What you have essentially done is put safety at risk in doing so.
So how is safety not put at risk if the RVR is exactly at minimum at 1000ft but is reported below minimum at 999ft and the approach is continued?
I would say that while the scenario presented by the OP is definitely counter to regulations, the threat to safety is moot.
The intent behind the regulation is to discourage an attempted approach where the chance of achieving the required visual reference at DH is slim. It is designed to prevent the “let’s have a look” mentality taking control, which might result in a continued approach, below DH with poor visual references. In the days of manual control and simple instrumentation, such a restriction was sensible.
Having said that, as long as a strict understanding and observance of the required visual reference is adhered to, a continued approach below 1000ft down to the DH with RVR below minimum is not, in itself, “unsafe”. It’s what you do at DH that might be unsafe.
Human nature being what it is, the rules are drafted to minimise the chances of a difficult decision having to be made close to the ground. The (first) decision point is shifted up to 1000ft. To cater for unfortunate variations in visibility, and because of the perceived risk in starting a go-around at a low height, the option to continue remains below 1000ft. Given the performance and reliability of modern autoflight systems, where a simple press of TOGA or an advance of the thrust levers results in a safe go-around manoeuvre, it seems a bit old-fashioned to prevent a crew from “having a look”, even if the RVR is below minimum above 1000ft; the rules allowing them to do so if the RVR reduces below that height.
There are definite risks to safety associated with assessing and responding to limited visual references close to the ground. More tailored use of high-fidelity simulators and visual systems would go a long way to manage and reduce this risk.
In my experience, during simulator training and testing, one either sees “nothing” at DH, or the required visual reference appears 50ft or more above DH. Perhaps we should introduce an element of “visual ambiguity” at this critical time and then train and test accordingly?