"Do not feed the trolls"

Dear Stef

Am 22.02.12 08:32, schrieb Stéphane Ducasse:
> I'm convinced that having support for multitouch event/ genie and others works (for iPad = $$$$) is important.

Without pretending to know the future, IMHO standalone apps for tablets
and mobile will disappear over time. Out of my perspective it would be
far more important to move in the direction of web-based technologies
also in this area. Looking at the success stories of Pharo and our own
strategy I do not see the advantages of having multi-touch support for
the development of such applications.

Or did I misinterpret your intention?

Cheers,

Chris

Dear Stef,

Am 21.02.12 21:26, schrieb Stéphane Ducasse:
> Thanks a lot Christophe!
> I love your trust :)
> It puts such responsibility on us but we accept it and we will do our best to make sure you can deliver
> robust and cool applications.

Trust is the best motivation... :)

It was not my intention to put pressure to you or the community. I
believe that Pharo improved our situation a lot in the last years and
that the project is on the right track. For us the lively and very
intelligent community that built and maintains a stable Open Source
Smalltalk IDE for the development of (web-based) commercial application
is the most important thing. If you listen to the wishes of commercial
users and we appreciate and support your work then Pharo will become
even more successful, attractive and fun.

Cheers,

Chris

Hi Christoph!

Jobs dropped FLASH for several important reasons:

1. Too much polling in there. Processor load was high, eating up batteries.
2. Flash games with "onMouseOver" can't run on tablets.
3. Event model was so oldfashioned, that Adobe decided to drop the VM.

FLASH was the most successful virtual machine ever, hundreds of thousands of applications, billions of users, dominating the whole market.

Dead within only 2 years! Severe marketing and technical design mistakes by the Adobe management, IMHO.

Pharo suffers similar problems: GUI is not tablet-ready. Compare to Android 4.0: Android has joined the 2.3 line for mobiles with tablet line and has invested much much brainpower in finding out, how apps can be designed, that they can comfortably be used on different resolutions, portrait as well as landscape. Well done IMHO, even suited for desktop apps.

Pharo also has too much polling code, is eating up batteries as well.

As long as Apple does not allow fullsized interpreters in Appstore, there is definitely no chance ever for Pharo to be brought onto tablet.

I see no chances for Pharo on ARM, Tablets, whatever ... never ever!

This increasing market with hundreds of millions hardware units is completely lost for Smalltalkers, once again.

Tablet apps will very soon even dominate the desktop market!

regards, Guido Stepken

Pharo is an IDE. IDEs don't run on tablets.



--
Milan Mimica
http://sparklet.sf.net

S, Christoph Wysseier piše:

> Stéphane Ducasse:

>> I'm convinced that having support for multitouch event/ genie and
>> others works (for iPad = $$$$) is important.

> Without pretending to know the future, IMHO standalone apps for tablets
> and mobile will disappear over time. Out of my perspective it would be
> far more important to move in the direction of web-based technologies
> also in this area. Looking at the success stories of Pharo and our own
> strategy I do not see the advantages of having multi-touch support for
> the development of such applications.

I agree completely that web based techologies are where we should go
because we have advantage here with Smalltalk while for any kind of GUI
we are off. Including for mobile apps. Making a web app and package it
as a standalone mobile app is much easier than making any decently
looking GUI app with Pharo. Adding only multitouch is not enough, entire
look&feel is what counts.

And I don't see a change to better in near future on GUI field.

It is better to concentrate improving Look&Feel (User Experience) of
development tools and overal stability of Pharo including fundamental
redesign of internals, those which cause instabilities. To prepare Pharo
for server based deployments and for web services of different kinds,
from web apps to REST API services. And a work the Gemstone/VMware guys
are doing adding Smalltalk to Cloud Foudry [1] is what we should support
as much as possible, this is the future.

[1] Adding Smalltalk to Cloud Foundry
http://programminggems.wordpress.com/2012/02/17/adding-smalltalk-to-cloud-foundry/

Best regards
Janko


--
Janko Mivšek
Aida/Web
Smalltalk Web Application Server
http://www.aidaweb.si

Hi Guido

Am 22.02.12 10:06, schrieb Guido Stepken:
> Tablet apps will very soon even dominate the desktop market!

I agree!

 > This increasing market with hundreds of millions hardware units is
 > completely lost for Smalltalkers, once again.

I disagree! ;)
It might be lost for standalone applications, right. But this is true
for a lot of other technologies too, like the mentioned Adobe Flash,
which also shows the difficulties of such a strategy. But, IMHO, the
door to this market for Smalltalkers is wide open as long as you insist
to use web-technologies (HTML5, AJAX, SVG et. al.) where Pharo and
frameworks like Seaside or Aida/Web offer you a wide range of
possibilities to develop platform-independent, innovative applications.
Also Adobe justified their decision to drop Flash with the evalution of
the future of these technologies. I am convinced that the future of such
($$$) business or community applications will be the browser.

Just my 5 cents and may of course be completely wrong.

Cheers,

Chris

Le 22/02/2012 10:34, Janko Mivšek a écrit :
> I agree completely that web based techologies are where we should go
> because we have advantage here with Smalltalk while for any kind of GUI
> we are off. Including for mobile apps. Making a web app and package it
> as a standalone mobile app is much easier than making any decently
> looking GUI app with Pharo. Adding only multitouch is not enough, entire
> look&feel is what counts.


Sometime I wonder if writting application like DrGeo could be
reasonnably doeable with html5? Ideally I would like to write Smalltalk
code transated to the host language plateform.
Whatever, html5 are really desktop application downloaded at runtime. We
are still moving in circle since 40 years.


--
Dr. Geo -- http://www.drgeo.eu

Am 2012-02-22 um 10:30 schrieb Milan Mimica:

> Pharo is an IDE. IDEs don't run on tablets.

Then see the lively kernel…

Am 22.02.12 10:34, schrieb Janko Mivšek:
> It is better to concentrate improving Look&Feel (User Experience) of
> development tools and overal stability of Pharo including fundamental
> redesign of internals, those which cause instabilities. To prepare Pharo
> for server based deployments and for web services of different kinds,
> from web apps to REST API services. And a work the Gemstone/VMware guys
> are doing adding Smalltalk to Cloud Foudry [1] is what we should support
> as much as possible, this is the future.

Just 100% my opinion too!

But, to not shock the devs on this list, Pharo may offer much more than
just an IDE with a headless operation on the server. Innovation on the
server part is as important as using modern web-technologies for the
user interface.

Cheers,

Chris

On 2/22/12, Hilaire Fernandes <[hidden email]> wrote:
It would be worth giving it a try with a simple exercise...

with Amber Smalltalk
http://amber-lang.net/
and Morphic.js
http://astares.blogspot.com/2011/03/morphic-in-javascript.html

What is implemented so far
http://chirp.scratchr.org/dl/experimental/JsMorphic/morphic.txt
No idea if that offers enough for Dr. Geo....

> Whatever, html5 are really desktop application downloaded at runtime.
Yes there is now the option of 'one web page applications'.

This is done when going for an Amber solution actually.

There is some discussion on Morphic  on the Amber Smalltalk Google group;
in short

- you may add any JavaScript library to Amber.
- you may call any JavaScript code just by entering it between <  >
(JavaScript code as 'primitive calls')
- there are simple mapping rules so that you can call an existing
JavaScript API with Smalltalk expressions.


The new HTML5 <canvas> tag seems to offer a lot of possibilities.
http://www.canvasdemos.com/

No idea however how it compares in terms of performance with a direct
solution in Pharo.

> are still moving in circle since 40 years.
In which sense do you mean this?

The web browser is actually an universal client these days.

Please note that Amber is still quite early in development --- version 0.9.1
And you have to shift gears mentally to get used to the file based
approach (github). But actually it still feels like Smalltalk.

For doing some experiments and proof of concept Amber is good enough.
In case something does not work in Amber you can just do it in JavaScript.

Smalltalk is no longer a closed world but running on top of a
JavaScript engine where many external libraries are available.

I think this is a very interesting development.

--Hannes

>
> --
> Dr. Geo -- http://www.drgeo.eu
>
>
>

Am 22.02.2012 10:31 schrieb "Milan Mimica" <[hidden email]>:
>
> Pharo is an IDE. IDEs don't run on tablets.

You're right so far. Principally IDEs best run on desktop computers.
I tried several new concepts to develop on tablet (Bluetooth keyboard, mouse) during travels:

Jens Moenig's ELEMENTS:
http://www.chirp.scratchr.org/blog/?p=24

Fun seeing Smalltalk code in colored blocks, pushing them around like in EToys, generating/refactoring code.

Much better designed, because of LISP + OpenGL is:

http://blocky.io

See the videos. Blocky makes no difference between Morphs and Code. This is pure fun, developing programs, games on tablet, and its incredibly fast!

Have fun!

Guido Stepken

>
> On 22 February 2012 10:06, Guido Stepken <[hidden email]> wrote:
>>
>> Hi Christoph!
>>
>> Jobs dropped FLASH for several important reasons:
>>
>> 1. Too much polling in there. Processor load was high, eating up batteries.
>> 2. Flash games with "onMouseOver" can't run on tablets.
>> 3. Event model was so oldfashioned, that Adobe decided to drop the VM.
>>
>> FLASH was the most successful virtual machine ever, hundreds of thousands of applications, billions of users, dominating the whole market.
>>
>> Dead within only 2 years! Severe marketing and technical design mistakes by the Adobe management, IMHO.
>>
>> Pharo suffers similar problems: GUI is not tablet-ready. Compare to Android 4.0: Android has joined the 2.3 line for mobiles with tablet line and has invested much much brainpower in finding out, how apps can be designed, that they can comfortably be used on different resolutions, portrait as well as landscape. Well done IMHO, even suited for desktop apps.
>>
>> Pharo also has too much polling code, is eating up batteries as well.
>>
>> As long as Apple does not allow fullsized interpreters in Appstore, there is definitely no chance ever for Pharo to be brought onto tablet.
>>
>> I see no chances for Pharo on ARM, Tablets, whatever ... never ever!
>>
>> This increasing market with hundreds of millions hardware units is completely lost for Smalltalkers, once again.
>>
>> Tablet apps will very soon even dominate the desktop market!
>>
>> regards, Guido Stepken
>>
>> Am 22.02.2012 09:39 schrieb "Christoph Wysseier" <[hidden email]>:
>>
>>> Dear Stef
>>>
>>> Am 22.02.12 08:32, schrieb Stéphane Ducasse:
>>>>
>>>> I'm convinced that having support for multitouch event/ genie and others works (for iPad = $$$$) is important.
>>>
>>>
>>> Without pretending to know the future, IMHO standalone apps for tablets and mobile will disappear over time. Out of my perspective it would be far more important to move in the direction of web-based technologies also in this area. Looking at the success stories of Pharo and our own strategy I do not see the advantages of having multi-touch support for the development of such applications.
>>>
>>> Or did I misinterpret your intention?
>>>
>>> Cheers,
>>>
>>> Chris
>>>
>
>
>
> --
> Milan Mimica
> http://sparklet.sf.net

Am 22.02.2012 10:36 schrieb "Christoph Wysseier" <[hidden email]>:

...

Yes, agreed so far. There is another new technology upcoming: WebGL.
On my Nexus the Google Chrome browser *is* a WebGL app itself. Thereunder the V8 Javascript interpreter, which internally creates metaclasses e.t.c. The two core developers were Smalltalkers! :-)
What can we expect from Google Chrome? Separation of Browser and JavaScript interpreter, the interpreter then resides on the server, sending OpenGL directly via websockets into the dumb OpenGL terminal.
And Google Chrome interpreter also has some (new) programming languages under the hood, Smalltalk included. C++ interpreter already works.

Have fun, Guido Stepken

Attatched screenshots from Samsung Nexus.

Dear Hilaire

Am 22.02.12 10:38, schrieb Hilaire Fernandes:
> Sometime I wonder if writting application like DrGeo could be
> reasonnably doeable with html5?

Without knowing your feature set in detail, but have a look for instance
at the following SVG editor:
-> http://code.google.com/p/svg-edit/
Although this does not nearly cover the features of your software, it
shows amazingly what can be done with the HTML5 and SVG.

> Whatever, html5 are really desktop application downloaded at runtime. We
> are still moving in circle since 40 years.

In some way you are right, some sort of things are being reinvented for
new underlying technologies over and over again. Nevertheless, IMHO the
development also exposed several sustainable advantages. As Hannes
mentioned, the browser became a universal client for any applications,
reducing the need to adapt applications for different operating system
or other basic technologies. As far as I understand for instance Googles
vision, it should even become the runtime environment of the user
interface of the operating system.

Besides, and this is especially true for mobile devices, using a
client-server architecture the computation power and data storage shifts
more and more to the server/cloud instead of the desktop.  This is a
large advantage for developers and system engineers for maintenance and
operation especially when deploying a new version of a software as there
is no software installed on the client directly.

Cheers,

Chris

Am 22.02.2012 11:45 schrieb "Christoph Wysseier" <[hidden email]>:
>
> Dear Hilaire
>
> Am 22.02.12 10:38, schrieb Hilaire Fernandes:
>
>> Sometime I wonder if writting application like DrGeo could be
>> reasonnably doeable with html5?
>
>
> Without knowing your feature set in detail, but have a look for instance at the following SVG editor:
> -> http://code.google.com/p/svg-edit/
> Although this does not nearly cover the features of your software, it shows amazingly what can be done with the HTML5 and SVG.

Works fine on Nexus with Chrome Beta. :-)

Perfect Cloud app and faaaast!!!! :-) OpenGL 2D canvas hardware support!

Guido Stepken

Or maybe web application disappear over time.
Most device providers creates own app stores which includes native application for devices (not web applications). And specialised devices market always grow.
All customers want see there client frontend applications at each market.

I think standalone computers will become less popular each year. And so classic web applications too.

Hi,

Germán,

On 22 Feb 2012, at 13:28, Germán Arduino wrote:

> Example: xmlrpc died before to born, worked only in 1.1.1, now I must find time to update it to zinc and 1.3 but with the fear that again will not work in 1.4.

I think I said it before (but I can't remember it must be a long time ago): I am more than willing to help you port it. I am not a fan of XMLRPC, but it is an existing standard that has its place. Furthermore, it was designed to be very simple, so it would be a shame to let it rot away.

Sven