Google ARC Welder

[etm001]

Hi,

I recently read about Google's App Runtime for Chrome (ARC) and the newly released ARC Welder, which packages Android APKs for use in Google Chrome OS and/or Chrome browsers. I'm curious to hear your thoughts about this technology, and whether you see it as the means through which Pleco could be brought to the desktop one day.

Thanks!

 

[mikelove]

In theory, but at the moment our doing anything to explicitly encourage it might be seen as violating several of our license agreements with publishers - we'd have to either work out terms with them for a desktop version or explicitly exclude their dictionaries from an ARC Welder release in order to make it viable license-wise.

Also, to be honest, even with ARC Welder it would take about the same amount of work to port our iOS app to Mac OS X as it would to optimize an Android app for ARC Welder and for desktops (in both cases the core code changes are minimal but there's a goodly amount of UI work), and while it would support fewer users I think the end result with the native Mac version would be much more satisfying. (I'm not sure if either one would be as useful as a web-based version of Pleco, though)

 

[jaykunming]

Does that mean you're considering a web-based version? If so, that would be great!

 

[alex_hk90]

In theory, but at the moment our doing anything to explicitly encourage it might be seen as violating several of our license agreements with publishers - we'd have to either work out terms with them for a desktop version or explicitly exclude their dictionaries from an ARC Welder release in order to make it viable license-wise.

Also, to be honest, even with ARC Welder it would take about the same amount of work to port our iOS app to Mac OS X as it would to optimize an Android app for ARC Welder and for desktops (in both cases the core code changes are minimal but there's a goodly amount of UI work), and while it would support fewer users I think the end result with the native Mac version would be much more satisfying. (I'm not sure if either one would be as useful as a web-based version of Pleco, though)

For what use cases are people requesting a desktop version? If it's mostly for when they are reading Chinese on the web then something like a Firefox plugin would have the widest audience (Windows, Mac and Linux).

 

[Shun]

I'm assuming a desktop/web version would be reduced to a dictionary/Reader, without Flashcards, OCR, or handwriting input.

I'm looking forward to the Flashcards overhaul, is it already time to hold one's breath for the first big release?

 

[alex_hk90]

I'm assuming a desktop/web version would be reduced to a dictionary/Reader, without Flashcards, OCR, or handwriting input.

I don't see why it would be reduced at all: OCR would be useful for (image) PDFs or similar, and handwriting input for touchscreens. Similarly for Flashcards Anki has a web interface.

 

[Shun]

Yeah, OCR would work for image/PDF files, but not for Live OCR. For handwriting input, if the touchscreen is lying on the table, perhaps, but not if it's facing the user. (too much strain on the arm ) Yes, Anki does, but integrating flashcards the way it has been done on iOS/Android would probably be rather hard for a web/desktop version.

 

[mikelove]

@jaykunming - we're thinking about something online, yes; the most popular use case for a desktop version of Pleco actually seems to be creating and editing flashcards and a web-based version would naturally be better for syncing with mobile than a standalone app. (bigger uphill battle licensing-wise, though)

@alex_hk90 - correct me if I'm wrong but AFAIK there's no way to do effective DRM in a Firefox/Chrome plugin unless you're hosting your content on a server anyway. So if we did a web-based app we certainly might accompany it with a browser plugin, but we'd need to have the web version in place first.

@Shun - there's 3.3 to get through first, with the new version of PLC + some improvements to support other impending dictionaries + (probably) a long-awaited overhaul to our purchase / registration management system to deal with the fact that something like 80% of our support email relates to purchase reactivations. 3.4 will probably be the first release with something that feels like a flashcard overhaul, though it'll actually be spread out over several release (probably even numbered will be flashcard-heavy and odd-numbered heavy on other areas).

 

[Shun]

I see, thanks! Do you think it would be possible to expedite the development process for various projects by employing additional very good programmers, or would the quality of the software then be bound to suffer? Of course, the amount of functionality gained would also have to increase the user base/revenue per user correspondingly. Then again, if one reads how Apple/Google have to fight for their talent, such developers may be few in number, let alone with a knowledge of Chinese. I'm not one, anyway.

 

[alex_hk90]

@alex_hk90 - correct me if I'm wrong but AFAIK there's no way to do effective DRM in a Firefox/Chrome plugin unless you're hosting your content on a server anyway. So if we did a web-based app we certainly might accompany it with a browser plugin, but we'd need to have the web version in place first.

I don't know but it would be against at least Firefox's principles so I guess you're right.

 

[mikelove]

@Shun - yeah, to be honest, that's a bit like the old Steve Martin joke where he explains how to become a millionaire and step 1 is "get a million dollars." We could indeed expedite our development process by hiring some very good programmers, but the difficulty of doing that is precisely why our development process is generally slower than we'd like.

A web-based version might be a bit more outsourcing-friendly than a native app, though.

 

[Shun]

@mikelove - Thanks for your honesty, I see that can be tricky. The risks of doing that all-out would probably overshadow the potential benefits. Perhaps a partly outsourced web-based version would make a good test balloon for outsourcing in general. I guess flexibility is essential here, i.e. allowing you to pull out at any time if it doesn't work.

 

[mikelove]

It might, but I kind of feel like after this series of flashcard updates we'll be approaching a point with Pleco where we can focus a lot of our future growth on content- rather than coding-driven features - creates far more user benefit for the money spent, frankly - so I'm not necessarily sure if we're going to need to ramp up + continue with a larger programing workforce long term.

 

[mikelove]

Official recognition of ARC Welder might still happen, BTW - the response to this has been rather impressive and we're currently reaching out to the licensors with no-desktop-version clauses (thankfully not too many) to see how many of them will consider this a problem.

It should theoretically be possible for us to detect + block some dictionaries from use in ARC Welder, and while that might be slightly irritating for people who are using those dictionaries happily in ARC Welder now I would tend to think that the benefits from official recognition (bug fixes, official packages, maybe x86-accelerated native libraries, etc) would offset that. And if we can tell the problematic licensors that we've seen an average increase of X % in sales from ARC Welder users (if such an increase materializes) it might persuade them to drop their restrictions too.

A *fully* optimized ARC Welder version - one that plays nicely with mice, desktop filesystems, etc - would be a whole different order of project and for that one my iOS-to-Mac argument comes back into play, but y'all seem sufficiently delighted with this rudimentary tablet-oriented interface that it feels like a waste not to make at least some effort to officially support it.

 

[Shun]

It might, but I kind of feel like after this series of flashcard updates we'll be approaching a point with Pleco where we can focus a lot of our future growth on content- rather than coding-driven features - creates far more user benefit for the money spent, frankly - so I'm not necessarily sure if we're going to need to ramp up + continue with a larger programing workforce long term.

Yes, very true; Pleco isn't like an operating system or a huge application like Photoshop.

 

[HuShifang]

A *fully* optimized ARC Welder version - one that plays nicely with mice, desktop filesystems, etc - would be a whole different order of project and for that one my iOS-to-Mac argument comes back into play, but y'all seem sufficiently delighted with this rudimentary tablet-oriented interface that it feels like a waste not to make at least some effort to officially support it.

While I certainly wouldn't want to discourage development of an official ARC Welder version, I have to say, I'm finding that it plays extremely nicely with my mouse and filesystem as it is. While the lack of a "back" button in ARC Welder is a very minor hitch, I'm finding the text-selection and mouse-handwriting functionality almost totally natural. And, when I used the ARC Welder interface to restore backed-up app settings, I was automatically kicked me over to the desktop filesystem -- I was able to seamlessly import in the XML file I'd saved to Google Drive from my tablet. (And I've had zero issues with clipboard hand-offs.)

 

[mikelove]

Cool, glad to hear it. Biggest benefit of official support would be performance improvements - we rely heavily on native ARM code now and if we added an x86 binary it would make some parts of Pleco an order of magnitude faster.

 

[alex_hk90]

Cool, glad to hear it. Biggest benefit of official support would be performance improvements - we rely heavily on native ARM code now and if we added an x86 binary it would make some parts of Pleco an order of magnitude faster.

Out of interest, how is it that Pleco works on Google ARC Welder (running on an x86 system) if it uses native ARM code?

 

[mikelove]

There's an emulator - based on VirtualBox, I believe.

 

[dcarpent]

I just came across this thread about running Pleco using ARC Welder. I now have it running on my PC, which is great. My question is what to do about add-ons. I have Pleco licensed for my iPhone/iPad, so I assume that add-ons for the Android version would be new purchases. If this is correct, I'm wondering if it is possible and safe (from a financial investment point of view) to purchase add-ons from within the Android version running on my desktop, via the Pleco web site. Will this work? And if so, would this be wise? For instance, if you decide that running Pleco in this way violates some agreements, could I lose the use of add ons that I buy?