Hi,

I'm posting this on bhalf of a group of 4 students from the final
year of Department of Computer Science and Engineering,University
of Moratuwa, Sri Lanka. We are required to do a 8 months long
project in our final year. The
project should consist of both a research and development
component and
an implementation. Our group is interested in areas such as,

   * Information security
   * Cloud
   * Parallel computing
   * Compilers
   * Operating Systems
   * Distributed systems


We’d like to know if the D programming language project has any
project
ideas that falls within our interest areas.

For more information about the members of our group, please refer
the
following links.

Ishan Thilina Somasiri -
http://lk.linkedin.com/pub/ishan-somasiri/25/bab/951/
Amila Surendra -
http://www.linkedin.com/pub/amila-surendra/25/9a3/392
Hasala Surasinghe -
http://www.linkedin.com/pub/hasala-surasinghe/34/927/b2
Rajith Siriwardana - http://www.linkedin.com/in/rajithsiriwardana

We’d be really glad if you could extend your support on this.

Thank you for your valuable time.

Cool !
It's like 'Google summer of code' ...

Ishan Thilina Somasiri:

> We are required to do a 8 months long project in our final
> year. The project should consist of both a research and
> development component and an implementation. Our group
> is interested in areas such as,
>
>    * Information security
>    * Cloud
>    * Parallel computing
>    * Compilers

8 months is a lot of time, so projects like the projects of Google Summer of Code risk being too much small.

Some smaller projects are:
- Use LLVM to Just-in-time compile just the longer running compile-time functions (CTFE), for the LDC compiler (the functions that run quickly at compile-time are better left to the interpreter, because LLVM JIT has lot of overhead, it's not fast).
- Implement Windows exceptions for the LDC compiler.
- Implement efficiently the vector operations of the D language, and related matters.

A larger and more rearch-y project is to try to plug Z3 (http://z3.codeplex.com/ ) in the D front-end to run some contracts at compile-time. And generally to study how to improve a fortified version of the D language ("Strong-D"?) to be used in high integrity programming, where Ada/SPARK are used.

Another medium sized project is to use some parts of the C/C++ static analyser of Clang (http://clang-analyzer.llvm.org/ ), to create a similar tool for D-LDC.

Surely others will give you more ideas. Surely there is no lack of stuff to work on, because compared to the universe of tools/compilers for C/C++ the D ecosystem is almost a greenfield.

Bye,
bearophile

On Wednesday, 2 January 2013 at 16:15:00 UTC, Ishan Thilina
Somasiri wrote:
> Hi,
>
> I'm posting this on bhalf of a group of 4 students from the final
> year of Department of Computer Science and Engineering,University
> of Moratuwa, Sri Lanka. We are required to do a 8 months long
> project in our final year. The
> project should consist of both a research and development
> component and
> an implementation. Our group is interested in areas such as,
>
>    * Information security
>    * Cloud
>    * Parallel computing
>    * Compilers
>    * Operating Systems
>    * Distributed systems
>
>
> We’d like to know if the D programming language project has any
> project
> ideas that falls within our interest areas.
>
> For more information about the members of our group, please refer
> the
> following links.
>
> Ishan Thilina Somasiri -
> http://lk.linkedin.com/pub/ishan-somasiri/25/bab/951/
> Amila Surendra -
> http://www.linkedin.com/pub/amila-surendra/25/9a3/392
> Hasala Surasinghe -
> http://www.linkedin.com/pub/hasala-surasinghe/34/927/b2
> Rajith Siriwardana - http://www.linkedin.com/in/rajithsiriwardana
>
> We’d be really glad if you could extend your support on this.
>
> Thank you for your valuable time.

I think the strongest need for D right now is a better GC. A very
good start would be to make the GC precise on the heap. Compiler
support have been implemented some time ago, but the GC
implementation is lacking behind.

I'd be happy to personally help on that one. Other people here
are very competent as well.

On 2013-01-02 17:14, Ishan Thilina Somasiri wrote:
> Hi,
>
> I'm posting this on bhalf of a group of 4 students from the final
> year of Department of Computer Science and Engineering,University
> of Moratuwa, Sri Lanka. We are required to do a 8 months long
> project in our final year. The
> project should consist of both a research and development
> component and
> an implementation. Our group is interested in areas such as,
>
>     * Information security
>     * Cloud
>     * Parallel computing
>     * Compilers
>     * Operating Systems
>     * Distributed systems
>
>
> We’d like to know if the D programming language project has any
> project
> ideas that falls within our interest areas.

* Since you ask and mentioned compilers I would say a lexer/parser/front end written in D. This should be usable as a library which the following could be build upon:

* Syntax highlighter
* Compiler
* Refactoring
* Static analyzer
* Various other features usually found in an IDE

There are several other developers working on lexers/parsers, perhaps helping out with existing project could be a good idea.

* Implementing support for extern (Objective-C), aka ABI compatibility with Objective-C. This is already worked on but the project has staled:

http://michelf.ca/projects/d-objc/

* Implementing support for the new WinRT in Windows 8.

* I think it would be really cool if D could get support for AST (Abstract Syntax Tree) macros.

Other ides for Phobos are:

* New XML module
* Finishing an allocator design + new containers
* A package dealing with cryptographic

-- 
/Jacob Carlborg

On Wednesday, 2 January 2013 at 17:39:06 UTC, Jacob Carlborg wrote:
> * Since you ask and mentioned compilers I would say a lexer/parser/front end written in D. This should be usable as a library which the following could be build upon:
>
> * Syntax highlighter
> * Compiler
> * Refactoring
> * Static analyzer
> * Various other features usually found in an IDE
>
> There are several other developers working on lexers/parsers, perhaps helping out with existing project could be a good idea.
>

Indeed, I work on SDC, and a lot of work need to be done. However, I'm not sure how to staff 4 person on it, and i'm sure the GC things is more beneficial for D in general (and is in the scope).

> Other ides for Phobos are:
>
> * New XML module

Tango.xml is known to be one of the fastest xml lib ever made. Maybe we should simply use it (or change its interface to be more phobosish while keeping the internals).

On 2013-01-02 22:26, deadalnix wrote:

> Indeed, I work on SDC, and a lot of work need to be done. However, I'm
> not sure how to staff 4 person on it, and i'm sure the GC things is more
> beneficial for D in general (and is in the scope).

Just need to figure out what needs to be done and assign different tasks.

> Tango.xml is known to be one of the fastest xml lib ever made. Maybe we
> should simply use it (or change its interface to be more phobosish while
> keeping the internals).

I would prefer that, Tango also has a package for cryptographic. But this will most likely never happen. The licensees are not compatible and there has been some trouble around this is the past.

-- 
/Jacob Carlborg

On Wednesday, 2 January 2013 at 17:06:09 UTC, deadalnix wrote:
> On Wednesday, 2 January 2013 at 16:15:00 UTC, Ishan Thilina
> Somasiri wrote:
>
> I think the strongest need for D right now is a better GC. A very
> good start would be to make the GC precise on the heap. Compiler
> support have been implemented some time ago, but the GC
> implementation is lacking behind.
>
> I'd be happy to personally help on that one. Other people here
> are very competent as well.

+1 for the GC. Making it precise (and maybe generational) would easily be a nice project for 2 people at least. As well as allocators.

For other projects, D still doesn't have a nice unified interface to databases.

As for distributed computing, a port of Apache Hadoop and HBase to the D language that shows  better performance and memory characteristics than the Java version would be nice, and it would shed the light on the language.

On Thursday, 3 January 2013 at 04:40:59 UTC, SomeDude wrote:
>
> For other projects, D still doesn't have a nice unified interface to databases.
>

Actually, there is a project called adbi for the database interface.

A full featured precise, generational, concurrent GC would be a nice project for 3 or 4 people.

On Thursday, 3 January 2013 at 04:40:59 UTC, SomeDude wrote:
> +1 for the GC. Making it precise (and maybe generational) would easily be a nice project for 2 people at least. As well as allocators.
>

Generational seems completely unrealistic.