In the future it would be nice to be able to add custom example sentences with audio.

I read that a future implementation may support tatoeba, but this would be a problem for audio that has copyright.

In cases where we bought a book (for example) with audio CD's, we are entitled to make personal recordings under copyright "fair use" for educational purposes, however we are not allowed to publish them. So, uploading the audio with the sample sentence to Tatoeba is not an option.

It would be great to be able to study the words in the context of the course materials that we are using. This would be invaluable for comprehension exercises.

I understand that this might require allocating storage for each user, which might be prohibitive.

A couple of alternatives spring to mind, would it be possible to reference audio sample files on local disks?

A completely seperate approach would be to choose a good comprehension text-book or audio-book and licence it for all users. This might actually appeal to a number of publishers, as it would encourage demand for their books.

I think Skritter is a superb tool. With support for comprehension it would make a complete self-study system.

I recently sent a related E-mail to them about the iBook 2. I thought it would be really good if there was a way for flash cards made in the iBook 2 to be imported into Skritter. If there is a way to make audio flashcards (I don't know if there is) with the audio supplied by the book then that would make importing them even more useful.

But for right now Skritter desperately needs to get the example sentences from Tatoeba.org as GDFC (God Dam Fucking Context) is the only way of knowing the correct reading needed for Kanji in Japanese.

I think you might still be dissappointed by the example sentences on tatobea for Japanese. I have been doing a few translations there recently (from Japanese to English), and found that in many Japanese example sentences the context was underspecified. This is perfectly normal in Japanese, but without a larger piece of text, there is often no definitive translation. I think you may well still be frustrated.

The "Read Real Japanese Fiction" series has audio, bi-lingual text and grammer notes. An electronic version integrated with skritter would be superb for Japanese study. Maybe the iBook2 concept could support something like this in the future.

At least for personal use you could rip the audio and type in the vocab lists manually as a quick solution, if only we had the ability to integrate our own audio samples. This is exactly the kind of thing I am looking for in Chinese. I have the audio tracks (legitimately purchased), I can type in the vocab, but I what I can't do is upload my audio as context sentences. However I couldn't upload the audio to tatoeba without infringing copyright, so tatoeba does not help.

Interesting suggestion! Hadn't thought of that one before. Will keep it in mind for when we start the example sentence system overhaul, to see whether it makes sense to work audio uploads into it.

Example sentences are great, but I also have found Tatoeba to be unreliable sometimes, either from lack of context, unnatural sentences, or flat-out errors. Don't get me wrong though, it's a lot better than nothing.

Anyways, I came across a good example of what ddapore99 is talking about today. The kanji compound 都合　（つごう） can mean "convenience" when serving as a subject, topic, or modifier; "circumstances" or "reasons" when marked with the particle で; "fortunately" when used in the phrase 都合よく; or "to arrange"/"to make arrangements" when used as a verb, as in 都合する.

Sure, you can put all of these meanings on the same flash card, but it's difficult to remember all of those different associations with the same compound unless you have some kind of context to distinguish them in your mind. Think of a multipurpose verb like 掛ける:

[to hang (e.g. picture), to hoist (e.g. sail), to raise (e.g. flag); to sit; to be partway (verb), to begin (but not complete); to take (time, money), to expend (money, time, etc.); to make (a call); to multiply; to secure (e.g. lock); to put on (glasses, etc.); to cover; to burden someone; to apply (insurance); to turn on (an engine, etc.), to set (a dial, an alarm clock, etc.); to put an effect (spell, anaesthetic, etc.) on; to hold an emotion for (pity, hope, etc.); to bind; to pour (or sprinkle, spray, etc.) onto; to argue (in court), to deliberate (in a meeting), to present (e.g. idea to a conference, etc.); to increase further; to catch (in a trap, etc.); to set atop; to erect (a makeshift building); to hold (a play, festival, etc.); (after -masu stem of verb)indicates (verb) is being directed to (someone)]

Without knowing the direct object of this verb or at least some kind of context, there's no way to narrow down the meaning. So being quizzed on a card which has all or some of these meanings on it really isn't doing a whole lot of good. Better would be to somehow have different cards with one meaning each on them that's inside of a sentence, so you had to figure out the answer for that specific context. I guess we could always make our own decks and write the sentences for them, but that would be quite the endeavor.

I agree completely with you on the need for context. Certainly a sentence is better than no context at all. I guess my point is that in Japanese, in many natural cases the context may well be underspecified even at the sentence level. A coherent chunk of text is even better than a stand-alone sentence. I know that Japanese is at one end of the spectrum in this regard. Some support for larger chunks of text, would be great in Japanese, and also useful in other languages.

My personal focus at the moment is Chinese. The audio samples sentences on Skritter are ... less than ideal. I have lots of audio sources, just no way of integrating them with Skritter.

I think using full sentence audio is the best since a single word out of context will often get a different inflection in a sentence. I used to study "Japanese the Spoken Language" at college and even though I hated the book (because of the font, the old version of romaji they used, the dull dialogues, dull drills, and the dated 60s japanese) I have yet to see it's equal when it comes to pronunciation. It showed the pitch for everything in the book and had native audio recordings for everything in the book and video in the first book.