Before I get to the old saying 'hard cases make bad law', I must start with the apocryphal story of the indulgent jury or magistrate who, not wanting to ruin the life of a poor young defendant who has made a silly mistake, returns the verdict "Not guilty" - and then adds sternly "but don't do it again".
This case (Crinion & Anor v IG Markets Ltd  EWCA Civ 587) is the closest to that story that I have seen in real life. A first instance judge (HHJ Simon Brown QC) heard the evidence in a trial and asked the two sides' barristers to produce written closing arguments. So far, so good. He then took the claimant's barrister's argument, fiddled around with it a little, and made that his judgment. It seems that apart from a cut and paste job from somewhere else, 94% of the words of the judgment came from the claimant's written argument. Indeed, the Court of Appeal records: "The Appellants point out that in the "properties" file in the Word version of the judgment the "author" is shown as "SChirnside" [i.e. the Claimant's barrister]." A jury point perhaps, but none the worse for that.
The defendants appealed, essentially on the basis that that was no way for a judge to write a judgment and so the judgment should be set aside. The Court of Appeal, however, decided that the little bit of fiddling done by the judge - the extra 6% - showed that in fact he had done his job of weighing up both parties' arguments and deciding fairly between them.
So it was a 'not guilty' for the judge. But the Court of Appeal added some stern 'don't do it again's. Underhill LJ said "In my opinion it was indeed thoroughly bad practice for the Judge to construct his judgment in the way that he did" and "the judgment was defective, even seriously so", while Sir Stephen Sedley added "I hope that a judgment like the one now before us will not be encountered again". Words such as "wrong" and "unacceptable" were also used. Strong stuff indeed.
At first sight, that all seems very sensible. However, there is something unsatisfactory about this outcome, at least for the legal purist. The defendant in the apocryphal story was, strictly speaking, guilty. If the judgment in this case was, strictly speaking, "defective" or "unacceptable" then surely it should not have been allowed to stand? On the other hand, if the 6% of the judgment added by the judge was, strictly speaking, enough to make it a good and proper judgment, then why shouldn't that be enough in other cases too, saving a huge amount of judicial time and effort in the future?
Here is where the phrase 'hard cases make bad law' comes in. You could say that this was a hard case, at least in the Court of Appeal: Underhill LJ's decision was reached, as he said, "not without some hesitation". But really the problem was that this was an easy case: the judge at first instance took the view that one side was clearly right and the other side clearly wrong; even on appeal Underhill LJ said that he had seen "nothing ... that leads me to think that the Appellants ever had a real defence to the Respondent's claims".
In an easy case, where the correct answer is obvious, there is a great temptation to take a shortcut to reach the right outcome. Quite right too, you might think: getting the right outcome is what justice is all about. The net result, however, is that the position in English law seems to be that adding 6% of the words to a judgment was good enough for HHJ Simon Brown QC, sitting in the Mercantile Court in Birmingham in 2012, but not good enough for any other judge in England and Wales, ever again. That might not exactly be bad law, but it is at least a little odd.