But with the widespread connectivity that we have it is very easy to launch a new language and to gain some small level of adoption. Programming languages are like kids: you can launch them in a night but it takes a lifetime of dedication and support to make them successful and given how hard it is to distinguish a small effort going nowhere from a future winner all this does is saddle up outsiders with the cost of maintenance or ultimately a rewrite in some other, hopefully wider supported language.
It would be great if people would just label their languages clearly as 'experimental, not meant for production' and 'long term supported' up front. Some languages, notably, 'Haskell' are successful in spite of actively holding back large scale adoption. This is a much healthier approach. It shows that the language really has merit.