Thanks all! Ing. Pablo Digonzelli Software Solutions IP-Solutiones SRL Metrotec SRL 25 de Mayo 521 Email: pdigonzelli@softsargentina.com pdigonzelli@gmail.com Cel: 5493815982714 ----- Mensaje original ----- De: "Hilaire Fernandes" <hilaire.fernandes@gmail.com> Para: pharo-dev@lists.pharo.org Enviados: Lunes, 9 de Junio 2014 13:27:52 Asunto: Re: [Pharo-dev] vm over android I have been using it for DrGeo https://code.google.com/p/squeakvm-tablet/wiki/JenkinsBuilds Hilaire Le 09/06/2014 16:50, Pablo R. Digonzelli a écrit :

Hi all, i want to know if it is planning develop a vm over android? Time frame ?

TIA

**** ------------------------------------------------------------------------ *Ing. Pablo Digonzelli* Software Solutions IP-Solutiones SRL Metrotec SRL 25 de Mayo 521 Email: pdigonzelli@softsargentina.com pdigonzelli@gmail.com Cel: 5493815982714

-- Dr. Geo http://drgeo.eu iStoa - https://launchpad.net/istoa

I have to admit Esteban is absolutely right. DrGeo on Android and iPad were very interesting experiences but it is not up to the expect level in term of user experience on these devices. Hilaire Le 09/06/2014 20:00, Esteban A. Maringolo a écrit :

Just a comment: With the exception of something like Dr. Geo which has a complex custom drawn UI or for games, I wouldn't advice anybody to develop a "traditional" application using Pharo on Android.

My company develops native Android apps (backed with Pharo servers), and unless the app is rudimentary or you need to run on all devices... I wouldn't recommend html5 (js and friends) either. The technology "isn't there" yet (but it's moving forward).

Regards!

Esteban A. Maringolo

2014-06-09 14:25 GMT-03:00 Pablo R. Digonzelli <pdigonzelli@gmail.com>:

...

Thanks all!

Ing. Pablo Digonzelli Software Solutions IP-Solutiones SRL Metrotec SRL 25 de Mayo 521 Email: pdigonzelli@softsargentina.com pdigonzelli@gmail.com Cel: 5493815982714

----- Mensaje original ----- De: "Hilaire Fernandes" <hilaire.fernandes@gmail.com> Para: pharo-dev@lists.pharo.org Enviados: Lunes, 9 de Junio 2014 13:27:52 Asunto: Re: [Pharo-dev] vm over android

I have been using it for DrGeo https://code.google.com/p/squeakvm-tablet/wiki/JenkinsBuilds

Hilaire

Le 09/06/2014 16:50, Pablo R. Digonzelli a écrit :

...

Hi all, i want to know if it is planning develop a vm over android? Time frame ?

TIA

**** ------------------------------------------------------------------------ *Ing. Pablo Digonzelli* Software Solutions IP-Solutiones SRL Metrotec SRL 25 de Mayo 521 Email: pdigonzelli@softsargentina.com pdigonzelli@gmail.com Cel: 5493815982714

-- Dr. Geo http://drgeo.eu iStoa - https://launchpad.net/istoa

-- Dr. Geo http://drgeo.eu iStoa - https://launchpad.net/istoa

2014-06-09 17:08 GMT-03:00 phil@highoctane.be <phil@highoctane.be>:

Le 9 juin 2014 20:40, "Hilaire Fernandes" <hilaire.fernandes@gmail.com> a écrit :

...

I have to admit Esteban is absolutely right. DrGeo on Android and iPad were very interesting experiences but it is not up to the expect level in term of user experience on these devices.

Yes, same feedback here.

For games, Monkey is a good system. http://monkeycoder.co.nz

IMO, Pharo can't compete with tools for the creation of native mobile apps like Xamarin, RubyMotion or Appcelerator Titanium. Mobile is a moving target (pun unintended), and requires tons of money to at least enter into the competition. :) Esteban A. Maringolo

I am really close to finish the android VM, only based on the Unix source (just decorate with less than 20 lines of codes, I need a entry point for the interpreter loop and some convention that cannot change via the android builder). + one graphical plugin and a C file for communicate between C and Java. CogDroid, project is not maintain since 3 or 4 years. But still seems work (just have a completely separated source). Peoples give me all the key to revive it. I take the source of Cogdroid for understand what they do and redo it changing less thing possible in unix source. My VM run without the graphics and with a subset of plugin (file, socket and some of them I cannot remember). Road map: - Full compiling and packaging process: Done. - Fix interpreter headless: Done. - be able to generate and load external lib on the fly from android: done - Graphical plugin: In progress. - reintroduce slowly and fix all the removed plugins: todo. - Nice android connectivity: todo. (all event manage, list of image, blablabla) - Create a CMakeGenerator extension to have a really nice way to generate project. Once a graphical version come out I promise some picture and a website. But again I do that on my free time :-). On 09 Jun 2014, at 18:27, Hilaire Fernandes <hilaire.fernandes@gmail.com> wrote:

I have been using it for DrGeo https://code.google.com/p/squeakvm-tablet/wiki/JenkinsBuilds

Hilaire

Le 09/06/2014 16:50, Pablo R. Digonzelli a écrit :

...

Hi all, i want to know if it is planning develop a vm over android? Time frame ?

TIA

**** ------------------------------------------------------------------------ *Ing. Pablo Digonzelli* Software Solutions IP-Solutiones SRL Metrotec SRL 25 de Mayo 521 Email: pdigonzelli@softsargentina.com pdigonzelli@gmail.com Cel: 5493815982714

-- Dr. Geo http://drgeo.eu iStoa - https://launchpad.net/istoa

Best Regards Dr Arnaud jbaptiste.arnaud@gmail.com

Jean, Can you share what is the intended use of the android vm that you're building? Regards! Esteban A. Maringolo 2014-06-18 9:56 GMT-03:00 Jean Baptiste Arnaud <jbaptiste.arnaud@gmail.com>:

I am really close to finish the android VM, only based on the Unix source (just decorate with less than 20 lines of codes, I need a entry point for the interpreter loop and some convention that cannot change via the android builder). + one graphical plugin and a C file for communicate between C and Java.

CogDroid, project is not maintain since 3 or 4 years. But still seems work (just have a completely separated source). Peoples give me all the key to revive it. I take the source of Cogdroid for understand what they do and redo it changing less thing possible in unix source.

My VM run without the graphics and with a subset of plugin (file, socket and some of them I cannot remember). Road map: - Full compiling and packaging process: Done. - Fix interpreter headless: Done. - be able to generate and load external lib on the fly from android: done - Graphical plugin: In progress. - reintroduce slowly and fix all the removed plugins: todo. - Nice android connectivity: todo. (all event manage, list of image, blablabla) - Create a CMakeGenerator extension to have a really nice way to generate project.

Once a graphical version come out I promise some picture and a website. But again I do that on my free time :-).

Hi Clement, 2014-06-19 12:52 GMT-03:00 Clément Bera <bera.clement@gmail.com>:

2014-06-18 15:39 GMT+02:00 Esteban A. Maringolo <emaringolo@gmail.com>:

...

Can you share what is the intended use of the android vm that you're building?

- Deploying application on Android - Proving to big customers Pharo can run on ARM processor on the contrary to several other smalltalks

What kind of applications? Is there an business interest of running on ARM processors? Is there a sale advantage of having this? I don't want to sound too inquisitive nor pedantic, those are real questions. As stated in previous mails, running on the platform is a technical challenge per se (I couldn't make it if I wanted), but it's just a very small part of "deploying to Android". And I'm not talking about app stores. Even though Android (AOSP) is Linux based, the memory/battery constraints makes that apps or services lifecycle completely different to regular unix processes. And I don't see how this fits into the whole android environment. Last time I tried (+1yr) it the VM was a permanent process, without access to device sensors, etc.[*] I have a genuine interest in this topic, because my company depends both on Pharo and Android (native) software. But how I see this, the advantage of Pharo running on Android for devices other than phones/tablets, more kind of "internet of things" devices. The advantage is also that the image is an asset of the vm, so you can update your app without having to reinstall it through the platform app management. Please don't let this stop you guys from doing this VM even for the fun/sake of doing it. But as you are making this public, I feel allowed to ask questions and add comments. You can simply ignore them. :) Regards, Esteban A. Maringolo [*] Several years ago I took a private course of "mobile squeak", and even compiled a modified squeak VM for WinCE, running MVC based UIs. Back then Squeak UI was way better than WinCE's, even for "business apps". Now I think mobile expectations changed.

I got an iPhone for my father recently and he was blown away by the capabilities of the device I was describing him. Then I was surprised by the fact he was surprised. Then I took a step back and tried to be in his place. Then I was blown away too. Here is this 70 year old man, he is used to mobile phones which are basically phones that call other mobile phones and ..... well... thats it! Compared to an iPhone which is thousands times more powerful than my first computer. He told me to get him the adaptor for TV, so he can see things from his iPhone to his TV (watch football matches and youtube videos). I had already done a search and I found plenty of keyboards and mouses for iPhone. It makes you sit back and ask "Mr Desktop / Laptop .... why we need you ?" For me it makes sense to have a desktop or a laptop because I do 3d graphics and I need the power of a powerful desktop GPU which can run circles around any GPU used for smartphones and tablets. So desktop/laptop is a mandatory. But I am the exception , I am "the freak". If computational power is not an issues, which is not for the vast majority of users out there. The size of the screen is not an issue because you can connect it to any modern screen and TV you want, the lack of keyboard is not because you can connect an external keyboard and lack of mouse is not because you can connect a mouse as well. And not to sound out of topic VM on Android is the single most important thing for the future of Pharo. Mobile market wont become smaller it will become much larger. Because smartphones are really lately that have turned into a full blow computers and people slowly realize it. How long before most of those people realize that they dont need their desktop ? Software migrates/ported constantly. I definetly would love to have Pharo on my Nexus 4 one day and run smoothly. I know its a lot of work, I don't expect it to happen any time soon, but when the time will come my mobile phone will be "Pharo Powered" ;) On Thu, Jun 19, 2014 at 8:52 PM, phil@highoctane.be <phil@highoctane.be> wrote:

A good place to run Pharo for IoT would be on an ARM-based Synology box.

That would be a killer niche.

Phil

On Thu, Jun 19, 2014 at 6:35 PM, Esteban A. Maringolo < emaringolo@gmail.com> wrote:

...

Hi Clement,

2014-06-19 12:52 GMT-03:00 Clément Bera <bera.clement@gmail.com>:

...

2014-06-18 15:39 GMT+02:00 Esteban A. Maringolo <emaringolo@gmail.com>:

...

...

Can you share what is the intended use of the android vm that you're building?

...

- Deploying application on Android - Proving to big customers Pharo can run on ARM processor on the contrary to several other smalltalks

What kind of applications? Is there an business interest of running on ARM processors? Is there a sale advantage of having this? I don't want to sound too inquisitive nor pedantic, those are real questions.

As stated in previous mails, running on the platform is a technical challenge per se (I couldn't make it if I wanted), but it's just a very small part of "deploying to Android". And I'm not talking about app stores.

Even though Android (AOSP) is Linux based, the memory/battery constraints makes that apps or services lifecycle completely different to regular unix processes. And I don't see how this fits into the whole android environment. Last time I tried (+1yr) it the VM was a permanent process, without access to device sensors, etc.[*]

I have a genuine interest in this topic, because my company depends both on Pharo and Android (native) software.

But how I see this, the advantage of Pharo running on Android for devices other than phones/tablets, more kind of "internet of things" devices. The advantage is also that the image is an asset of the vm, so you can update your app without having to reinstall it through the platform app management.

Please don't let this stop you guys from doing this VM even for the fun/sake of doing it. But as you are making this public, I feel allowed to ask questions and add comments. You can simply ignore them. :)

Regards,

Esteban A. Maringolo

[*] Several years ago I took a private course of "mobile squeak", and even compiled a modified squeak VM for WinCE, running MVC based UIs. Back then Squeak UI was way better than WinCE's, even for "business apps". Now I think mobile expectations changed.

"I disagree here. As an example, Android is switching its JIT runtime for an AOT one, because with JIT the performance (mostly related with UI perception) is slower than its iOS counterpart even when running on better hardware. Most of the popular apps use native multithreading intensively to maximize responsiveness. Pharo (as an organization) can't compete with that level of optimization, it doesn't have the budget for it. " Its not about competition, its about "hey guys may have a tiny slice of the cake ? I can ? thanks I will be here in my corner" Also the more powerful the hardware will become the less need would be to have true multithreading. I am not saying its very important but lets face it even on a single core you can do pretty amazing stuff. Right now mobile devices are quite slow even for basic stuff but the hardware only gets faster. "Do you really want to have Pharo (for development) in a 5" screen? Or even in a 10" tablet? Have you tried using it? I did, and played around for a while. It was impressive that it ran bit identically. But that's all. " As I said you can connect your mobile device to any monitor you want, you want 20'' fine , 30'' sure, 40'' why not , 1000'' oh yeah as long it has an hdmi port. The bottom line is that almost everyone will have a smartphone phone . Is this a market as a pharo developer you want to tap in ? It already has happened in USA and most of Europe , even here in Greece with big economical crisis people buy mainly smartphones and some even spend substantial amounts. iPhones actually are pretty popular here too. "I'm not aware of a trend to have IntelliJ/Eclipse and/or XCode running on mobile devices, even when there is less "technology impedance" between their programming language and the supporting platform." This is why I said we are not there yet, but is getting there. For example there is an Emacs port , though it requires external keyboard. Its not that it cannot be done, its more about moving a very big code base takes a lot of time and effort. But once again remember that its about the users not the developers. Developer did not move to the web because they like, actually most people I have talked to dislike and some even hate it. But its a very big market and a market with a very solid future in a world with economic uncertainty. 4 years ago, I remember arguing in music forums that music mobile apps will become serious tools even for professionals, back when ipad was first released, many people laughed at me, other though it was a possibility , some were sceptical . I was almost certain. 4 years later here we are, loads of music apps on mobile devices and their even artists that use solely mobile devices for music production with some help from Dekstops toos for final touches. What will happen in the next 4 years ? I was not successful trying Pharo , I tried Squeak but the multitouch was problematic as hell. Pharo claims "its modern" , but for me no language or software can claim that unless its taps to the mobile market. Unless the software is so demanding that it needs loads of raw performance that a mobile device cannot offer. But even that last front will slowly fall. Its not that Desktops will not be the fastest / most powerful option but mobile devices will become loads more powerful. Imagine what will happen if the smartphone of the future is 1000 times more powerful lets say in 30 years , imagine what you could do with that technology. Should Pharo see 30 years in the future ? If it has learned anything from the mistakes of Smalltalk , it should. About the cloud option, I have to say I am not a huge believer in cloud technology. I think for developed countries its a very good solution BUT many countries out there, including mine, have extremely unreliable internet connections. Browsers have seen this need and loads of options for offline storage has been offered, though it has become quite a confusing mess. So yeah cloud with offline storage is a good recipe , Pharo on the cloud is definetly a good option. Pharo + Amber is something I have recommended many times. So yes I agree with you there, but that does not make me dislike the idea of a VM on android and iOS. Why ? because web technology is a huge pile of mess, its a pain in the ass. Are Amber people going to wrap all this technology for us to make it easier to use ? probably not. So I still like to fully use Pharo. But Amber is a very good option too. So yes I think Pharo on mobile devices, however that happens is not the near future Pharo but its absolutely the distant future.

Hi,

Yes there is a business interest of running on ARM processors. Some of the big companies use smalltalk on their own hardware for very specific tasks, and they are now constraint to use x86 processor due to smalltalk.

There are other smalltalks that run on non-x86 && non-amd64 machines. Smalltalk ~~ Pharo :-) Jan

2014-06-20 5:06 GMT-03:00 Clément Bera <bera.clement@gmail.com>:

2014-06-19 18:35 GMT+02:00 Esteban A. Maringolo <emaringolo@gmail.com>:

...

What kind of applications? Is there an business interest of running on ARM processors? Is there a sale advantage of having this? I don't want to sound too inquisitive nor pedantic, those are real questions.

Yes there is a business interest of running on ARM processors. Some of the big companies use smalltalk on their own hardware for very specific tasks, and they are now constraint to use x86 processor due to smalltalk.

In addition for some big projects management checks how the environment is compatible to other platforms, and they can check "running on ARM, Android, iOS, ..."

I don't know about other PMs. But of the two companies I know that evaluated Pharo to develop "big projects" (10+ current smalltalk developers) one discarded it not because OS support, but rather because of simpler, mundane, library related, things. (However this is a different topic).

...

As stated in previous mails, running on the platform is a technical challenge per se (I couldn't make it if I wanted), but it's just a very small part of "deploying to Android". And I'm not talking about app stores.

I agree. I want to run on ARM, not necessarily and Android.

Fair enough. IMO, and from the PR point of view, saying it "runs on Android" and then finding out it is an alien process in the platform is not a good sales point.

...

But how I see this, the advantage of Pharo running on Android for devices other than phones/tablets, more kind of "internet of things" devices. The advantage is also that the image is an asset of the vm, so you can update your app without having to reinstall it through the platform app management.

There is also this direction. Running on ARM is very important to be cross platform. That includes connected objects, the raspberry pie, and many other devices.

That's totally reasonable, and a platform where something like a self contained platform like Pharo can do "magical" things. Regards!

Can you share what is the intended use of the android vm that you're building?

-_^ ? Run StackVM on an android is not enough from intellectual point of view ? I am not sure to understand the question. Why I do that, "Because I can :-)" I am based on the StackVM unix, my main objective is to change less sources as possible to keep that easy to maintain. I used a fork of PharoVM because the process is simpler for merging and reviewing my changes. And because I using the Pharo-contribution (jenkins) infrastructure to generate VM as I do for RaspberryPi. The first objective is to have something usable on Android, which can be easily update if for whatever reason I decide to stop or if someone want to do it own android App. First step: -Having that running, with graphic based on unix source (with some change but less possible). Next step: - have fun with it and wrap it with a nice android app. - a lot of cool idea that people can do because everything will be public. If it isn't precise enough, be more direct :-). On 18 Jun 2014, at 18:44, Hilaire Fernandes <hilaire.fernandes@gmail.com> wrote:

Hi Jean-Baptiste,

Dmitry wrote some hook to call the Android Virtual keyboard from the image, and may be some other I forget about.

I get the source of CogDroid from him. :-) Most of the java code and the hook that he did is conserved they are extract from the platform android source. And I create a External plugin display for android.

Hilaire

Le 18/06/2014 14:56, Jean Baptiste Arnaud a écrit :

...

CogDroid, project is not maintain since 3 or 4 years. But still seems work (just have a completely separated source). Peoples give me all the key to revive it. I take the source of Cogdroid for understand what they do and redo it changing less thing possible in unix source.

-- Dr. Geo http://drgeo.eu iStoa - https://launchpad.net/istoa

Best Regards Dr Arnaud jbaptiste.arnaud@gmail.com