I think the idea of cross language is that an IO is only in one language and others can use that IO. My feeling is that the idea of “what language is this IO in” becomes an implementation detail that folks won’t have to care about longer term. There are enhancements needed to the expansion service to make that happen but that’s my understanding of the strategy.
On Tue, Jan 10, 2023 at 7:40 AM Austin Bennett <aus...@apache.org> wrote: > This is great, thanks for putting this together! > > A related question: are we as a community targeting java to be the > canonical/target IO language if an IO does not currently exist? If that is > not the case, then I would imagine we are hoping that we might eventually > also wind up with good examples for implementing IOs in other languages as > well [ not suggesting that you/John address that, but that we add GH Issues > as that might be worthwhile to hope others take on ]? > > > > On Mon, Jan 9, 2023 at 8:58 AM John Casey via dev <dev@beam.apache.org> > wrote: > >> Hi All, >> >> I spent the last few weeks of December drafting a "How to write an IO >> guide": >> https://docs.google.com/document/d/1-WxZTNu9RrLhh5O7Dl5PbnKqz3e5gm1x3gDBBhszVF8/edit# >> >> and an associated code sample: https://github.com/apache/beam/pull/24799 >> >> My goal is to make it easier for a new IO developer to create a new IO >> from scratch. This is intended to complement the various standards >> documents that have been floating around. Where those are intended to >> prescribe structure of an IO, this is more focused on the mechanics of >> internal design. >> >> Please take a look and let me know what you think, >> >> John >> >
