Words can be flexible, they can be changed. Language is dependant upon usage, and so new words can crop up or old ones can be co-opted to new meanings. With games, a new language had to be created from scratch once those first games appeared. As they became more popular, the language evolved and terms were borrowed or created from wherever was required.
Even now, such terms are often not entirely fit for purpose. We call certain games "First Person Shooters", where they might previously have been known as "Doom-clones" or just "Shooters". All these terms are descriptively opaque, unless you are aware of the culture from which they sprung (and even then, they don't completely convey the full nature of the game that they attempt to describe). They're a shorthand, but they're imperfect, and no doubt a lot of these sorts of terms will change over time as better terms are coined.
"Rogue-like" is one such term that I thought was long past its sell-by-date. The game "Rogue" was first released in 1980, and it didn't take long for others to copy various elements of its design. Such games were termed "Rogue-likes" to convey their similarity to the original Rogue. These games often copied many of the same design elements: Characters and monsters based loosely on Dungeons & Dragons role-playing standards, procedural generation, game over after a single death, multi-level dungeons, and so on.
The term Rogue-like went out of favour for a while, with that type of game being overshadowed by the RPG - a catch-all term for a genre that is so wide that it would require an article to itself. The general idea was to create an approximation of the tabletop role-playing game using a computer to do the dice-rolling and paperwork and leaving the player to enjoy a more crafted and directed story, with well defined rules and mechanics. The fact that games as diverse as Mass Effect, Pillars of Eternity, Dark Souls, Fallout and Skyrim can share the same label somewhat dilutes it's usefulness.
The labels we give games are attempts to provide descriptive terms, but they vary in their application and their usefulness. Sometimes a game can be described in terms of another: Rogue-like. Or it may be defined in terms of mechanics: Real-Time Strategy. Or it may be described in a more abstract term: Adventure Game. Sometimes these terms are useful, sometimes they are only useful in context, and sometimes they rely on prior knowledge.
In the past decade or two, there has been an explosion of independent games development. People chose to makes games that reflected the experiences they had when they were younger, often harking back to styles and mechanics that had gone out of fashion in mainstream game development. Rogue-likes had remained popular throughout the years, but there was now a potential audience like never before. Some of these new games were recognisably like Rogue, their history easily traced back. Others used certain mechanics, but wanted to make something very different.
The terms "Rogue-like" and "Rogue-lite" were used, no matter how ill-fitting they might be. They were used as umbrella terms, much like RPG, where there are certain common characteristics but broad enough that it allowed for so many games to occupy that same ground.
The problem arises when you consider that the vast majority of people who play games will never have heard of Rogue, let alone played it. They at best will associate it with certain mechanics, if only by association with other games of its type.
Like so many terms in gaming, it feels like we're a long way from creating the terms we need to fully describe the immense variety and complexity of the medium. It's a young medium, with plenty of room to grow, so this isn't a pressing need. With the advent of VR experiences, things are only going to get increasingly complex, and it may be some time yet before gaming has the level of descriptive language about itself as literature or film does.