"The starting-point ... is that the giving of reasons is a normal (albeit not universal) incident of the judicial process. That is because 'the duty is a function of due process, and therefore of justice'. Fairness requires that the parties should know why they have won or lost. A requirement to give reasons is likely to produce a more soundly based, rational judgment. The requirement also furthers judicial accountability.
Where there is a right of appeal, the reasons must be sufficient to give effect to that right. The basis for the decision must be apparent, as otherwise the losing party cannot know whether there has been a mistake of law or of fact. Just what that will involve depends upon the nature of the case. Some cases turn upon a simple contest of credibility between two witnesses. Others involve detailed and complex factual and legal issues requiring close reasoning and analysis.
Reasons need not be lengthy and elaborate; nor do they need to refer to all the evidence led in the proceedings. However, relevant evidence should be referred to (albeit not necessarily in detail) and, where there is conflicting evidence of significance to the outcome, both sets of evidence should be referred to. Where one set of significant evidence is preferred over another, the trial judge should set out findings sufficient to explain why. Similarly, where a dispute involves a form of 'intellectual exchange, with reasons and analysis advanced on either side', the judge 'must enter into the issues canvassed before him and explain why he or she prefers one case over the other'.
Inadequacy of reasons does not necessarily amount to an appealable error. An appeal court will only intervene when no reasons have been given in circumstances in which they were required, or when the inadequacy is such as to give rise to a miscarriage of justice. Nor does an appealable error arising from inadequate reasons necessarily result in a new trial. The appeal court is entitled to consider the matter and, if it can do so (where, for example, only one conclusion is reasonably open on the available evidence), it may itself decide the matter." (Citations omitted)