The issue of legacy systems & the uselessness of SW to help with legacy systems needs has been put on this list repeatedly, by at least JFS & myself... to resounding sounds of silence.
Consider this... from Capers Jones... there are some 50 "major" software languages, 1,500 "minor" languages & a new language is introduced at the rate of 1 per month. I assume the SW stack is on the far tail of the 1,500 minor list.
This is the sort of "diversity" the Global 2000 (or whatever company size label you want) is wrestling with.
I've spoken with a large firm that's been in the news recently & one of their senior technical officers told me they have some 28 DBMS engines in house. That's not databases, that's database engines. Personally I could only count to about 15.
How will the SW help with that mess?
From what I've seen over the past 8 years the SW is just going to add yet another layer of opaque complexity to the legacy systems. Intellectually stimulating? Absolutely. Useful? Dubious.
Jim Hendler told me to my face there was nothing in the SW for legacy systems... which I entirely believe.
One final rant... do think about the fast approaching Retirement Brain Drain.
My generation of mainframers who built these reliable, long-lived but difficult to grok legacy systems is headed to retirement. The Millennials mostly cannot spell JCL, much less know what it is (while they reinvent it).
So who's going to keep the Systems of Record running?
How is something in the SW going to help keep the legacy systems working?