Polishing D - suggestions and comments

Don Clugston dac at nospam.com.au
Mon Jan 28 08:54:36 PST 2008


Lars Ivar Igesund wrote:
> Bill Baxter wrote:
> 
>> Leandro Lucarella wrote:
>>> Robert Fraser, el 25 de enero a las 17:31 me escribiste:
>>>> Lars Ivar Igesund wrote:
>>>>> I know there are a few places in Tango where an additional line may be
>>>>> needed (and many where you'll need quite a few less), but without exact
>>>>> examples of what people think is a problem, it is hard to make
>>>>> qualified decisions on where to make improvements.
>>>> Not to be negative, but I think no matter how many
>>>> tests/examples/whatever show that Tango is comparable or better in speed
>>>> and efficiency to Phobos, the stigma of a feature-rich, strongly
>>>> abstracted/modular standard library reminds of Java and .NET . I think
>>>> the fear is less based on logic and more based on association between a
>>>> modular standard library and VM-based languages.
>>> I second that. Phobos is closer to C/C++ stdlib, Tango to Java/.NET.
>>> I think it would be great to have 2 "compatible" standard libraries. One
>>> minimalist for embeded and such (phobos) and one for "big" (or not that
>>> big) desktop applications (tango). Of course both should be compatible
>>> and it had more sense if the "big" library were a super-set of the
>>> "small" one.
>> However, when it comes to the low-level parts of the library (gc,
>> threading, etc), I don't really see anyone arguing.  Tango's seems to be
>> better.  It seems like those improvements should just be rolled back
>> into Phobos.  Then Tango could go back to being a regular library that
>> doesn't require you to "get religion" first.
> 
> Tango require you to "get religion" ? 

Also, Tango has never been a "regular"
> library, if that means a library without its own runtime.

It was in the Mango days. I don't see any reason why the low-level stuff 
couldn't replace the Phobos stuff.




More information about the Digitalmars-d mailing list