Part of the problem is that they *DON'T* have all the source code. Also, many systems were not well documented and the original developers are gone.
On top of that, we don't really know what development platform the old system used or what they are using in the new design. I spent my career in IT starting with IBM mainframes in the late '60s, going through the PC revolution in the 80's/90's, Y2K, Internet based web applications, smartphones. I lost count of the number of conversions, rewrites, application obsolescence and abandonment, false starts, etc. that I was involved with. But major upgrades/rewrites are never trivial, and often there is little beyond conceptual design that can be salvaged from the old system. I'm not saying IBM/WU is blameless in all the criticism they are getting … just that the fixes are probably not as simple as they seem like they should be looking in from the outside.
… this is just a 2 cent observation from an old f@rt who doesn't really know what is going on at WU. I don't even own a weather station yet (waiting for the Acurite Atlas Elite before I decide what to get), but I sure hope WU is successful. I use WU all the time to monitor other stations in my neighborhood, and for the most part WU is reliable for that purpose. I suppose I may have a different opinion once I have a weather station of my own.