Via Larry Solum's excellent Legal Theory Blog, I found Paul Horwitz's 2000 article in the Osgood Hall Law Review entitled, "Law's Expression: The Promise and Perils of Judicial Opinion Writing in Canadian Constitutional Law".
Horwitz's argument is that the style of judical writing makes a difference for the quality of law, particularly constitutional law. The typical judicial "opinion" (note to PH: in Canada, they are referred to as "reasons for judgment") takes an omniscient and dogmatic tone, states the obvious and irrelvant at length and sets out lots of "tests" and "hurdles" that rarely do much of the work of deciding the case. Horwitz not only thinks that this style is boring, but that it is also bad for the law, and although he doesn't make much of a case for his position, I tend to agree.
Instead, Horwitz would like to see a style of "open-textured minimalism." The Pithlord likes the minimalism part, but to the extent I understand what "open-textured" means (Socrates meets Solon, I suppose), I doubt that most judges are really up to the task. Judges are successful lawyers who have avoided creating powerful enemies-- intelligent and hard-working, usually, but not prophets. A few of them -- like Oliver Wendell Holmes or Richard Posner -- have original minds, but even these people are only acceptable as judges to the extetnt they suppress their most original ideas when on the bench.
Horwitz doesn't care for the Oakes test, and presumably would decry the Delgamuukw decision in which Lamer goes on and on at Russian novel length setting out impractical and many-stage tests, while never deciding any issue actually between the parties. So far, the Pithlord can add little more than "Amen" and "Hallelujah".
The Pithlord gets crankier when Horwitz reveals what he thinks of as skookum judicializing. Horwitz is a big fan of the Secession Reference, in which the Court held that Quebec couldn't unilaterally separate, but that a "clear majority on a clear question" would trigger a duty to negotiate the terms of secession. By Lamer-era standards, the decision is a model of clarity and pith. And the underlying political tradeoff is defensible. However, it seems to me that this case shows a bit of a weakness in the Horwitz approach, since the style cannot hide the substantive trickery of the decision. Our Constitution has a detailed set of provisions for its own amendment. Referenda, whether clear or opaque, have no role in those provisions. Legally, the question the Court was asked in 1997 wasn't hard at all: Quebec couldn't secede (except through revolution) unless at least the federal Parliament and six other provinces agreed, and there is no legal requirement for those other entities to consider a Quebec referendum at all.
Whatever its stylistic merits, then, the Secession Reference was lawless. That strikes me as the bigger point.
Technical note: The University of Montreal website with Supreme Court of Canada decisions seems a bit wacky right now, so I haven't tried to hyperlink the decisions referred to. I may get around to it someday.