On Monday, 30 March 2015 at 08:53:15 UTC, Ola Fosheim Grøstad wrote:

same theme. I pick them based on what they+ecosystem is good at, not the language by itself. So basically, you have to be best at one particular application area to do well. Go is aiming to have a good runtime for building smaller web-services, and they are getting there. Because they focus.

It is necessary to be appealing to Ola by Ola's standards for a language to appeal to other people?

I think how it actually works is that you have to find a small but focused group of people to love you lots. Then they tell other people and over time you get better at appealing to those for whom you weren't ready before. So that's similar to what you suggest in one sense, except that the chicken and egg problem is smaller. Sociomantic didn't consider the ecosystem when selecting D (or at least were not put off by its immaturity). But if in five years time their competitors realize the possibilities for doing things better, they will certainly benefit from the work Sociomantic has done on improving D (even purely as a demanding use case, but it's more than that). [And Sociomantic won't lose, in my uninformed estimation, because edge is dynamic].

Similarly in the tiniest of ways, I didn't weight the library situation very heavily in picking D. I have written a couple of bindings (painfully, before I got Dstep to work or knew the language very well!) and wrappers and if anyone like me arrives subsequently then it will be that little bit easier. So that's one more reason why it can take a couple of decades for something to be an overnight success - it takes time for paths and habits to be formed, and there are threshold effects, beyond which there is a phase change. So D's long-term prospects will be shaped by how it responds to the challenges of growth. Looks good to me right now.


Reply via email to