Boy did I come in late. It might be a bit dated by a couple of weeks, but Ghinzdra doesn't know what he's talking about. The SRS is not an immutable system of godly memory requisition which, if gone against, cease to function on all levels until the offense has been corrected.
Sometimes, putting in a sentence which you don't fully understand will allow you to grasp it. Going through math problems proofs you don't understand from different angles over and over again will allow better understanding as well.
Long sentences are a fact of life. The difference between vocabulary words and sentences is to understand the context. If you shorten a sentence to the point where it loses context while trying to get it to qualify as a "sentence", then you're doing it wrong. Understand the reason behind why shorter sentence are generally better, and you can make the decision on a daily basis.
Thirdly, the SRS is almost entirely useless. But in order to see my argument, you have to understand the context with which I present it: Namely, the rise of efficiency in pre-mined sentences, subs2srs and the very thought of workflow within learning the language. I submit that these efficiency oriented approaches to learning a language will not hold the test of time.
It is difficult to keep up a system so mundane.
Reaching thousands of cards in your system, while quite the feat, I do not think is the primary reason for your increased skill level. Let's consider the effect of Khatzumoto's site, the koohii community, Krashen's theories and an overconfidence in the SRS. While all of these stress the importance of such things as input before output, sentences are better than vocabulary, headphones on all the time, etc., there is a secondary effect to it all. Reinforced by the community here, there are many among us who have been involved in a breathtaking inspiration. The unshakable belief, which I admit to have been seduced by as well, that learning Japanese is not only possible, but easy if the system is followed. But what is the system? That is what we still argue and debate over endlessly today. The SRS workflow, the "correct" method of collecting sentences, the efficiency in using pre-mined sentence, the efficiency in using such things as subs2srs are all a part of the system.
But after weeks, months, or even years, many of us fall. That is not to say we were not dedicated enough, we were. It was simply a matter of sustainability. Is this system, in all of its glorious perfection, really a sustainable solution? If you examine the methods of Khatzumoto, Krashen, Antimoon etc., you will find that while the outside semantics, the details, change somewhat, the internal system they preach stays the same. This means not that they have a perfect system which requires the same interpretation as something like the Bible is thought to require, but rather that their systems are not those created out of imitation. They are created with a personal customization which is simply not within reach of the rest of us.
But, to get back to my original point, there was a second factor beyond the methods. That factor was one that had us dedicate so much time to our methods, to our Japanese. It was in this dedication with which we were able to triumph. With this, even the most imperfect system will still reach its goal. The reason our systems tend to be unstable, is precisely because we get caught up in the system. We feel as if we need to do things right, we trip up over mistakes and try to refine them. Even going so far as to dedicate more time to refining the method than putting it into practice.
The idea of SRS workflow, laws and the like is ridiculous. Consider instead the exploration of Japanese, the literature and media. You will reach your goal of fluency despite the naysayers, as these are the things which you desire. This idea of maintaining an efficient and perfect system is a fallacy and must be discarded. Right or wrong you move forward and learn not by what the others tell you, but by your own experiences.
.
