What to do with 30 extra MB...

[mikelove]

So Apple just bumped the 3G download size limit from 20 to 50 MB, which means we now have 30 extra MB of built-in data file space to play around with. We're going to use about 5MB of that for basic 2.3-related stuff (like the code libraries for our new text-to-speech feature, and a few new graphical assets) but that's still 25MB that aren't spoken for.

So: what sorts of data files would you like to see us use this extra space for? Pretty much anything can be an add-on download too, including our new font files, so this doesn't really impact which features will be in the product - it's mostly an issue of which features / functions are important enough to deserve to be built into our app from the moment that people download it.

For example, should we build in a version of CC-CEDICT to make up for the vocabulary gaps in the PLC dictionary? (11 MB) We could still make it updatable between app versions. Should we put in all of Unihan so there isn't any more fiddling around with the "extended version"? (2-3 MB) Should we build in basic single-syllable audio (9 MB), or add a demo version of OCR so people don't have to download that (2 MB)? Maybe build in some useful flashcard vocabulary lists (1 MB), or some interesting reader documents (1 MB)? Should we build in the instruction manual (13 MB, but only 1 if we make it text-only)? Should we build in a Kai font file and default to it for headword display? (13 MB)

 

[rcloud]

Great news Mike. My vote is overwhelmingly for the Kai font.

The OCR is a really cool feature so a built in demo seems like a good idea that should lead to some more sales for you.

 

[character]

I agree it should drive sales -- my SWAG is that basic audio and OCR demo would be the way to go. Having audio present from the start might enable Pleco to give an even better first impression.

Audio from Pleco always gets a good reaction when I demo it; OCR leaves people amazed and trying to fit an iPad and Pleco into their budget.

 

[mikelove]

Great news Mike. My vote is overwhelmingly for the Kai font.

Well it's a big one but it would make for some awfully pretty screenshots.

The OCR is a really cool feature so a built in demo seems like a good idea that should lead to some more sales for you.

Indeed - my biggest worry with that is that people will be less likely to discover the "Add-ons" tab if they download our app looking for OCR and it's already built in / easy to get to, and hence have a hard time figuring out how to purchase it (or other add-ons) when they're ready to buy them.

Audio from Pleco always gets a good reaction when I demo it; OCR leaves people amazed and trying to fit an iPad and Pleco into their budget.

Audio is a nice easy-to-understand demo - male or female?

 

[character]

Indeed - my biggest worry with that is that people will be less likely to discover the "Add-ons" tab if they download our app looking for OCR and it's already built in / easy to get to, and hence have a hard time figuring out how to purchase it (or other add-ons) when they're ready to buy them.

The demo OCR could display a dialog indicating it's a demo, and point out how to get the full version. The app itself could put up a dialog on each launch until the user opens the Add-ons tab at least once. Such a dialog could rotate through the various bundles/add-ons available. You're giving away the base app for free, most people expect some effort to sell them something. The trick is to not tick off too many of the entitled 'all your hard work should be free' jerks who will give you a bad rating for doing so.

OTOH, how about embedding an OCR demo video instead of OCR?

Audio is a nice easy-to-understand demo - male or female?

Female, assuming the majority of your customers are male.

 

[scykei]

If you're going to include voice support, be sure to fix the possibly annoying pop up when you try to have it pronounce words with two or more characters. Instead, you should make a less obscuring message that does not make it feel like you're trying so hard to prompt the user into purchasing the add on. I always feel that unless you have already acquired any sort of add on for any app, the button for it shouldn't show up. Having a button which you can easily access but not being able to be used until you actually buy the feature will make the app feel like a cheap demo.

It would be a different story if the user knowingly downloads this demo add on. That way, it will not affect the user's first impression on the app.

Pleco seems to handle all of these stuff pretty well right now though, like how the flashcards and OCR options aren't there until the user buys them.

Just something to keep in mind if you're going to include them.

 

[mikelove]

The demo OCR could display a dialog indicating it's a demo, and point out how to get the full version. The app itself could put up a dialog on each launch until the user opens the Add-ons tab at least once. Such a dialog could rotate through the various bundles/add-ons available. You're giving away the base app for free, most people expect some effort to sell them something. The trick is to not tick off too many of the entitled 'all your hard work should be free' jerks who will give you a bad rating for doing so.

We actually do do that with the startup alert already - not specifically to go into Add-ons but to access the menu button at least (at which point the way to access Add-ons should be pretty obvious). But a more aggressive demo alert might be an effective replacement, yes.

OTOH, how about embedding an OCR demo video instead of OCR?

No way we'd have enough space for that, unfortunately. We could see about embedding a scrollable OCR-friendly still image and letting people try out the full recognizer with it, but I think most users would be highly skeptical unless they had the opportunity to try it out with their own content.

Female, assuming the majority of your customers are male.

Not that big a majority, but yes - the fact that the female audio isn't quite as high-quality worries me a bit, though.

 

[mikelove]

If you're going to include voice support, be sure to fix the possibly annoying pop up when you try to have it pronounce words with two or more characters. Instead, you should make a less obscuring message that does not make it feel like you're trying so hard to prompt the user into purchasing the add on. I always feel that unless you have already acquired any sort of add on for any app, the button for it shouldn't show up. Having a button which you can easily access but not being able to be used until you actually buy the feature will make the app feel like a cheap demo.

We're actually dropping that restriction altogether in 2.3 - you'll still have to pay for the extended audio databases (which are where most of the money / time went when we originally developed them anyway), but syllable-by-syllable will now be free. (mainly driven by the introduction of our new text-to-speech add-on)

It would be a different story if the user knowingly downloads this demo add on. That way, it will not affect the user's first impression on the app.

Quite right, we don't want too many demo warnings popping up before they've actually downloaded a demo (and hence presumably also discovered where they can go to delete it).

Pleco seems to handle all of these stuff pretty well right now though, like how the flashcards and OCR options aren't there until the user buys them.

Just something to keep in mind if you're going to include them.

Very good point to remember, thanks!

 

[yoose]

I will throw in my 2 cents. I think it is fine right now and anything that isn't essential shouldn't be added. Of course having optional demos and fonts would be nice, but a lot of people are on limited data plans and that extra 30mb can make a big difference. also the download time will increase as well which can also be a problem.

 

[goldyn chyld]

Yes, fonts would be awesome... As for CC-CEDICT, it would be nice to have it built in, but what I'd love even more would be the option to have it updated more frequently than on a 2-month basis. 8)

 

[sych]

I'd suggest putting in some flashcard lists, perhaps HSK and/or one of the common textbook sets. That would enable you to also include some sample flashcard profiles, like "Learn to read HSK Level 1" or "Learn to write Textbook XX vocabulary". I think it would go a long way towards making flashcards accessible to novices.

 

[mikelove]

I will throw in my 2 cents. I think it is fine right now and anything that isn't essential shouldn't be added. Of course having optional demos and fonts would be nice, but a lot of people are on limited data plans and that extra 30mb can make a big difference. also the download time will increase as well which can also be a problem.

Fair point - we may do some experimenting with a prompt to download some of those add-ons on startup first; in the case of CC-CEDICT I worry that our app ends up looking bad compared to other CC-CEDICT-based ones since there are words you can't find in PLC that you can find in CC. (then again, if the definitions are more accurate / have more examples it would actually look good, so the question is whether we want to emphasize that difference or not)

Yes, fonts would be awesome... As for CC-CEDICT, it would be nice to have it built in, but what I'd love even more would be the option to have it updated more frequently than on a 2-month basis. 8)

Still not quite clear if the demand is out there for that - to be honest, looking at the actual diffs between CC-CEDICT versions, it doesn't seem like there are massive enough changes in a typical week to justify doing weekly updates.

I'd suggest putting in some flashcard lists, perhaps HSK and/or one of the common textbook sets. That would enable you to also include some sample flashcard profiles, like "Learn to read HSK Level 1" or "Learn to write Textbook XX vocabulary". I think it would go a long way towards making flashcards accessible to novices.

That one feels to me like one of the most appealing uses for the extra space, yes.

 

[iphneusr]

Why is KaiTi still not included?!?

 

[mikelove]

Why is KaiTi still not included?!?

We only just signed a license for it, and that license hasn't even taken effect yet. We'll probably introduce KaiTi support in our next major update, though it may be as an add-on download instead of built into the app.

 

[Lurks]

My God, KaiTi font... you know where I stand on this

I wonder if there isn't something out there like a tagged dataset of Chinese with linguistic information, parts of speech and semantic sets, hyponym/hypernym relations etc. Lots of stuff around for English but I've yet to find good sets for Chinese and I guess it might not be hugely useful to non-linguists without a pretty fancy UI front end anyway?

Mat.

 

[mikelove]

My God, KaiTi font... you know where I stand on this

Indeed, though since we already support installing custom fonts on Android (and have for a while now) I imagine you've already got it working on that device at least.

I wonder if there isn't something out there like a tagged dataset of Chinese with linguistic information, parts of speech and semantic sets, hyponym/hypernym relations etc. Lots of stuff around for English but I've yet to find good sets for Chinese and I guess it might not be hugely useful to non-linguists without a pretty fancy UI front end anyway?

There may be a couple of commercial ones floating around but we haven't seen a good one that we can use yet. I'm actually surprised the Chinese government isn't funding more work into this - it seems to me like they have a vested interest in promoting Chinese as a global commercial language, and providing good open-source linguistic data would probably help with that.