A friendly retort for Linux build client
Rob,
To restore my +1 and negate yours (& thomas6’s) -1s and , I have to give this a +3.
It is all about what tools folks are comfortable with and the ability to commune with their community (for support) to enable their productivity. Everyone works differently, having different perspectives. Remember Apple’s mantra, “Think Different”.
Personally, I’m an OS guy. I’ve used ~ 2 dozen OSes in 20 years. I’d like to develop for Android first with Corona to test the waters. If things go well, I’ll move onto the requirements for iOS.
First, we are obviously in the same camp about Linux or at least Ubuntu with regard to Windows - though I don’t know if for the same reason (so I won’t assume). And as a long time Mac user I understand why folks want alternatives. (I’ve a retro collection of working: Quadra 650, 7100, G3, PowerComputing PowerBase, Duo 210 with dock - running their MacOS’s and NetBSD) Some folks need, want or like computers that just work - even power users - it could be a Mac or commodity hardware running Linux like me.
And like I mentioned previously, there’s the startup costs associated with trying to develop for iOS or Android - such as affording a $1000+ computer vs parts + Linux OR Windows.
I’ve digressed, allow me to retort on your points 1 & 2.
- I understand Ansca would need to hire at least 1 if not a couple of Linux FTE’s to make a Linux Client. Then again, for the initial port, they have these folks call contractors
But your math regarding selling Ubuntu subscriptions doesn’t make sense to me. For Android, the license count in the end would be the same whether for Windows OR Linux OR the Pro version. Yes, Ansca would probably need to show there’s a viable demographic in Ubuntu (or Linux in general) to merit a Corona port to increase their license count.
I would hope being “close enough” to the target demographic count would be enough to hire some contractors for the port. About this point, the _more_ Ansca can license = the more revenue there is for Ansca – regardless of the OS. Adding a Linux build client would benefit their license count & revenue.
2. You point out that porting would take away important cycles from improving the Corona client platform. I don’t think it would. (This also addresses thomas6’s comments. There’s nothing wrong having self interest. Hell, I’m pushing for a Linux client 
Here’s why:
a. Corona is a client for a _build platform_ - at least regarding Android. All builds occur on their servers, not your local desktop. So the only thing needing to be ported is the local emulator/client. See http://developer.anscamobile.com/content/building-devices-android - to quote:
“In particular, you do not need to install the Android SDK! The entire Android build process can be handled by our servers.”
In fact, until you subscribe, all your remote builds are huge non-optimized compiles.
b. Based on a. that means the linux port wouldn’t necessarily take away from the “Corona/Lua engine improvements” - they don’t have a substantial direct relation. The folks working on the Corona Engine and the Build Servers - the essence of the platform in general - more than likely wouldn’t be working on the Linux GUI anyway.
Ansca has abstracted the client & build process from each other so it’s already OS agnostic. Ansca doesn’t even provide an IDE, though they have recommendations. So there’s not porting of an IDE either. (If you already use Eclipse, it runs on Linux brilliantly.)
c. Linux users have tried getting Corona to run under Wine without much success. Maybe Ansca could work with CodeWeavers to make a custom Wine install for Corona until a full port can be made. Then they wouldn’t need to do a full-on port, however …
d. Many commercial, enterprise companies program their proprietary apps, agents, client and drivers for Linux/Unix first for a stable “foundation” code base, then enhance for other OSes. nVidia does it. So does BMC, HP, IBM, etc. It’s really common practice. Where to you think all those MacPorts come from? … BSD & Linux user space. What about Mac OS X & iOS, their foundation is Darwin/OpenBSD. Safari from WebKit.
e. Numerous commercial companies use QT as the client UI framework. Anything made with it receives native OS feature sets automagically. I don’t know if Ansca uses QT, but it would be a smart choice. Build with QT on Linux, then have OS UI enhancements for Mac OS X & Windows. Remember, Lua, SQLite & Box2D builds on anything. (Check out Moai, Love, etc.) There’s nothing magically about it. (Hell, Love 2D has even been ported to my GP2x Wiz handheld game Emu and the PSP.)
f. Back to point b. Pirating Corona on Linux should be a problem either. You _always_ have to log into the Corona client to use it and submit code for building on their servers (again for Android at least).
All in all Rob, I don’t think you’re Debbie Downer. I just think there’s still more opportunity to evangelize and spread the benefits of Corona. Folks want it. Ansca has to make sure if they create a survey of interest, they promote the survey in the right places … like: XDA forums, Phoronix, AndroidPolice, OMG Ubuntu, PlanetUbuntu, etc. Even include the Fedora folks (the SuSE folks would use their RPM packages.).
Wouldn’t it be amazing if Corona was in the Canonical Ubuntu Store for an instant install. (Other commercial apps with subscriptions appear there already). Then students and more common users would try it out and come up with some amazing Apps. (which adds to the whole Corona eco-system … apps, third party tools & libraries, community membership, etc.). And of course, it would come back to benefiting Ansca & the rest of us.
Cheers. Thanks for listening. M.
[import]uid: 68781 topic_id: 20682 reply_id: 111263[/import]
