The other thread seems to have incredibly low visibility so I'll try to use this one to get a little feedback on what features your ideal iPhone/Touch/iPad RevTK application would include. (Before I go down a road no one cares about...)

This project included writing a REST backend to the site so the progress has been back and forth but here is what the initial 'slightly' working version looks like.

Boot screen:


Login and News:






Kanji Study Video:
http://krnl.net/videos/KanjiStudy_conv.html




Kanji Practice Video:
http://krnl.net/videos/KanjiStudy2_conv.html


One of the things I definitely want is the ability to write on the flashcard the kanji. Besides that I have no other goals than to make a very error-free application.

Anyways, thanks in advance for any suggestions!

Also if anyone was available to test (requires the freely available iPhone SDK) I'd really appreciate it. While it's not quite at that level yet, I'd appreciate it nonetheless.
John

[ファブリス]

Btw, when we say iPhone app it also means iPad app these days right?

The app would need to be written with iPad in mind to take advantage of the higher resolution, otherwise it will just be scaled.

That is what I gathered from the keynote anyways.

Does this really need a full-blown App though? Wouldn't it be better implemented using a webapp and offline database (like iAnki)?

From what I can gather, unless you have a jailbroken iPhone, the iAnki still requires a localhost server running on the network to use (or available over the internet).. I'm not sure what the benefit is of this? ( Let me know if I'm missing something huge )

There is nothing stopping the application from running entirely in an offline mode, but I just chose to initially have it cache the current boxes (Leitner) and query against the server to see if the there is a more recent review of cards through the website than what is cached on the client.

And since I never developed an iPhone application, I figured I'd give it a try. I'm not sure how feasible it would be to do things like handling drawing on the screen with iAnki.

[ファブリス]

Using the browser as the "engine" and adding offline support are two different things that always get mixed up for some reason. Both would be nice, and both have been requested often in the past.

Considering the "engine" question, that's a good call. A web app has the advantage that it would run on other mobiles, although the support would still be limited to the really modern phones with a browser that supports modern XHTML and Javascript, Ajax,..

With that said a native iPhone app is great too, and can certainly pave the way for other mobile support. For example John has implemented a REST api, which AFAIK is what the "mobile" web app would have used. Also I assume that iPhone apps can render WebKit views with custom html/css, in that case a lot of that will be reusable for a web app version if need be.

So my 2 cents is John should work on whatever motivates him best!

PS: news post'ed.

[ファブリス]

johnrb Wrote:There is nothing stopping the application from running entirely in an offline mode

Yeah, srsly. Let John get started already

johnrb Wrote:From what I can gather, unless you have a jailbroken iPhone, the iAnki still requires a localhost server running on the network to use (or available over the internet).. I'm not sure what the benefit is of this? ( Let me know if I'm missing something huge )

Once synced, iAnki is entirely offline despite running in the webbrowser.

Thanks for the news-post-assist!

So far I haven't use any Webkit views but I'm it's possible to do some very cool things with them. (At least more than my anticipated usage of "Letting you see the news post in it's full glory... with <b/> and <a href/> !")

Fortunately (or unfortunately?) I get to prioritize this all by myself so I have come up with the coolest animated leitner boxes graph :-P

If you guys want to write a mobile web app, by all means! The REST interface should be even easier to consume than it was in Objective-C. (I'd hope... it being JSON and all)

I'll have to download the iPad beta SDK to see what happens. I'm not familiar enough with iPhone to even try to do a multi-targeted release the first time. My main goal is just to get it out as fast and as stable as possible.

Ahh, I see what you mean about iAnki and the syncing. At some point I can tackle that but my focus for the first release will be on a feature complete consumption of RevTK's REST services. (And in the process fleshing out the REST services themselves)

Hey there, I'm happy to hear about this app! I'd really love the ability to download/sync rather than having to be connected. I only have an Ipod Touch and more often than not don't have web access when I'm using it.

Thanks!

Pretty soon I'll have no choice but to implement this! I haven't looked at how the servers algorithm works but I'll check it out if there is enough interest. I guess I was a bit naive thinking everyone would be using iPhones

Actually, I guess I misunderstood. I was under the impression that you were creating an offline version of RevTK that could be synced to the website.

If it's online-only then I think it's even BETTER of an idea to make it a webapp instead (a user would be presented with the webapp upon visiting kanji.koohii.com, instead of the normal PC interface). See wikipedia for an example. I don't really see what being an appstore app gains you (other than some experience writing native apps).

I use this website on my iPhone everyday, all the time (train, lunch breaks, breaks between classes) to review etc.

Other than having a slightly friendlier interface, (save zooming etc.) it would be great if it could be made to work offline, so that when I'm out of range (e.g. on the subway) I can still use it. This would be great for iPodTouch people too.

頑張って！

[ファブリス]

@Jarvik7: I agree in general but I think the native app may prove to be much more feasible in the short term than a html web app providing the same user interface experience. I am guessing that the native app SDK offers standardized icons, buttons, and other widgets which facilitate the development and provides a guaranteed user experience quality . With a web app, you need a lot of experience with HTML and Javascript and also SVG to make the most out of it. You can use Javascript frameworks but none of them provide widgets with the MacOS native feel.

Fab: I think Apple provides a JS toolkit to make webapps look native. Check out the iPhone keynotes from before there was an appstore. The demonstrated webapps all look pretty native and consistent with each other.

John,

I think the iPhone app is a great idea. I have a few apps on the store right now and am quite familiar with Cocoa and Objective-C. I've would love to have a native iPhone version. If you need any assistance, please let me know. I'd like to help!

David

I also think that the best thing about an iPhone app would be an offline mode. The regular website works well on the iPhone as is - an app or webapp that would make it slightly less cumbersome would be welcome, but not necessary!

Thank you for the effort!

[ファブリス]

@Jarvik: It seems that SproutCore is linked to Apple in some way, though I don't know if that's the framework you meant.

SproutCore looks good! It's on Github too, very interesting.

Damn, I'd try that for the main site but it says HTML 5, can someone say if the demos run on IE? I have yet to install an emulator on the iMac..

PPS: it looks pretty amazing what you can build with SproutCore. I saw it a while ago and didn't get it, because their demos are really boring. MobileMe appears to be made with SproutCore.

While I'm not adverse to making it with offline capabilities, how will things like stories, etc work? Will only some features work when you are online? Otherwise you cannot enumerate the available stories when you are reviewing your failed cards.

We could make it so that the client only caches 'his' story offline I suppose. I'm up for suggestions.

Also I like the idea of the RevTK server (kanji.koohii.com) being able to push to the phones that new cards are available and such using the whatever-it-is-called-push-type-service.

ファブリス Wrote:Damn, I'd try that for the main site but it says HTML 5, can someone say if the demos run on IE? I have yet to install an emulator on the iMac..

Quote:MobileMe is accessible on the web via Safari 3, Internet Explorer 7, and Firefox 2 or later.

I'm guessing HTML5 is only required for certain features.

Also, while Sprout is probably a part of it, I think there are some iPhoneOS specific HTML/JS extensions and perhaps some local graphics resources you can rely on.

I haven't coded any serious HTML for many many years so don't take my word for it though.

re: offline mode

johnrb Wrote:how will things like stories, etc work? Will only some features work when you are online?

Here's my suggestion for how offline works:

The iPhone stores locally:
1) the state of every card (stack/last review)
2) your story for every card
3) the top X highest rated stories for every card that you don't have a story for (or for all cards)

Then, offline, you can review cards as much as you want - the app would use the same expiration logic as the site, with sync when you return online. You'd have to work out some conflict logic for the sync - probably just use the most recent state for each card if there are differences, with a report ("Sync successful. 10 cards updated. 7 cards had been reviewed more recently at koohii.com and were not updated.")

Offline, you can also study the stories from the local database. For simplicity, the study section could be read-only offline. Or it could be editable, with a similar sync logic.

Thoughts?

It sounds good to me!

That is sort-of how I had planned on doing it my head. The only thing that I'm not sure about is whether ファブリス minds so much of the sites database being pulled to the clients machines. If it is okay with him, then that is how I'll approach it.

How would you handle two cards, both being reviewed but differing values? (One reviewed on the PC, the other on the iPhone). Should you just go with the latest value? Or by default count it as a miss...

Anyways I appreciate the suggestions!

John

You could reduce the amount of database query and server load by including a database of the top 5 stories for each kanji in the app itself. Then, when you view any kanji you could have a 'refresh' button which checks if the top 5 stories have changed and downloads the new ones if there has been a change. The app would then only have to download a user's personal stories upon login.

As for two conflicting card review data, you should take the lowest scoring result, or an average of the two. This avoids the pitfall of the SRS thinking you've remembered a kanji well when in fact you just saw it twice in a short space of time.

I'd be interested in doing testing. I don't develop in Obj-C but I can read it. I'm an iTouch guy, though.

Not that I'm being selfish... but making the app an off-line synchronizing app would open the door to a LOT more users.

Yeah, sorry for the delay in updates.. I was at a wedding in Phoenix.

I've gone ahead and focused on making it "off-line"-able.

I should have something fairly soon for you to test as long as you can compile the source. (Mind you the testing will only work in the simulator)

After it's a bit closer I'll start allocating actual testing licenses so people can test it on their actual device.

I really appreciate your help!

John