Consider an integrated switch like TPS1H100. Basically no support parts needed. Very robust.
You can make parts of one yourself, but it will take a fair number of components. To run an N-channel MOSFET at DC, you need at least some microamperes of supply above the input voltage -- because Vgs(on) is always gate-to-source, and the source pulls up to input voltage, so the gate voltage is above it, and so too your gate driver/controller.
Easiest way to generate this, is a little oscillator like a 7555, driving a charge pump (basically a half-wave voltage doubler) stacked on top of the input voltage.
Then you need a level shifter, to get the ground-referenced logic input translated up to the source-referenced voltage (which is near ground when off, and Vin when on). This is usually done with current sources (so the signal doesn't vary with voltage), preferably balanced current sources so that capacitance is canceled out. There need not be much gate drive (peak) current delivered -- the gate is a capacitor, and it switches only as fast as you can dis/charge that capacitor; but we aren't talking switching power supplies here, so the switching speed can be slower, and this also helps deal with capacitive loads by bringing them up slowly, and inductive loads by letting them down slowly.
As mentioned, the extra voltage is only required for NMOS. PMOS, the source is referenced to Vin, and Vgs(on) hangs "below", so you don't need an extra supply. You still need the level shifter though. NMOS is preferred in general, because it has about 2.5 times better performance than PMOS -- if you just want a switch, it doesn't matter much, but if you need a switch that's faster, lower power, and lower cost, the NMOS is preferred.
So this gives you a switch, but it has no protection whatsoever. The slightest short circuit and, kaboom! Transistors die in say 1/10,000 of a second. Fuses are no use: they blow in 1/100 of a second or thereabouts.
Current limiting can be provided by adding a sense resistor, and a transistor that pulls down the gate voltage when source current rises too high. Downside, this has a relatively high voltage drop (0.6V at the onset of limiting, with a single BJT sense circuit). Current limiting isn't all that useful anyway, because trying to switch, say, a 20V supply with 1A current limit and a short-circuited load, that's 20W continuous you're dissipating; tolerable with a TO-220 and heatsink, but that's a lot of bulk, and anything without a heatsink will quickly burn up at this power level.
So you might consider adding thermal limiting, but mind that this acts very slowly (maybe 1/10 of a second, at best), so you can't handle too much power dissipation -- this isn't a scalable solution. Say if you want to handle 60V and 30A with a single switch -- easily within the switching rating of a single transistor, but that's 1800W short circuit, enough to pop it in a 1/1000 of a second.
Alternately, you might detect that current limiting is active, and start a timer; when the timer runs out, a flip-flop is toggled off, keeping the transistor off until the input is turned back on again. But this has a number of limitations.
So the design of one of these things is actually quite involved, and you are getting an excellent deal from the, probably hundreds of transistors that are integrated into one of those protected switches! The main downside is that those integrated switches don't absorb much energy, and they aren't made very large; you can get 24V at 30A, but you aren't going to find 120V at 10A, or 400V at 30A, or more. (Or actually, much of anything above 80V I think? They're mostly for automotive purposes, so 30-60V ratings are super common. Or for USB port power, where 5-20V ratings are super common.)
So, if you need higher voltages or currents, or higher dissipation (more short-circuit capability, say to charge bigger capacitors?), you're pretty much on your own. There are a few controller chips out there, so you can use bigger external transistors, but I'm not aware of any that deliver very high performance (especially high voltages).
Tim