This another question/comment where the response grew really long. Here goes:
A quick question when you get back from your break.
When I started using my SRS, I entered the English sentence in the question field and the Japanese translation in the answer field. But after reading your method on this site, I no longer try to translate – as you say, this is too failure prone, so now I just put the Japanese in the answer field and leave out the translation.
The only concern I have in doing this is that I am not actively using my brain to come up with the sentence I want to say – I am just reading a sentence that has already been created. It seems like the easy way out. It’s like for example asking somebody to write a Japanese email for you and you may understand it perfectly BUT you would have learned and retained a lot more if you had gone through the thought process by writing the email yourself.
I hope this makes sense?!
Hey Mark. Thanks for your comment.
It seems like the easy way out.
It does. It really does. And I had the exact same reservations as you.
you would have learned and retained a lot more
This is the power of input. Input has a lot going for it: Input actually allows you to learn MORE, and more correctly, than just output. There is simply a limit to how long of a sentence (and how many sentences) you can actively recall in a single day, but with input you can learn to read sentences of a much higher number and greater length. Part of why this is important is, Japanese has the structural power to create very long (yet very understandable) sentences, compared to English. Output would tend to bias your Japanese toward short sentences.
[a] When you get into real Japanese, it’s nothing like English structurally, and nothing like the textbooks. I fear that outputting would lead a person to remain tied to English structure.
[b] There are many (hundreds) of ways to say the same thing in Japanese. These ways are not superfluous; they each carry a different nuance. I worry that output-focused practice would force one to either have to create a lot of context for each SRS entry, or learn a limited range of vocabulary/phraseology/usage.  There are words you need to be able to understand but never need to be able to use. Input lets you catch these where output would lead you to side-step them.  Writing an email is not practicing a language. It is using a language. To paraphrase the folk at AntiMoon, if output is so much practice, why don’t we go practice some Dzhongka right now. Come on, let’s write an email in Dzhongka! Practice makes perfect! Writing an email is a demonstration of ability, not a creator of ability as such. What, are you going to create the text, format and conventions of Japanese email out of thin air? Of course not, you have to read (input) them first. There comes a time to “draw” (output), but first you must look at the subject. And the more accurately (native-like) you wish to “draw” the subject, the more closely you must look.  Input vocabulary/phrasing/structure eventually bubbles its way up to your active vocab. It just does, naturally. Just like when you watch (just WATCH) a movie over and over again until you can talk along with the characters. You never practiced speaking with them, it just starts to come out:
“Little ducks, there’s trouble in Russia. So they called us. And we’re going over there and bringing the most lethal killing machine ever devised. We’re capable of launching more firepower than has ever been released in the history of war. For one purpose alone: to keep our country safe.”
“Did you order the ‘code red?!!'”
“You don’t have to answer that question!”
“I’ll answer the question. You want answers?”
“I think I’m entitled.”
“You want answers?”
“I want the truth!”
“You can’t handle the truth!”
“Perhaps it is fate that today is the Fourth of July, and you will once again be fighting for our freedom…Not from tyranny, oppression, or persecution…but from annihilation. We are fighting for our right to live. To exist. And should we win the day, the Fourth of July will no longer be known as an American holiday, but as the day when the world declared in one voice: We will not go quietly into the night! We will not vanish without a fight! We’re going to live on! We’re going to survive! Today, we celebrate…our Independence Day!”
I watched that famous scene in A Few Good Men (YOU CAN’T HANDLE THE TRUTH! Who doesn’t love that line?) and all of Crimson Tide and Independence Day so many times that lines like this are stuck in my head, even though it’s been some 10 years since I last saw CT or AFGM. As you use input in conjunction with an SRS, patterns and phrases will start to appear in your active repertoire, you’ll start using them, perhaps without even fully knowing where they came from; this is something I am experiencing with increasing intensity: I am using more and more new Japanese words and phrases correctly with less and less conscious awareness of where I’m picking them up from; they’re just there, as if by instinct — of course, it’s no coincidence that I remain surrounded by Japanese audio, video and text. If you just keep doing SRS reps without trying to force it, whole phrases, whole patterns, will come out by themselves, as wholes and correct. You’ll be there one day, taking your clothes out of a washing machine, and realizing that you put bleach on your favorite black cargo pants, and you’ll go: “何じゃゴリャあああ?!”. Or your friend will be pasting mayonnaise onto his kosher hot dog like there’s no tomorrow, and you’ll go “お前、どんだけマヨネーズ掛けるっちゅうねん！”, all because you’ve seen and heard these things used in the right contexts. It’s not magic or anything, it’s just some kind of neurological process at work, I don’t fully know what goes on, I just know it works. It’s unconscious, so it’s hard to test, so schools don’t like it. But it’s happening, if you just keep inputting, working diligently but patiently. Humans created language before there were governments or schools to regulate it. We do it naturally. It takes a while to build/develop this “instinct”, but input is the best way I have come across so far. Input is not all passive: for example, you still have to actively recall kanji readings.  Incorrect output, if not corrected, leads to bad habits. Massive input automatically leads to correct output with a very low error rate. Perfect for the person who isn’t surrounded by enough people who have the combination of time, compassion and condescension to correct her at every turn.  I’ve found myself more able to retain and produce Japanese as time has gone on. This is due to massive input. So, input has helped me not only to output the things that I have input, but to build my ability to recall and output new things. I know what words fit where, where one word would be appropriate and another wouldn’t. And this is crucial. To use English as an example: It’s a “building site” NOT a “building place”. It’s “a place where I belong” NOT “a site where I belong”. Now, there are some contexts where these examples wouldn’t work, but what I’m trying to show is that words that mean almost the exact same thing can still be quite inappropriate to interchange. Massive input will lead you to see these patterns because of the quantities involved. I don’t know whether output can or will do that for you.  You’re of course free to try doing things any way you want. Some people like doing output-centered work because they get usable vocabulary quicker. This is a valid point. BUT, what about their comprehension? What good will it do you to ask for directions if you can’t understand the response of that old man who speaks in a regional dialect and, in fact, cannot or does not switch to “standard” speech? As I see it output-centered practice offers narrow, short-term benefits at considerable expense to the wider and more long-term issue of complete fluency — reading, listening, writing, speaking. But that’s just me. Let me suggest that you (and anyone else), give input a chance and not ignore it just because it’s counter-intuitive and seems too “easy”. It took me a long time to come around to it, and I was initially more than a little skeptical (“it can’t be this easy”), so much so that even when I started doing input, it was with curiosity (“let’s just run this for a while and see how it turns out”) rather than faith (“omigosh this is SO gonna work, and a little while from now I’ll start a website about it!”). I think this skepticism arises at least in part because:
[a] Output, especially speaking, is more dramatic and impressive that input. Anyone can see and hear you speak, but if you just say you understand something…
[b] Input generally precedes output in learning — we learn to understand before we learn to use, therefore it seems like a “lower” level. Why not just jump straight into the driver’s seat and start outputting, right? (Answer: because you’re going to cause an accident!)
[c] Passive memory — the ability to understand but not produce — is a stage of forgetting a language; it’s part of the process of losing proficiency, so again, there’s the temptation or tendency to misunderstand it, to look down on it since it’s associated with decay and regression, rather than progress and improvement.
[d] Our society (whatever that means) values being “pro-ACTIVE”, it values being a producer, going OUT and putting OUT, a leader, not an observer, not a fence-sitter, not a watcher…it doesn’t so much value the receptive, analytical, contemplative aspects involved in input. This doesn’t mean that people don’t spend vast sums of time and money on TV and movies, it just means that these things are not highly valued. Admittedly, this is fuzzy and a huge generalization, but…it’s just something that hits me. 
I am just reading a sentence that has already been created
Perfect. This is not a bad thing. In fact, the more, the better. Remember, you need native-like Tanaka Tarou Japanese, not Mark Quinn Japanese. The time for Mark Quinn’s idiolect will come. But it will come after he knows what he’s doing, after he’s observed (heard and listened) and internalized Japanese-like Japanese. Like the writer who breaks stylistic conventions and rules for effect, first you have to know what the rules ARE. There’s a huge difference between linguistic rule-breaking brought about intentionally and rule-breaking due to ignorance. One is genius, the other is, well, ignorance. You have no business creating your own Japanese until you see how masters of Japanese — native-level speakers — create theirs. Each email/piece of output is like a dish (of food). If you want to make it edible (legible) to Japanese people, you first need ingredients and a recipe (better yet, taste some, see what it looks like, watch someone make it, and cetera!), and early in your cooking career, you need to stick to the recipe rather than substituting the ingredients. If you hope to make something acceptable to a Japanese person, then you can’t just wing it right from the start — that’s leaving your success to random chance, and your chances of success in such a situation are slim to none. Get the right ingredients, follow the recipes. And when you can consistently make something good according to a pre-created recipe, THEN start making your own variations, THEN start winging it. I’m sure this metaphor falls apart at some point, but hopefully it makes sense where it needs to.
Note that the way to learn these rules I’m discussing is not to go look up a list of rules of Japanese. Most of the rules I’m talking about are not explicitly written, but they do exist; they are real and tangible. They’re hard to express, but people instantly know when they’re being broken. In comedy they’re often intentionally bent.
so now I just put the Japanese in the answer field and leave out the translation.
Yeah. You can have [Japanese question field] and [Japanese answer field + (if you are a beginner) an authoritative English translation from the source text]. You will, of course, want to move away from English altogether as soon as poss.
Sorry for the over-long response; I’m wanting to make the issue clear for whoever else comes to read. Have fun with your studies and feel free to ask any other questions.