The Host Software Component and ACSLS worked to control the tape libraries, but had multiple problems. HSC has been described as, “A 1990s product built in the 1980s with 1970s technology.” StorageTek was having a real problem hiring assembler language programmers by the mid-1990s.
So a recurring hope was to replace HSC, and perhaps ACSLS, with a whole new set of software. At one point, one of our lead engineers proposed writing a new system in Java. (Note: This was well before I learned Java and I was not involved.)
They selected a hand-picked group of developers and hired an OO consultant, since STK had absolutely no expertise. The consultant gathered the team in conference rooms for months while he taught them OOAD. What we heard mostly was that they were “drilling for objects.”
After a long while, they started development…on MicroSoft’s Java compiler and Windows PCs. They planned to do the development there, then just load the code into IBM’s JVM to run on the mainframe. After all, with Java you could “write once and run anywhere.”
Eventually, they had an initial subset of the software running on the PC against the library simulator. They were ready to try it on MVS!
The results were disastrous! Apparently, STK was the very first customer to write any serious software on IBM’s JVM. It was not implemented correctly, and the software that ran well on MicroSoft’s JVM did different things on IBM’s. There was no way to fix it other than to redo the development on IBM.
This was a very expensive fiasco, and in the great STK tradition, lots of people lost their jobs.
President & CTO at Judge Consulting and Copley Consulting Group
2wThis is pretty cool!!