The word legacy as mostly frame with negative connotations. The is unproductive.

Because it is also (I got these insights from person.swizec-teller):

• Legacy code is code that works.
• Legacy code is legacy because it works. Your code either dies a prototype or it becomes legacy.
• That legacy code is paying your salary.
• Legacy is an asset

So legacy code is actually not what most often mean by it. What we probably more often meen by it is it being a big-ball-of-mud.

It also often comes down to the question what to do with legacy code or how to think about it or relate to it. The second question is a better one, I think, because it gives you a better ground to stand on when you go for the first question.

Often the question comes down to "should re rebuild the whole system and by that get rid of all the legacy code"

I think first we have to differentiate between legacy code and a bit ball of mud. If we have identified that most of what is in from of us is legacy code my hunch is that no we should probably not rebuild the whole thing. There is to much value having received legacy that needs to be appreciated not for what it has done for the current state of affears but specially for what it can do now too.

Can be looking at legacy code as something that fosters a sense of beauty?

How does lecacy code relato to abundance, wealth, productivity and maturity?

What is quaint code?

How to introduce domain-driven-design into legacy systems with rebuilding the whole system. person.eric-evans talks about creating bubble context.