We're very happy and honored to had Adam Wilson on board as a GSoC 2012 mentor. Adam brings solid project management experience and has a specific interest in the Mono-D project.

Please join me in welcoming Adam to the ranks of GSoC mentors!


Thanks,

Andrei

On Monday, 26 March 2012 at 15:27:29 UTC, Andrei Alexandrescu
wrote:
> We're very happy and honored to had Adam Wilson on board as a GSoC 2012 mentor. Adam brings solid project management experience and has a specific interest in the Mono-D project.
>
> Please join me in welcoming Adam to the ranks of GSoC mentors!
>
>
> Thanks,
>
> Andrei

On 03/26/2012 08:27 AM, Andrei Alexandrescu wrote:
> We're very happy and honored to had Adam Wilson on board as a GSoC 2012 mentor. Adam brings solid project management experience and has a specific interest in the Mono-D project.
> 
> Please join me in welcoming Adam to the ranks of GSoC mentors!
> 
> 
> Thanks,
> 
> Andrei

Welcome Adam and congratulation Alex.

I am using Mono-D and I almost enjoy it.  One thing is for sure:
Code LookUp /
"Intellisense" is great in Mono-D, ,Code outline simply rox, and
MNono-D is (in this regard) light years ahead of Visual D.

The pure speed  in which Alex's code analyzer is working is just
xtreme amazing. Alex ? Benchmarks ?

But it is a GTK# and  MONO based project and this means it is
finally a C# project.

I am pretty sure that we will have a complete wxWidgets 2.9.3
binding in a few days/weeks.  (and we will have a TOOL to create
almost automatic wxWidgets 2.4. 2.5, 3.0 bindings)  incl. say Gtk
3.0 and iOS support)

So. wouldn't make more sense to ask Alex to port and enhance  his
code analyzer into D2  as GSOC project  to become part of a wxD2
driven IDE ?
I think, Yep.
Despite that,
Alex, thanks for Mono-D, very well done.
My 2 cents, Bjoern

On Mon, 26 Mar 2012 11:15:54 -0700, BLS <bizprac@orange.fr> wrote:

> On Monday, 26 March 2012 at 15:27:29 UTC, Andrei Alexandrescu
> wrote:
>> We're very happy and honored to had Adam Wilson on board as a GSoC 2012 mentor. Adam brings solid project management experience and has a specific interest in the Mono-D project.
>>
>> Please join me in welcoming Adam to the ranks of GSoC mentors!
>>
>>
>> Thanks,
>>
>> Andrei
>
> On 03/26/2012 08:27 AM, Andrei Alexandrescu wrote:
>> We're very happy and honored to had Adam Wilson on board as a GSoC 2012 mentor. Adam brings solid project management experience and has a specific interest in the Mono-D project.
>>  Please join me in welcoming Adam to the ranks of GSoC mentors!
>>   Thanks,
>>  Andrei
>
> Welcome Adam and congratulation Alex.
>
> I am using Mono-D and I almost enjoy it.  One thing is for sure:
> Code LookUp /
> "Intellisense" is great in Mono-D, ,Code outline simply rox, and
> MNono-D is (in this regard) light years ahead of Visual D.
>
> The pure speed  in which Alex's code analyzer is working is just
> xtreme amazing. Alex ? Benchmarks ?
>
> But it is a GTK# and  MONO based project and this means it is
> finally a C# project.
>
> I am pretty sure that we will have a complete wxWidgets 2.9.3
> binding in a few days/weeks.  (and we will have a TOOL to create
> almost automatic wxWidgets 2.4. 2.5, 3.0 bindings)  incl. say Gtk
> 3.0 and iOS support)
>
> So. wouldn't make more sense to ask Alex to port and enhance  his
> code analyzer into D2  as GSOC project  to become part of a wxD2
> driven IDE ?
> I think, Yep.
> Despite that,
> Alex, thanks for Mono-D, very well done.
> My 2 cents, Bjoern

I think that the best thing that we can do right now is to focus on bringing the parser to completion. It's still missing some key features of D, especially in terms of code-completion and syntax highlighting. It's also missing UFCS from 2.058, which is a pretty big deal I think. For a full list of tasks that Alex would like to get done please see this list: https://github.com/aBothe/Mono-D/blob/master/MonoDevelop.DBinding/Remaining%20features.txt

As to an IDE written in D, that's a HUGE project and well outside the scope of what can be accomplished in a GSoC project. It takes millions of lines of code to make a *DECENT* IDE. Not to mention that UI design is something that will always polarize the community, some basically want a glorified VIM/EMACS while other will settle for nothing less than a Visual Studio clone, still more people will want a radically different UI from anything previously seen (I personally am intrigued by Code-Bubbles for instance). Plus why bother with that when we can integrate into existing solutions like MonoDevelop or Visual Studio *much* quicker.

I personally think that Mono-D represents the most capable path forward for D IDE's right now, maybe later that might change as D grows, but for the moment we need an complete IDE fast, and integration can deliver that.

-- 
Adam Wilson
IRC: LightBender
Project Coordinator
The Horizon Project
http://www.thehorizonproject.org/

On Monday, 26 March 2012 at 19:43:56 UTC, Adam Wilson wrote:
>
> I think that the best thing that we can do right now is to focus on bringing the parser to completion. It's still missing some key features of D, especially in terms of code-completion and syntax highlighting. It's also missing UFCS from 2.058, which is a pretty big deal I think. For a full list of tasks that Alex would like to get done please see this list: https://github.com/aBothe/Mono-D/blob/master/MonoDevelop.DBinding/Remaining%20features.txt
>
> As to an IDE written in D, that's a HUGE project and well outside the scope of what can be accomplished in a GSoC project. It takes millions of lines of code to make a *DECENT* IDE. Not to mention that UI design is something that will always polarize the community, some basically want a glorified VIM/EMACS while other will settle for nothing less than a Visual Studio clone, still more people will want a radically different UI from anything previously seen (I personally am intrigued by Code-Bubbles for instance). Plus why bother with that when we can integrate into existing solutions like MonoDevelop or Visual Studio *much* quicker.
>
> I personally think that Mono-D represents the most capable path forward for D IDE's right now, maybe later that might change as D grows, but for the moment we need an complete IDE fast, and integration can deliver that.

And one of the very nice things about Mono-D is that the parser is completely standalone. It would not be difficult to integrate into Visual Studio in the future. Both are done in C#, and both are somewhat similar to code for. Instead of making a D specific IDE, we can just use a very nice plugin for both Visual Studio and Mono-D, with being able to use the same code-base for the logic.

On 03/26/2012 01:11 PM, Kapps wrote:
> And one of the very nice things about Mono-D is that the parser is
> completely standalone. It would not be difficult to integrate into
> Visual Studio in the future

Well, I am almost on Windows.(Not valid for all of us)
(AFAIK) almost everyting to integrate D into Visual Studio is done in D. (incl. IDL stuff) correct if I am wrong. So yes... Alex's code Analyser should fit. as NET assembly But as well as D shared linrary

Writing a State of the Art D2 IDE will not necessarily require  a million lines of code .

I am convinced that developing in wxD2*** code  will be very close to what you do in wxPython, maybe even smarter.

But I am loosing the point. Even if Alex carries on in Mono-D during GSOC it is a good thing. And.. if we are not able to translate C# stuff into D2 than the D2 design fails..

On Mon, 26 Mar 2012 14:52:19 -0700, bls <bizprac@orange.fr> wrote:

> On 03/26/2012 01:11 PM, Kapps wrote:
>> And one of the very nice things about Mono-D is that the parser is
>> completely standalone. It would not be difficult to integrate into
>> Visual Studio in the future
>
> Well, I am almost on Windows.(Not valid for all of us)
> (AFAIK) almost everyting to integrate D into Visual Studio is done in D. (incl. IDL stuff) correct if I am wrong. So yes... Alex's code Analyser should fit. as NET assembly But as well as D shared linrary

Yes, and IMHO that is really holding it back because not everything that VS has to offer is available via COM. For example anything that wants to touch VS's WPF interface directly needs to go through .NET. In the case of integrations, building the integration in anything other than the language used to build the IDE itself is intentionally tying one hand behind your back in the name of 'purity'. I support Alex's choice to use C# to build the Mono-D binding, its the most sensible decision that can be made.

> Writing a State of the Art D2 IDE will not necessarily require  a million lines of code .

Mono is over a million, Visual Studio is almost as much as the Windows Kernel (5m+ IIRC), and Eclipse ... well I don't what they are doing wrong over there but the bloat is epic.
In other words, a good IDE is a massively complicated beast. Integrations are much quicker and we don't have to reinvent the wheel all over the place.

> I am convinced that developing in wxD2*** code  will be very close to what you do in wxPython, maybe even smarter.
>
> But I am loosing the point. Even if Alex carries on in Mono-D during GSOC it is a good thing. And.. if we are not able to translate C# stuff into D2 than the D2 design fails..
>

Actually, I'm porting the ANTLR Runtime from C# to D right now. The languages are VERY similar, where the whole thing falls apart is the standard library, or the fact that Phobos is brutally underpowered compared to the .NET BCL. I wrote a List(T) class just to make the pain stop.

-- 
Adam Wilson
IRC: LightBender
Project Coordinator
The Horizon Project
http://www.thehorizonproject.org/

Hi there,

Yeah I'm very grateful that Adam wants to be a mentor for GSoC this year -
Nevertheless I'm still not sure which feature(s) I want to focus - since there are so many features that sound interesting but are obviously complex and time-intensive (like showing all possible methods e.g. after a string literal, so it'd be then that you're typing "asdf". and all available/matching methods will pop up)

-- things like CTFE and pre-compile time mixin resolution also are interesting and surely features which hard to fit in a relatively strict time table.

So my actual problem/goal is to fill those 3 months efficently. - My application to GSoC and other formal things are going to follow 'later on' - so I guess in a couple of days.


Oh btw, there's a new Mono-D version :D

Hi,
to make it absolutely sure !! I hope that Alex's project will make it.

(and as one who has worked on a concrete project with Alex, having several private phone conversations, I am sure that Alex will deliver pretty cool stuff. Most probably more than one might expect.)


On 03/26/2012 03:00 PM, Adam Wilson wrote:
> ono is over a million, Visual Studio is almost as much as the Windows
> Kernel (5m+ IIRC), and Eclipse ... well I don't what they are doing
> wrong over there but the bloat is epic.
> In other words, a good IDE is a massively complicated beast.
> Integrations are much quicker and we don't have to reinvent the wheel
> all over the place.
>

IMO this is questionable. What do you count as required LOC ?
Say this is what could be done by Plug-Ins.

SVN / GIT  support,
Database Explorer,
ER Designer
UML Designer
XML/XSL support
SOAP/REST support
etc.

So the core IDE has to support a flexible Doc/View Model a Plug-In Architecture, and Source code analysis.
Maybe an internal project-management that supports a build/make tool.
Debug Support.
Period.
All that visual stuff, say panel docking, gui persistence has not to be written from the scratch.. It is part of the GUI lib.

Exotic stuff,  You want the best ever Ultimate Development Environment.
Say you want Realtime developer collaboration/Video conferencing ... a piece of cake in Python (using async IO/ XMPP ....) No rocket science at all.


>> I am convinced that developing in wxD2*** code  will be very close to
>> what you do in wxPython, maybe even smarter.
>>
>> But I am loosing the point. Even if Alex carries on in Mono-D during
>> GSOC it is a good thing. And.. if we are not able to translate C#
>> stuff into D2 than the D2 design fails..
>>
>
> Actually, I'm porting the ANTLR Runtime from C# to D right now. The
> languages are VERY similar, where the whole thing falls apart is the
> standard library, or the fact that Phobos is brutally underpowered
> compared to the .NET BCL. I wrote a List(T) class just to make the pain
> stop.

Well, here I definitely should shut up.. std.collections... Anyway from time to time I think it would make sense to port the MOMO/NET collection stuff into D. Simply to make porting of dot net code possible without too much pain.  but that's an other story.

Thanks for being a Mentor for this Project.

On 2012-03-27 00:00, Adam Wilson wrote:

> Mono is over a million, Visual Studio is almost as much as the Windows
> Kernel (5m+ IIRC), and Eclipse ... well I don't what they are doing
> wrong over there but the bloat is epic.
> In other words, a good IDE is a massively complicated beast.
> Integrations are much quicker and we don't have to reinvent the wheel
> all over the place.

I agree that an writing an IDE will be a massive project. But it doesn't have to be as complicated as Eclipse or MonoDevelop. These support plugins to add support for new languages, Eclipse contains support for UML diagrams and similar things. Not something that is needed to make a fully usable D IDE.

-- 
/Jacob Carlborg

On Tue, 27 Mar 2012 09:38:33 -0700, Jacob Carlborg <doob@me.com> wrote:

> On 2012-03-27 00:00, Adam Wilson wrote:
>
>> Mono is over a million, Visual Studio is almost as much as the Windows
>> Kernel (5m+ IIRC), and Eclipse ... well I don't what they are doing
>> wrong over there but the bloat is epic.
>> In other words, a good IDE is a massively complicated beast.
>> Integrations are much quicker and we don't have to reinvent the wheel
>> all over the place.
>
> I agree that an writing an IDE will be a massive project. But it doesn't have to be as complicated as Eclipse or MonoDevelop. These support plugins to add support for new languages, Eclipse contains support for UML diagrams and similar things. Not something that is needed to make a fully usable D IDE.

To be a fully useable *D* IDE this is true, but that's not really an Integrated Development Environment, its just Yet Another Specialized Development Environment. I'd argue that the whole point of the "Integrated" part of IDE is that everything you might possibly need to do your job is one place specifically so you don't have to go hunt down that other software package you only need every couple of months. And I'm not saying that we shouldn't have an IDE written in D, just that it's not the best path at the moment, and regardless of the purity folks "everything must be written in D!" tirades, integrating D into MonoDevelop represents the best way to get devs using D right now. Also, the D GUI situation leaves a lot to be desired in terms of complex UI's like IDE's.

Besides, Mono-D has more pressing issues than a potential stand-alone IDE ... CTFE/mixin parsing anybody?

-- 
Adam Wilson
IRC: LightBender
Project Coordinator
The Horizon Project
http://www.thehorizonproject.org/