"Jarrett Billingsley" <jarrett.billingsley@gmail.com> wrote in message news:mailman.186.1221682076.19733.digitalmars-d@puremagic.com...
> On Wed, Sep 17, 2008 at 12:39 PM, Rayne <DiscipleRayne@gmail.com> wrote:
>> Agreed.
>
> ..wait.  I thought you said that "its [sic] a matter of choice, not
> like your [sic] really missing out on anything"?  Now you agree to the
> opposite?

You've convinced him!
Surprised by your own skills?

L. 

On Thu, Sep 18, 2008 at 4:04 AM, Lionello Lunesu <lionello@lunesu.remove.com> wrote:
>
> "Jarrett Billingsley" <jarrett.billingsley@gmail.com> wrote in message news:mailman.186.1221682076.19733.digitalmars-d@puremagic.com...
>>
>> On Wed, Sep 17, 2008 at 12:39 PM, Rayne <DiscipleRayne@gmail.com> wrote:
>>>
>>> Agreed.
>>
>> ..wait.  I thought you said that "its [sic] a matter of choice, not like your [sic] really missing out on anything"?  Now you agree to the opposite?
>
> You've convinced him!
> Surprised by your own skills?

More along the lines of "first time I've ever seen anyone change their opinion on the internet" :P  Unless it's just a simple mistake.

> Yes, it is. But I think the most important is that the Tango group and Phobos cooperate about the core. Because if the third one do this things would be very hard. Because of Phobos is in developing and Tango is in developing too. So if we seperate the core from Phobos and Tango, and then Phobos and Tango is based on the Core, then they can working together. This will be a great success in the D community.

I absolutely agree. Having a single core is essential.

On Wed, 17 Sep 2008 08:26:22 -0400, Jarrett Billingsley wrote:

> On Wed, Sep 17, 2008 at 8:18 AM, Rayne <DiscipleRayne@gmail.com> wrote:
>> They both pretty much accomplish the same damned thing, I don't really understand what all his fuss is about. Use tango if you like it or use phobos, its a matter of choice, not like your really missing out on anything.
>>
>>
> Except for all the libraries that are developed for only one stdlib. And
> for library writers, it's hell for just that reason -- either you target
> one stdlib and exclude yourself from those who use the other, or you
> (god forbid) try to make your library work with _both_ stdlibs.
>  It's a pointless bifurcation in the community.

agreed

Graham St Jack wrote:
>> Yes, it is. But I think the most important is that the Tango group and
>> Phobos cooperate about the core. Because if the third one do this things
>> would be very hard. Because of Phobos is in developing and Tango is in
>> developing too. So if we seperate the core from Phobos and Tango, and
>> then Phobos and Tango is based on the Core, then they can working
>> together. This will be a great success in the D community.
> 
> I absolutely agree. Having a single core is essential. 

And this would give the opportunity to improve the core.  For example, what about pluggable garbage collectors?  So you could change the GC type, perhaps not at runtime, but at least have more say into the GC technique used.  It might also give more people a chance to work on GC technique if it were an easy plugin architecture.  I know I've felt the wanderlust to try my hand at GC programming.

Disclaimer: I know next to nothing about GC implementation.

Moving onward, that would also allow us to take parts of Phobos and Tango that coincide and put them into the new core, consolidating the implementations into one location, perhaps unifying the libraries in the distant future and ending the "Tango vs. Phobos" """"Discussions"""" once and for all?  I am unsure, but it's a possibility to consider.

Chris R. Miller wrote:
> Graham St Jack wrote:
>>> Yes, it is. But I think the most important is that the Tango group and
>>> Phobos cooperate about the core. Because if the third one do this things
>>> would be very hard. Because of Phobos is in developing and Tango is in
>>> developing too. So if we seperate the core from Phobos and Tango, and
>>> then Phobos and Tango is based on the Core, then they can working
>>> together. This will be a great success in the D community.
>>
>> I absolutely agree. Having a single core is essential. 
> 
> And this would give the opportunity to improve the core.  For example, what about pluggable garbage collectors?  So you could change the GC type, perhaps not at runtime, but at least have more say into the GC technique used.  It might also give more people a chance to work on GC technique if it were an easy plugin architecture.  I know I've felt the wanderlust to try my hand at GC programming.

Already done in the Tango runtime.  Pluggable at link-time, but I can't envision a reason to have dynamically pluggable GCs.

> Moving onward, that would also allow us to take parts of Phobos and Tango that coincide and put them into the new core, consolidating the implementations into one location, perhaps unifying the libraries in the distant future and ending the "Tango vs. Phobos" """"Discussions"""" once and for all?  I am unsure, but it's a possibility to consider.

Yup.


Sean