how much pre-studying might be worthwhile before diving into the text.
I haven't done much research into it but the 98% number gets through around a lot as an ideal place to be to pick up the remaining words through context.
edit: ah, just a few sentences down: "As a rule of thumb, learners can not do much with a text if they know fewer than 90% of its words. From 90% to 95% the text can be used intensively (for dictionary work, contextual inference, re-reads, etc.). From 95% to about 98%, the text can be used for fluency building. Above 98% the text can be used for 'reading to learn' rather than 'learning to read.'"
Another tip is to pull vocabulary lists from reading material, and then to review them you just re-read that text until it becomes easy.