As a plugin author, I’m completely in the dark about the actual benefit to me that comes with migrating an existing plugin the new split mode paradigm.
Are there any metrics regarding what percentage of the JetBrains IDE user base is actually using IntelliJ’s remote development capabilities that can be shared with the community?
Without that information, it’s hard to justify investing in split mode given the additional complexity and implementation overhead it entails for any sizable plugin.
Feels the same here, I don’t know if I need to migrate to split mode or how to do it.
If my memory is correct, somebody shared some metrics: 30% of users do remote development. Honestly, it’s hard to believe 
. At least, we would need more information, because we can easily manipulate these metrics.
Also, the remote-dev APIs are not finalized, and it’s not clear if paid plugins are supported.
Finally, there are still some GUI glitches (at least with Swing) when running in spitted-mode.
I may share another aspect of usage: our top 10 large enterprise customers use Remote Development extensively, although unfortunately I can’t disclose their names.
Just in case, my current customer (one of the top 10 banks worldwide) moved almost all their development to devpod and some other big customer are moving or moved already to the same paradigm.
There are several advantages (cost, uniform development environment, using Linux but still having the control of Windows, etc.) and some problem.
About migration I wrote a blog post with my experience https://medium.com/@m.sciachero/my-journey-on-intellij-plugin-v2-1c15ce97f34a
My path was a little different from what JetBrains suggest, but how the plugin was implemented and how the modules are organized in IJ that was the only way I found to split it.
But it’s not simple and you will easily find out that you changed too much to be safe.
@yuriy.artamonov Thanks for sharing that additional information. Do you know if the 30% number @jonathanlermitage.1 shared is roughly accurate? That would certainly help ground the discussion around what the true addressable market is for plugins that properly implement support for remote development.
Jonathan’s comment on the verifier thread ( Stricter plugin verification in IntelliJ Platform Gradle Plugin 2.15.0 - #13 by jonathanlermitage.1 ) regarding uncertain interpretation of the relevance of the top 10 largest enterprise customers using remote development to 3rd party paid plugin developers is also a fair point. I had a similar reaction, wondering if those enterprises actually allow third party plugins outside a small number of pre-approved OSS options.