On 2013-02-28 00:37:50 +0000, Andrei Alexandrescu <SeeWebsiteForEmail@erdani.org> said:

> Hello,
> 
> 
> Walter and I have had a long conversation about the next radical thing to do to improve D's standing. Like others in this community, we believe it's a good time to consider bootstrapping the compiler. Having the D compiler written in D has quite a few advantages, among which taking advantages of D's features and having a large codebase that would be its own test harness.
> 
> By this we'd like to initiate a dialog about how this large project can be initiated and driven through completion. Our initial basic ideas are:
> 
> 1. Implement the dtoh standalone program that takes a D module and generates its corresponding C++ header.
> 
> 2. Use dtoh to initiate and conduct an incremental port of the compiler. At given points throughout the code D code will coexist and link with C++ code.
> 
> 3. At a point in the future the last C++ module will be replaced with a D module. Going forward there will be no more need for a C++ compiler to build the compiler (except as a bootstrapping test).
> 
> It is essential that we get support from the larger community for this. This is a large project that should enjoy strong leadership apart from Walter himself (as he is busy with dynamic library support which is strategic) and robust participation from many of us.
> 
> Please chime in with ideas on how to make this happen.

Actually, I think it'd be easier and faster to convert it all in one chunk. Perhaps I'm a little too optimistic, but I did port successfully a game from D to C++ once and it was not that difficult. I never bothered with having a half-translated version that'd work.

My impression is that trying to add some layer to make the intermediary state compile is more likely to introduce bugs than to help. The current architecture isn't modular enough to do that without many complications.

-- 
Michel Fortin
michel.fortin@michelf.ca
http://michelf.ca/

On 3/1/2013 7:43 AM, H. S. Teoh wrote:
> Wow. You make me feel really lucky that at my day job, I once made a
> request to use a particular piece of open source software, and the legal
> department actually replied with "the license is MIT, it should be OK,
> approved."

This is exactly why we are using a well-known license, rather than rolling our own.

On Saturday, 2 March 2013 at 02:35:34 UTC, Walter Bright wrote:
> On 3/1/2013 7:43 AM, H. S. Teoh wrote:
>> Wow. You make me feel really lucky that at my day job, I once made a
>> request to use a particular piece of open source software, and the legal
>> department actually replied with "the license is MIT, it should be OK,
>> approved."
>
> This is exactly why we are using a well-known license, rather than rolling our own.


Anyway the DDMD maintainer was asked about the license in 2010,
he never picked one but it seemed like he was open to
anything:
http://www.dsource.org/forums/viewtopic.php?p=25763#25763

On 3/1/2013 6:55 PM, Andrej Mitrovic wrote:
> On Saturday, 2 March 2013 at 02:35:34 UTC, Walter Bright wrote:
>> On 3/1/2013 7:43 AM, H. S. Teoh wrote:
>>> Wow. You make me feel really lucky that at my day job, I once made a
>>> request to use a particular piece of open source software, and the legal
>>> department actually replied with "the license is MIT, it should be OK,
>>> approved."
>>
>> This is exactly why we are using a well-known license, rather than rolling our
>> own.
>
>
> Anyway the DDMD maintainer was asked about the license in 2010,
> he never picked one but it seemed like he was open to
> anything:
> http://www.dsource.org/forums/viewtopic.php?p=25763#25763

I'd recommend Boost or GPL. Anyhow, it's a pity to see his work go to waste because of no license.

On Saturday, 2 March 2013 at 03:01:20 UTC, Walter Bright wrote:
> On 3/1/2013 6:55 PM, Andrej Mitrovic wrote:
>> On Saturday, 2 March 2013 at 02:35:34 UTC, Walter Bright wrote:
>>> On 3/1/2013 7:43 AM, H. S. Teoh wrote:
>>>> Wow. You make me feel really lucky that at my day job, I once made a
>>>> request to use a particular piece of open source software, and the legal
>>>> department actually replied with "the license is MIT, it should be OK,
>>>> approved."
>>>
>>> This is exactly why we are using a well-known license, rather than rolling our
>>> own.
>>
>>
>> Anyway the DDMD maintainer was asked about the license in 2010,
>> he never picked one but it seemed like he was open to
>> anything:
>> http://www.dsource.org/forums/viewtopic.php?p=25763#25763
>
> I'd recommend Boost or GPL. Anyhow, it's a pity to see his work go to waste because of no license.

I'll send him an e-mail to see if he's still around, maybe he'd be interested in this again.

On Fri, Mar 01, 2013 at 07:01:20PM -0800, Walter Bright wrote:
> On 3/1/2013 6:55 PM, Andrej Mitrovic wrote:
> >On Saturday, 2 March 2013 at 02:35:34 UTC, Walter Bright wrote:
> >>On 3/1/2013 7:43 AM, H. S. Teoh wrote:
> >>>Wow. You make me feel really lucky that at my day job, I once made a request to use a particular piece of open source software, and the legal department actually replied with "the license is MIT, it should be OK, approved."
> >>
> >>This is exactly why we are using a well-known license, rather than rolling our own.
> >
> >
> >Anyway the DDMD maintainer was asked about the license in 2010, he never picked one but it seemed like he was open to anything: http://www.dsource.org/forums/viewtopic.php?p=25763#25763
> 
> I'd recommend Boost or GPL. Anyhow, it's a pity to see his work go to waste because of no license.

I would personally go for GPL, but it does scare certain companies off -- I've personally witnessed that.  Just FWIW.


T

-- 
Computers aren't intelligent; they only think they are.

On Friday, 1 March 2013 at 06:57:31 UTC, Zach the Mystic wrote:
> So you're saying some of our dogfood is actually caviar then...
>
> I would divide the caviar into two groups, manifest and hidden. The manifest caviar is the easiest to sell. Hidden caviar is the benefits which are unexpected by at least a portion of the D community. Each piece of hidden caviar therefore needs one or more champions.
>
> Not that this is a perfect example, but the lexer being assembled by Brian and Dmitri seems to have a spark of the hidden caviar about it, lending weight to the "clean room" camp. The politics of "existing" versus "clean room" must be mastered because there's a lot of room for resentment there if the wrong choices are made, it seems to me.
>
> One thing both "clean room" and "existing" have, or should have, in common is the test suite, which is probably a better spec than the spec is. Perhaps a method can be devised which makes it easy to divide and conquer the test suite.

By "clean room" I really meant starting from scratch, regardless of license.

On Saturday, 2 March 2013 at 03:01:20 UTC, Walter Bright wrote:
> On 3/1/2013 6:55 PM, Andrej Mitrovic wrote:
>> On Saturday, 2 March 2013 at 02:35:34 UTC, Walter Bright wrote:
>>> On 3/1/2013 7:43 AM, H. S. Teoh wrote:
>>>> Wow. You make me feel really lucky that at my day job, I once made a
>>>> request to use a particular piece of open source software, and the legal
>>>> department actually replied with "the license is MIT, it should be OK,
>>>> approved."
>>>
>>> This is exactly why we are using a well-known license, rather than rolling our
>>> own.
>>
>>
>> Anyway the DDMD maintainer was asked about the license in 2010,
>> he never picked one but it seemed like he was open to
>> anything:
>> http://www.dsource.org/forums/viewtopic.php?p=25763#25763
>
> I'd recommend Boost or GPL. Anyhow, it's a pity to see his work go to waste because of no license.

I'm no copyright lawyer, but I think ddmd being a derivative work from dmd should probably inherit the license from it (although I prefer Boost as in my opinion it is more liberal that GPL).

If someone is willing to bring the project back from it's stale state - I'm more than willing to help (by both writing patches and explaining how the existing code works). He must also understand that ddmd is a couple of years behind dmd, and updating it is a monkey-work that requires little thinking but lot's of time.

It usually took me a couple of hours to generate a code diff between 2 subsequent dmd releases, apply it to ddmd and run tests (normally I was compiling druntime.lib and compare it against one produced by dmd - they should match byte-for-byte) - that is if everything went smoothly.

Sometimes they won't match so I compile druntime one file at a time, find what files don't match, reduce test case even further until I find the source of the problem and fix it. For me (someone who only has very basic understanding of the codebase) it could took another extra couple of hours to fix bugs. Recent releases also started containing more features and a lot more bugfixes (= larger diffs, more difficult to merge). Let's make it 5 hours per release.

20 versions behind x 5 hours of work ~ 100 man-hours to bring it up to date. A little optimistic probably, but it doesn't sound too bad :)

Anyway, feel free to take the sources and attach whatever license you want to it.

On Saturday, 2 March 2013 at 03:16:59 UTC, H. S. Teoh wrote:
> On Fri, Mar 01, 2013 at 07:01:20PM -0800, Walter Bright wrote:
>> On 3/1/2013 6:55 PM, Andrej Mitrovic wrote:
>> >On Saturday, 2 March 2013 at 02:35:34 UTC, Walter Bright wrote:
>> >>On 3/1/2013 7:43 AM, H. S. Teoh wrote:
>> >>>Wow. You make me feel really lucky that at my day job, I once made
>> >>>a request to use a particular piece of open source software, and
>> >>>the legal department actually replied with "the license is MIT, it
>> >>>should be OK, approved."
>> >>
>> >>This is exactly why we are using a well-known license, rather than
>> >>rolling our own.
>> >
>> >
>> >Anyway the DDMD maintainer was asked about the license in 2010, he
>> >never picked one but it seemed like he was open to anything:
>> >http://www.dsource.org/forums/viewtopic.php?p=25763#25763
>> 
>> I'd recommend Boost or GPL. Anyhow, it's a pity to see his work go to
>> waste because of no license.
>
> I would personally go for GPL, but it does scare certain companies off
> -- I've personally witnessed that.  Just FWIW.
>
>
> T

Yeah, GPL can have that effect.

One of the most friendly license I've seen is sqlite's
http://www.sqlite.org/copyright.html

--rt

On Sat, Mar 02, 2013 at 04:46:57AM +0100, Denis Koroskin wrote: [...]
> Anyway, feel free to take the sources and attach whatever license you want to it.

Unfortunately, I don't think copyright works like that. You (the author) must be the one who licenses it. Any license applied by others will most probably be invalid, since they are not the real copyright owner, and, in the case of actual legal disputes, such a license will be indefensible.


T

-- 
Береги платье снову, а здоровье смолоду.