Suppose there is a very expensive commercial software program. For example, we'll call it Zyzzy's Super Expensive Software. Assume the Zyzzy Software Corporation has a trademark on its logo, but no explicit trademark on Zyzzy's Super Expensive Software as a product name. Zyzzy Software's primary product, responsible for a majority of their business, is Zyzzy's Super Expensive Software - if that program were to go away, the company would dissolve with it.
The company itself occasionally abbreviates the program as "ZSES", but colloquially in the industry the program is used within, it's simply referred to as "Zyz" or "Z", e.g. "Use Zyz to do that" or "I just updated to the latest version of Z today." Assume also that the company never uses the terms "Zyz" or "Z" in "official" communications or documentation, but perhaps its Twitter account has whimsically acknowledged the colloquial term, for example: "Hey Zyz-a-maniacs!"
Suppose I want to write an open source version of ZSES. I'd be doing this entirely "clean" - i.e. by simply observing the behavior of ZSES and implementing it my own way. No access to code, reverse engineering, etc. - purely based on freely available public documentation and observing the behavior of the commercial product. The actual output of ZSES and my clone are not necessarily guaranteed to be exact and in fact likely won't be, but would be functionally equivalent such that my clone would be a very simple replacement for almost all use cases. I would use a unique UI, but functionally the program could perform the same functions, and also let's assume my UI might be improved over the original product. In other words, my clone could potentially have a significant impact on Zyzzy's Software Company's sales and bottom line.
Now assume I want to decide what to name my clone. Similar to classic examples like LibreOffice (arguably based on Microsoft Office) and even Linux (Linus + Unix?), I decide I want to name my program "LibreZ" or "FreeZyz".
Given that "zyz" is just a colloquial name, not officially used by the company itself and definitely not trademarked but instantly recognizable by anyone in the industry, am I putting myself at obvious risk of litigation, assuming the situation that Zyzzy Software realizes they're losing money and I'm the reason? Consider that anyone in the industry would immediately see "FreeZyz" and know that it must be a free version of ZSES.
If Zyzzy had trademarked the colloquial names, then this would likely be a simple case and could easily be argued to potentially cause confusion. But the key point in my scenario is that the company has not trademarked the colloquial names, but perhaps it might be assumed by a lay person that they are trademarked because they are so closely associated with the product.
Naturally I know the easy answer is "you can be sued for anything" but I'm thinking beyond that - would a company have a credible case against a developer doing this? Could a company use a "prior art" type of argument to retroactively register a trademark and then go after a developer in this scenario, since it could be very easily demonstrated that the colloquial term is strongly associated with the commercial product?
(Note: When I originally wrote this I had used "John" and "Johnson" for the software and company name, but I decided to change it to "Zyz" because "FreeJohn" could be argued to just contain a common name, but "Zyz" is unique enough that it'd be easy to understand how it might be uniquely associated with a specific industry. It's also not a "generic" term, as in, perhaps you might colloquially refer to an iPad as a "pad" but "pad" is a generic enough term; "Zyz" is definitely not generic despite not being trademarked.)