Don’t burry them yet
You should never let your software stockpile slip into a legacy space where number of omitted, compromised or defunct items grows so considerably leaving you with just one option in hand: to scrap the code and rewrite it completely from scratch. I stand quite firm with proper management of application lifecycles, where any properly maintained applications are always in between releases taking part in a process of continuus improvement. Properly managed applications rarely reach end of their lifetime and never reach the latest release. My perspective is that continuus improvement and constant application reengineering are absolutely critical factors in weekly application maintenance. Agreed: maintaining a piece of technology isn’t possible in every single case. Some technologies just go out of fashion with time. Many systems are no longer being supported and even programming languages are being replaced over time with new versions. There’s also a question of cost involved in each of the above scenarios. It might be expensive to maintain some of the older releases of the UNIX system, it might be more effective to replace one technology with another, or it might simply become to expensive to research engineers specialising in some legacy technologies.