For me, it sounds like an ad-hoc solution. I see many potential issues with this:
- the text will not be kept up to date (wrong information instead of none, which is worst)
- the text will be difficult to extend when the domain is extended. Natural languages are not as composable as a domain is.
- domain may be too complex to express in a sentence.
For me, it is better to work on improving the DomainParser to be able to stringify more domains.
And also to implement it on the server side to allow server to provide better error message also.