The Death of D. (Was Tango vs Phobos)

[Paul D. Anderson]

Robert Fraser Wrote:

> I see the D community heading for a split. One side will be D1 (hopefully a D 1.5) +tango, and the other side will be D2+phobos.  I don't think of this as a bad thing nessescarily, just two different directions.

First, let me say that I have the highest regard, both personally and professionally, for Walter (and Andre and Bartosz and the others who are working on D 2.0) and also for the Tango development team. Both groups have a lot of talent and have worked hard to give us (at no cost!) a great language and a great library. 

The Tango developers have done important and valuable work, providing sturdy, efficient and rigorous implementations of many common software components. They have not only provided the Tango library, but they've also wrung out the language and pushed its capabilities to give us all a better picture of what is possible. (I'd love to see what they could do with the new features in D 2.0.) 

And they have been careful not to step on Walter's toes while they do this. They discussed the problem with Walter at the first D Language Conference and agreed on a solution: a common runtime module. They have narrowed the incompatibilities between phobos and tango to a minimal set and are waiting for Walter to meet them there. To live up to his part of the bargain, to put it bluntly. A year later he still hasn't done so.

But Walter drives development his own way. Walter himself has said that the only difference between himself (and D) and a lot of other replacements for C/C++ is that he's stayed with it longer than they have. And so Walter will listen and respond to criticism, new concepts, alternative implementations and frantic pleas, but, at the end of the day, he will work toward what he sees as the best solution to the most pressing problems. This is mostly a good thing -- Walter has a long-term vision for the D programming language and he is determined to see it through. Of course, this also has a downside. Work that doesn't advance the main line that he has mapped out doesn't get his attention, or his time. It's not that he's unwilling, but he's doing a lot of work on developing what he sees as more fundamental things.

So I don't think it is likely that Walter will delay or disturb his current effort to get D 2.0 pushed out the door. And he's made it clear that he not interested in going back to "fix" D 1.0. But what I would really like to see (and I don't think it's asking too much) is a clear statement from Walter that he will indeed make the changes the Tango developers are asking for at some future date. I don't think he even needs to commit himself to a specific date, just a reference point -- when D 2.0 is final, or when the functional programming stuff is in place, or whenever. I think all that we really NEED at this point is assurance that the two groups are not working at cross purposes.

If this has already happened, if there is some agreement in place that I just don't know about, then I apologize for this rant. But it seems like little enough to ask.

Paul