
Traditionally, enterprises have considered old hardware, and the mainframe in particular, to be "legacy technology". But in a week in which there was legacy modernization tools and services news, it is worth reminding ourselves of exactly what is meant by the term.
Many think of legacy technology as the older stuff they can't get rid of because it still plays an important role. The bright spark who procured it for the company has almost certainly left by now, and in some cases left this mortal coil, too. Mainframes, mini-computers like the Vax, anything that runs VMS, anything that includes either of the names Sperry or Burroughs in its documentation: these are all almost certainly legacy technologies.
But can a Java application also be legacy? In fact it can, because in many cases older Java applications are like black boxes - no one knows what they were written for or which business users and processes remain reliant on them. The point being that it is not only mainframes and minis that are legacy technology today. The list is far longer than that, and it includes both software and hardware.
Only by acknowledging this are IT departments likely to keep on top of their ageing IT assets. Legacy technologies deserve slightly different treatment. A decision needs to be taken whether to maintain them, to turn them off, or to modernize them. Modernizing them can take several forms too, including rewriting, migrating, extending, or perhaps "wrappering" with more modern technologies like web services.
There are lots of things you can do with legacy technologies, depending of course on the size of your IT budget. But whatever you do to them, it's not a good idea to simply ignore them.
> anything that runs VMS [...] almost certainly legacy technologies.
HP (who owns VMS now) claims to have increased their installed base by 10% in 2005.
Of course, VAX systems are "legacy", since they haven't been made in 10+ years. But VMS (renamed OpenVMS) is available on both Alpha and Itanium processors. Hardly legacy...
Stanley you are absolutely right -- VMS was probably a bad example because as you say some companies may have recently deployed apps on VMS, which would make them quite up-to-date (and hence not legacy).
However, for the other 90% or whatever of companies running VMS, it probably *is* considered legacy: indeed companies like Jacada and Transoft specialize in VMS migration services for exactly that reason.
I am not saying there is anything inherently bad about VMS -- the reason for its ongoing popularity is its undisputed reliability. I heard that they had to extend the integer in VMS that stores the number of days between a reboot from something like 4 places to 6, because the thing just never falls over. I can't remember the exact figures but you get the drift.
Anyway many thanks for the comment -- always welcome.