I had my issues with George Klaus of Epicor (now gone after merger and cashing out), but I really respected when he told us at a presentation that he's willing to take crap from the analysts and investors in order to invest a significant amount into R&D so they can be competitive over the long haul. In other words, take a long term, Japanese like, view of business and not just manage to the quarter via cost cutting to make bonuses.
Now they are sitting here with an extremely broad, modern, flexible and deep product that can now be back filled on the quality and performance front which is what their latest release, that had a 50 man year investment, focused on.
MAS90 is reliable (for the most part), because fundamentally most of the logic is 25 to 30 years old! With the same design issues of detail, history files, etc to deal with the huge cost of 5 MB hard drives.
Before they had history files, they dumped the data monthly. Then storage came down, and instead of redesigning the software they cheaped out and didn't mess with the core logic and instead built 'history files' to shuttle off data to other tables (which then become out of sync due to the crappy flat file shared database). So in 2012 we have to deal with ""AR Open Invoice File"" (which contains way more than just open invoices now), vs. AR Invoice History File, (which contains AR Invoice History AND Open Invoices). And it seems many of us here just give this stupidity and extreme cheapness on Sage's part to invest the R&D to fix design issues dating back 25 years a pass...