No layer at all is impossible with aluminium. In the best vacuum producible by humans, the oxide layers forms within few minutes. This explains why the electron work function is so imprecisely known. Things look better with gold or iridium.
Specialized companies prefer to first remove much of the existing layer, which tends to be dirty.
Chlorine is extremely deleterious to aluminium (that is, to its oxide layer). Anyway, the process must bring oxygen atoms to aluminium, which NaCl doesn't: think at the ions present in the solution.
Other acids are commonly used, not only sulphuric. They give different colours and corrosion properties.
A good oxide layer is densified in water vapour after anodizing and cleaning. Anodizing alone leaves small pores in the oxide.
Round cathode: this mustn't be very important. Edges would emit more current there, but it spreads over distance. Flat electrodes and sponges retaining the electrolyte are used when the part is bigger than the bath.
A uniform thickness is obtained in electrolytic processes through proper current density, temperature, bath stirring and, much more astute, by reversing regularly the voltage. That is, thickening processes diverge on spikes that concentrate the current, while thinning processes polish the part by acting more on the spikes; this is amplified by current density but lessened by heat. Now, if the polarity is often reversed, and the reversed period has a different duration and a different current, you can obtain both a net deposition and a net polishing effect. Re-genial, isn't it?
Anodizing is much more easily uniform than an added metal layer, because its thickness defines locally a minimum voltage drop that allows current to flow (more or less a breakdown voltage). As this voltage is nearly uniform on the part, you get a uniform thickness, especially if the electrolyte drops little voltage.
This is how very thin and nevertheless uniform insulation layers are obtained, to produce electrolytic capacitors for low voltages and with a big capacitance. Aluminium is still used, tantalum as well, and more recently niobium. The anodized electrode can even be sintered to get cheaply a huge area, yet carry a uniform oxide layer.
Pure aluminium exists only in books. I had 99.9999+ when making electronic chips, but the purest available for mechanical engineering is 99.8%, whose corrosion and mechanical properties differ markedly from 99.5% and 99%.
These alloys (named AA1xxx) are excellent for anodizing, shortly followed by AlMgSi (AA6xxx) and AlMg (AA5xxx, good protection but lackluster grey). AlZn (AA7xxx) gets some acceptable resistance against corrosion only through anodizing but is dark grey then, unless you colour the oxide. AlCu (AA2xxx) is bad at anodizing. Heat treatment influences much.
The common claim "aluminium resists corrosion" is plain nonsense. Some alloys do (more or less the ones that improve through anodizing...), others don't at all (after putting AlZn5Cu1 one day in a humid soil, I removed 1mm oxidation crust with the nails) - just as for steel.
Tin: food is corrosive and traces of pollutants in food can poison consumers (lead) or induce allergies (nickel) so the choice is narrow, more so because lasting metals are expensive (gold, palladium...). Plastics now, yes.