Another user asked for a workaround for sqrt in the mentors forum a little while ago, and I directed him to the Babylonian method of approximating square roots. <!-- let's calculate sqrt(a^2 + b^2) shall we? --> <input type="hidden" disabled="true" name="attr_s" value="(@{a} * @{a} + @{b} * @{b})" />
<!-- Choose a different initial guess if you like, just make sure the result will be positive -->
<!-- The better the initial guess is, the more accurate the result will be -->
<input type="hidden" disabled="true" name="attr_x0" value="@{s} / 2" />
<!-- Add additional iterations if you need additional accuracy -->
<input type="hidden" disabled="true" name="attr_x1" value="((@{x0} + @{s} / @{x0}) / 2)" />
<input type="hidden" disabled="true" name="attr_x2" value="((@{x1} + @{s} / @{x1}) / 2)" />
<input type="hidden" disabled="true" name="attr_x3" value="((@{x2} + @{s} / @{x2}) / 2)" />
<input type="hidden" disabled="true" name="attr_x4" value="((@{x3} + @{s} / @{x3}) / 2)" />
<input type="hidden" disabled="true" name="attr_x5" value="((@{x4} + @{s} / @{x4}) / 2)" />
<input type="hidden" disabled="true" name="attr_x6" value="((@{x5} + @{s} / @{x5}) / 2)" />
<input type="number" disabled="true" name="attr_result" value="@{x6}" /> This method works by convergence of the arithmetic means of over- and under-estimates of your target number. For an example, look at sqrt(100) = 10: s = 100 x0 = s / 2 = 50 x1 = (50 + 100 / 50) / 2 = (50 + 2) / 2 = 26 x2 = (26 + 100 / 26) / 2 = (26 + 3.846) / 2 = 14.923 x3 = (14.923 + 100 / 14.923) / 2 = 10.812 x4 = (10.812 + 100 / 10.812) / 2 = 10.030 x5 = (10.030 + 100 / 10.030) / 2 = 10.00004 x6 = (10.00004 + 100 / 10.00004) / 2 = 10.000000000107445793143744907022 The worse your initial estimate is, the more iterations you need for a good approximation. Your initial guess (x0) can be literally any positive number, and it doesn't need to be based on s like mine (s / 2). The final value will always be an overestimate, because an arithmetic mean is always going to be greater than a geometric mean.
