I thought the article was a fairly honest appraisal of the ritual system, whether or not we approve of its design goals or execution.
Personally, my preference is to conceptually treat all magic as a ritual (slow, expensive, setting the baseline for what we think of as "effective", largely universal) and from there let classes or creatures with magical connections explore, for given subsets of magic, different facets of cheap, fast, and good (choose 2). For example, the first fireball ever cast wasn't necessarily done by a guy facing a few orcs, but perhaps the result of an involved process designed to break the grinding stalemate of an ancient siege. Although later refined into the elegant tool loved by arcane casters, there are still plenty of people around who want big fires, and don't need them within the next 6 seconds.
The absolute requirement for (tactical) combat magic is that fast casting, so the classes we traditionally think of as magical utilize an edifice (arcane training, divine favor, magical pacts) which achieve this. Perhaps anyone could raise the dead in principle, but the deities maintain the expressway for souls, and make sure they are all wearing their seatbelts. And certainly, "non-magical" characters will need to spend character resources in order to dabble. I like the idea of a rogue being able to cast a ritual which opens locked doors, but written in such a way that any rogue in a situation where he has to will feel he's already way beyond plan A. The wizard might be able to match a trained rogue's lock opening abilities by "upgrading" the ritual, but the opportunity cost to do so should be quite significant, and that usually means the wizard doesn't get there until higher levels anyway.
In my current game, which uses a homebrew system, the clerics can petition their deity for any ritual (i.e. miracles) but at rather significant personal cost. If they already know the ritual they can even do it quickly and waive larger component costs. Wizards, on the other hand, have a limited ability to perform the bulk of a ritual they know beforehand, and release it in an instant if desired. Druids use rituals faster on cheaper depending on the environment, etc. These are less common modes of accessing magic, however, and normal combat or utility spells are empowered rituals. I just didn't bother writing the "ritual version" for many since using them in that manner would be very rare.
Finally, I think 3.5 would have benefited (both in theme and balance) if clerics and druid could not prepare from such a large selection of spells every day, but had treated them more like 4e rituals. For example, a ritual similar to Find the Path would still be a potent ability if it required an hour to cast and had a component cost, but if a specialized cleric was the only one capable of casting it over a few rounds, he could use it to achieve qualitatively different outcomes. Of the homebrew designs I made for 3.5, my favorite is a divine class that chooses 3 domains, where the spells in those domains make up the entirety of its spell list. (Although primarily a warrior, it is a spontaneous caster with a slow slots/day progression.) Those restrictions make each one quite unique from deity to deity, and treating some 3.5 spells more like 4e rituals would have achieved that even more. I hope it will not surprise anyone that I'd like 5e to go in this direction, at least with divine casters.
).
