Nobody talks about dado joinery the way they talk about dovetails. There are no YouTube channels dedicated to people cutting perfect dadoes. Nobody is mounting their best rabbet on the wall.
And yet: the bookcase that's still square after fifteen years got that way because of dadoes. The cabinet whose drawer bottoms haven't split, the frame-and-panel door that survived three BC winters without cracking — all of that is the housing joint family doing its job without any fanfare.
Here's what you need to know.
Three names, one geometry
A dado is a channel cut across the grain. It runs from edge to edge (or stops short of one, more on that later) and receives a mating piece at right angles. Shelf into bookcase side. That's the model.
A groove is the same channel, but along the grain. It captures a panel. Drawer bottom in the drawer sides. Cabinet back in the case sides. The geometry is identical; only the orientation changes relative to the wood.
A rabbet is a two-surface cut at the edge or end of a piece. One shoulder, one cheek. Open on one side instead of two. Used for back panel reveals, drawer fronts, applied moldings, and overlay joints.
The dado and groove differ only in which direction they run. The rabbet differs in that it's open on the outside face rather than enclosed. One family, three applications.
Two numbers govern everything
Width equals the actual thickness of the mating piece. Not the nominal dimension. Not the stated spec on the cut list. The piece in your hand, measured with a caliper. Sheet goods vary. Dimensioned lumber varies. Measure it, then sneak up on the cut. Rout slightly narrow, test the fit, widen until the piece seats under moderate hand pressure. This applies whether you're using a dado stack, a router, or a hand router plane.
Depth is one-third of the stock thickness. On ¾" case sides, that's ¼". Hard rule: never deeper than half the stock. Go past half and you've removed too much from the piece the joint is cut into. The shelf seated in that dado is now bearing against a compromised wall. The shelf will be fine. The case side won't be.
Where movement makes it complicated
Dadoes across a case side are low risk. The shelf and the case side run in the same grain direction. They move together.
Grooves for panel capture are the critical case, and this is where people break things.
A solid wood panel in a grooved frame runs cross-grain to the frame members. When humidity climbs, that panel expands across its width. If the groove is too shallow or the panel is glued in, that expansion goes into stress. Panels crack. Frame joints blow open. The fix is simple: the panel floats. It isn't glued. The groove is deep enough to retain the panel at its minimum dimension while leaving clearance for expansion at its maximum. For most ¼" back panels, a ⅜" deep groove with ⅛" float on each edge is the working spec.
MDF and veneered sheet goods are the exception. A stable panel can be glued because it doesn't move. That logic does not transfer back to solid wood. Don't let it.
Drawer bottoms in solid wood run cross-grain to the drawer sides. The front edge is captured. The back edge is screwed, not glued, through the bottom into the drawer back — this is what allows the bottom to move seasonally without splitting. Minimum ⅛" float in a humid climate. If you're in the BC Interior, give it more.
Stopped versus through
A through dado runs the full width of the case side and shows on the front edge. Fine for painted work or enclosed cases. For a bookcase where the front edge is visible, you want a stopped dado: it terminates before the front edge, the shelf gets a small notch cut from the front corner, and the joint disappears. The router leaves a curved terminus that won't close cleanly — you chisel it square.
More work. Right answer for visible casework.
Where this family belongs — and where it doesn't
Fixed shelves in case sides. Cabinet back panels. Frame-and-panel doors. Drawer bottoms. Drawer front construction with lock rabbets for painted work. These are all housing joint applications.
Chair rail connections and table aprons are not. Those joints load in multiple racking planes and need the full mechanical interlock of a mortise and tenon. A dado on a chair leg rail will fail under normal use. This is not a "can you get away with it" question — it's a physics question, and the answer is no.
The housing joint family handles the load it's designed for. It doesn't try to be the dovetail. It doesn't try to be the mortise and tenon. That's why, when you use it in the right place, it just works.