On Fri, Mar 3, 2017 at 5:18 PM, David Kastrup <d...@gnu.org> wrote: > Frankly, I doubt that migration of large Python-based applications is > going to be a thing when nobody can even be bothered with immersing > himself in the problems with migrating LilyPond from Guile-1.8 to > Guile-2.
No, I don't think so. If we have guile-python3, the migration work becomes attractive to Guile community. Because each time you migrate a library, it can be used in all languages implemented on Guile platform. Write only once for all supported languages. Even if we don't care about Python (do we?), we have to write many libraries for Guile to make it more useful. It's the work Guile community has to do anyway. And LilyPond is not a good case here, since not everybody needs it. I think the best way to push a community is to provide convenient way to let users who care certain library to contribute it. But we don't have it now. For example, the documentation or tools to help 1.8->2.0. Python has tools for Python2->Python3 and documents for it. It is the management of Guile community, not technical problem. Fortunately, it's just management problem, and it's easier to improve than technical one, only if we found a persistent way to push and there's enough contributors.
