I feel like that if you get better at existing methods, alg trainers will make you faster in the short term though if a "quantum leap" in methods is developed, you would be better off learning the new method.
Fair enough, I probably made it seem like I'm somewhat against developing new methods - which I'm not. I might not have expressed that very well. I do think that
some people should explore existing methods though, rather than forever waiting for that "quantum leap" to arrive - there are definitely methods out there that deserve to be experimented with more. Who's to say that the quantum leap has already arrived, except nobody bothered to learn the algorithms?
As well as that, I think it is in the best interests, of anyone who is looking for that "quantum leap", to make algorithms easier to learn. If algorithm sets become easier to learn, the work required to speed gain ratio is decreased for every method, and it will allow people to switch faster. Perhaps it will also allow people like me to learn multiple methods faster and give my opinions on how they compare.
I should probably remove the line "Would that be more productive than making more new methods?". That's probably just wrong. But I do think it would be worth the time to work on a good algtrainer. After all, what point is there in making new methods if nobody bothers to learn them?
