[SailfishDevel] Mono for Sailfish - round #2?

Bob Summerwill bob at summerwill.net
Thu Jun 18 07:14:08 UTC 2015


PS. More info on the MonoDevelop add-in for Tizen, which would be the model
for the MonoDevelop add-in for Sailfish:

    http://kitsilanosoftware.github.io/MonoDevelop.Tizen/

Aside - The Tizen project is even more broken now than it was last year.
I now have a deep understanding of the reasons why Carsten and David
re-launched the Mer project.   Thank god for you guys doing that, because
Tizen has played out exactly as you expected, which is presumably exactly
the same as MeeGo did, just with Samsung taking the place of Nokia.
 There is absolutely no consideration of the community within the Tizen
Project, which Intel seem to have silently abandoned in the last few
months.    It is now a pure Samsung show and seems to be pretty
indistinguishable from Android in the last it is developed.   Everything
behind closed doors.   Source code drops at major releases.   No
opportunity for community involvement, and little to no information on what
is being worked on until after it is released in a product.

Samsung just recently did a preview SDK for the pending Gear watch (round
one) which is a move in the right direction, but this is very much on their
terms (and again a drop, rather than development in the open):


http://www.samsungmobilepress.com/2015/04/24/Samsung-to-Release-SDK-for-Next-Generation-Gear

The open governance for Tizen which was announced an incredibly long time
ago only applied to Tizen 3.0 development, not to the ongoing Tizen 2.x
development on which all of Samsung's products are built.   That is all
still happening behind closed doors.

Tizen 3.0 development was happening in the open, but was only for Tizen IVI
and Tizen Common.   They did get to a
https://wiki.tizen.org/wiki/IVI/Tizen-IVI_3.0 release, but as soon as that
happened Intel appeared to silently bugger off, as though they had a legal
obligation to ship Tizen IVI, and having met that obligation, they
high-tailed it out of Dodge.   Of course, they may not admit that reality
for months, or indeed ever, but that appears to be what has happened.

With that being the case, Tizen 3.0 is essentially a zombie codebase.
Some Samsung engineers might be working on Tizen 3.0 profiles, for Mobile,
or Wearable, or TV, but those don't correspond to the code which is going
into the Tizen-based products which Samsung are actually shipping - Tizen
SmartTV (all 2015 models are Tizen), Cameras, Gear smartwatches (Gear 2,
Gear S and now pending Gear A), and finally mobile - Samsung Z1, with Z2
and Z3 pending rumored.   I see little movement towards Samsung making
product based on Tizen 3.0.

With Intel having gone, I see no reason whatsoever why Samsung will ever
"play nice" and move to open governance and a Tizen 3.0 codebase for
Mobile, Wearable and TV.    They can harvest whatever features they like
from Tizen 3.0 (Wayland, Multi-User, SMACK, etc) back into their internal
Tizen 2.x codelines, and need never be subject to open governance, and
claim they never violated any agreement, because the open governance was
only for Tizen 3.0 development, and they aren't doing any Tizen 3.0
development.

They made a huge API change between Tizen 2.3 Alpha and Tizen 2.4 Beta
showing complete contempt for semantic versioning.
https://developer.tizen.org/fr/forums/native-application-development/huge-difference-native-environment-between-2.2.1-and-2.3?langswitch=fr.
   They dropped the C++ OSP framework which had been brought across from
Bada between "Alpha" and "Beta", breaking all existing Native applications.
  Everybody who had been working on native applications for the last 18
months or so was left in the dust.   The API which everyone had been told
to use, and which books had been written about, was just silently killed
with no forewarning or even explicit announcement - until developers
noticed the change and responded with WTF!

Given that relationship between Samsung and the community, I can see them
hiding inside Tizen 2.x forever.    The sad thing is that Tizen is maturing
quite nicely, and is covering a larger and larger range of profiles and
actual devices.   Samsung are taking Tizen very seriously.   It just is NOT
AT ALL development in the open or community-friendly.   It is just like
Android in that respect, just with a better, more GNU/Linux aligned
code-base.


Cheers,
Bob

On Thu, Jun 18, 2015 at 2:42 AM, Bob Summerwill <bob at summerwill.net> wrote:

> Greetings!
>
> Last year the Mono for Sailfish project was announced, development started
> and then withered and silently died.   That was mainly due to reasons
> related to my own personal situation (I lost a job and had to focus on
> job-hunting, not Kitsilano Software, etc) rather than any lack of technical
> merit of the project.
>
>    http://monoforsailfish.com
>
> http://www.mobilelinuxnews.com/2014/08/introduction-mono-sailfish-os-jolla/
>
> Anyway.  It is a new year, and circumstances have changed.   After several
> months in the doldrums, the winds have changed in our favor again, sailors!
>
> 1. Microsoft have open sourced .NET in a major way, and are supporting it
> on Linux and Mac OSX.   They announced that last November and in April of
> this year they made the first preview releases for OSX and Linux.   See
> http://venturebeat.com/2015/04/29/microsoft-releases-net-core-preview-for-mac-and-linux/.
>   The did the most amazing .NET Core demo "trick" during //BUILD, which was
> creating an ASP.NET 5 web app (ASP.NET5 is open-sourced too) in Visual
> Studio on a Windows PC, deploying that app into a Linux Docker container
> (so .NET Core assemblies on Linux with the ASP.NET5 assemblies on top of
> that) and then running that app and hitting a breakpoint and
> single-stepping through the app).    So debugging a .NET app running inside
> a container, running on a different OS.   Kind of cool.     .NET Core is
> going to be an even better base for getting .NET onto mobile Linux than
> Mono was, because it has the full weight of Microsoft support behind it.
> They want that .NET platform available for Linux to support ASP.NET apps
> inside Azure.   Mono on Linux wasn't supporting any business for Xamarin,
> so was a little unloved.   Their focus is on Android and iOS.
>
> Aside - Microsoft also released this -
> http://www.hanselman.com/blog/IntroducingVisualStudioCodeForWindowsMacAndLinux.aspx
> .
>
> 2. QtSharp (https://github.com/ddobrev/QtSharp), the project on whose
> completion Mono for Sailfish was dependent, has got funding as part of the
> Google Summer of Code, so will be brought to functional completeness on
> Windows, OSX and Linux this year.  That is fantastic, because I was
> personally bankrolling that non-Sailfish-specific work as part of Mono for
> Sailfish.   It moved along for a couple of months under Mono for Sailfish,
> but it was apparent that there was a lot of work more work to be done to
> get to that 1.0 version.   But that will now be moving ahead independently
> of Mono for Sailfish, which is great to see.   Dimitar Dobrev is the
> developer.  Hi, Dimitar, and congratulations on securing funding from GSOC!
>
> Deliverables: Improve the QT bindings generator to the point that they
> can be used for a non-trivial QT sample written in idiomatic C#.
>
>
> https://www.google-melange.com/gsoc/project/details/google/gsoc2015/ddobrev/5741031244955648
>
> https://trello.com/c/b34YKGIi/57-cppsharp-continue-mono-net-bindings-for-qt
>
> When the QtSharp GSOC project is over (when is that, Dimitar?) and we have
> a non-trivial Qt sample written in idiomatic C# working on Windows, OSX and
> Linux, I think we are in a position to look at rebooting this project,
> though it would be .NET Core for Sailfish now, not Mono for Sailfish.
>
> This new project would have much of the same flavor as the last one, but
> have a smaller scope of effort required to get to a 1.0 release:
>
> 1. Get .NET Core runtime for Linux working on Sailfish (should be similar
> scope to the work which Damien Diederen did for MonoTizen).   See
> http://monotizen.com.
>
> 2. Build MonoDevelop plugin for Sailfish (should be similar scope to the
> work which Damien Diederen did for MonoTizen).   See http://monotizen.com.
>
> 3. Build wrappers for Sailfish-specific Qt/QML components, so that apps of
> similar complexity to the deliverable of the GSOC project can be built on
> Sailfish.
>
>
> With regard to this third point, is there a Wiki page or other posting
> detailing the latest state of licensing for Silica?   Has that moved at all
> since last year?   Are more QML components being open-sourced?   And just
> to be clear, there is no "source code hiding" going on with Silica, right?
>   It is just that certain files are not under an open source license?
> Nothing that would hinder this binding work, eh?
>
>
> Cheers,
> Bob Summerwill
> Kitsilano Software
> (http://bobsummerwill.wordpress.com/about)
>
>
> --
> bob at summerwill.net
>
>


-- 
bob at summerwill.net
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.sailfishos.org/pipermail/devel/attachments/20150618/c8c2e288/attachment.html>


More information about the Devel mailing list