I think your points are complementary to the article. It does seem like if a tool reaches a certain popularity threshold, programmers start porting it to every

I think your points are complementary to the article. It does seem like if a tool reaches a certain popularity threshold, programmers start porting it to every language and environment under the sun. And that often takes the form of building things in to larger tools. But the academic tools this post highlights are typically standalone and haven't reached that threshold. My suspicion is that the author's point applies most to academic tools that have trouble reaching a critical mass.