Floating Point Number parsing: Is there a Catch All algorithm?

I think the best you can do in this case is to take their input and then show them what you think they meant. If they disagree, show them the format you're expecting and get them to enter it again.

The difference between 12.345 in French and English is a factor of 1000. If you supply an expected range where max < 1000*min, you can easily guess.

Take for example the height of a person (including babies and children) in mm.

By using a range of 200-3000, an input of 1.800 or 1,800 can unambiguously be interpreted as 1 meter and 80 centimeters, whereas an input of 912.300 or 912,300 can unambiguously be interpreted as 91 centimeters and 2.3 millimeters.

You can't please everyone. If I enter ten as 10.000, and someone enters ten thousand as 10.000, you cannot handle that without some knowledge of the culture of the input. Detect the culture somehow (browser, system setting - what is the use case? ASP? Internal app, or open to the world?), or provide an example of the expected formatting, and use the most lenient parser you can. Probably something like:

double d = Double.Parse("5,000.00", NumberStyles.Any, CultureInfo.InvariantCulture);

I don't know the ASP.NET side of the problem but .NET has a pretty powerful class: System.Globalization.CultureInfo. You can use the following code to parse a string containing a double value:

double d = double.Parse("100.20", CultureInfo.CurrentCulture);
//  -- OR --
double d = double.Parse("100.20", CultureInfo.CurrentUICulture);

If ASP.NET somehow (i.e. using HTTP Request headers) passes current user's CultureInfo to either CultureInfo.CurrentCulture or CultureInfo.CurrentUICulture, these will work fine.