Unearthed Arcana was released in February 2004 by Wizards of the Coast, a subsidary of Hasbro, for the Dungeons and Dragons system. It is a 224 page hardcover rulebook with a $35 price tag.
The original Unearthed Arcana was published in 1985, and was compiled chiefly from material published in Dragon magazine and material from published modules. It introduced the barbarian class, the Cavalier class (which has never returned), and the first true prestige class: the thief-acrobat. Also new there were Weapon Specialization, cantrips, and first real epic class: the Hierophant Druid.
While the first Unearthed Arcana made rules variants official that were previously optional variants, the new UA is a published book of optional variants. This isn't the first time rule variants have been introduced into official D&D products. This was done prominently when 2nd Edition AD&D was released, for example. If you flip through the 2nd Ed. PH, you will see the occasional blue column, which is a presentation of a rule variant.
What is different is the scale of it. 224 pages of rule variants is a lot, and I mean a LOT.
When you think of a 224 page book of rule variants you automatically think, "Well, of course, its obviously a lot." But try reading through it, and considering each variant. And deciding which ones you will use/permit. Suddenly it hits you. You have to pick up the rulebook and start reading it to appreciate what I mean.
The range of potential complexity this rulebook offers makes it impossible to support in official releases, such as Dungeon magazine. I would anticipate, at most, the occasional mention and employment of a rules variant here and there.
Racial Paragons are basically "race as a class". When I first saw them I thought they might be like basic D&D, where elf and halfling were classes, but essentially what they do is grant additional racial abilities in return for class levels.
The "Bloodlines" are not that different, and are actually much worse in execution, with things like "lycanthrope bloodline". I dislike this particular case because I think of lycanthropy as either a specific curse, or a magical disease transmitted by a bite. A "part-lycanthrope" makes no sense to me. You end up using character levels to "pay" for a bloodline. The stronger the bloodline, the more levels you have to pay. You put specific character levels into the bloodline, its not like Level adjustments.
For the most part, I liked the class feature variants. Giving the classes more interesting and distinct abilities help to differentiate characters and make them feel unique. An example of this is Whirling Frenzy as a variant of barbarian rage. Instead of a bonus to Con, you basically gain a bonus to Dex. I think this is how it should have worked in the first place. The "bare sarks" or berserks generally worked by being so aggressive the opponent had difficulty hitting. Its also true that they could ignore pain thanks to adrenaline and endorphins though, so the Con bonus is appropriate as well.
Facing rules return, in spirit, from 1st edition. The biggest problem I see with this is arguments about designated facing vs intended facing. For the right group its great, but it could turn into a another thing to argue about. In cases of dragons particularly, I feel it is appropriate, with tail slaps and so forth.
Metamagic components are nice. There are only a few metamagic feats I have ever used or considered using: Still Spell, Silent Spell, and Heighten Spell. The first 2 are obvious, if you are chained in a "silent chamber" its that or nothing. A Still, Silent Knock is one use for those 2 feats. And Heighten is useful in order to make a spell harder to save against.
Anyway, of the magic system variants I like this one the best. It can add a tremendous amount to the flavor of the game, in terms of the materials themselves. It encourages the player to actually care about spell components, rather than regarding it as a book-keeping headache. And it encourages people (like me) who would never even consider using metamagic feats to pay attention. The upshot is that it makes components a fun part of the game. Needing a spellcraft roll to determine which components can do what for each spell seems reasonable to me.
It also imposes costs. The components aren't cheap, and can't just be tossed around with no thought. Finally, it grants the user more flexibility. He can decide when he chooses to cast the spell how he wants it to be. The additional options add to the fun of the game. Also, it allows the characters to heroically triumph in a flash of inspiration by making the component available as needed by the DM (although you wouldn't want to be too obvious about this).
All in all, it seems like an excellent variant to me, and is destined to show up in my campaigns quite soon.
There is a magic point system present, with several variants. This type of variant has been around as long as I can remember, and it didn't strike me as particularly noteworthy.
Item familiars provide a good rationale for the existence of intelligent magic items, in addition to the "spirit of a dying person" thing. Add the fact that it can be inherited, and you have a source of adventure seeds right here.
New here is also a systems for contacts, honor and reputation, which adds roleplaying depth.
The additional complexity offered extends to the basic mechanic of the system: the d20. One rule variant includes using a 3d6 (bell curve probability) instead of a d20. I guess that would make it the 3d6 system instead of the d20 system, eh?
There is some redundant material here. The horror rule variants are lifted from the d20 Cthulhu manual. The variant damage rules are taken primarily from Star Wars d20. Rules for Taint are taken from LO5R.
Action points are like Force points or Last Resort points (Alternity), or Karma (Shadowrun), etc. Traits are like background options, and for the most part they act as a net penalty. Flaws work like Flaws from Shadowrun and similar systems in other games. Flaws allow you to take an additional feat for each flaw selected. Nothing all that new there.
This is an appropriate point at which to make an observation.
The art quality of D&D books has been dropping. The quality of art in the new Dragonlance setting book actually shocked me. The art in UA is better, but only marginally so. Its starting to look like a trend at WotC, which clearly does not augur well.
The editing, typing, and layout are all up to the high standards WotC has, so it looks as though WotC is simply trying to cut costs. The cover art is excellent, if a tad odd.
For this review, production quality can be examined objectively, but the content is a different story. Do you like the rule variants? Do you not? Do you even like the idea of a book of rule variants?
I personally don't think its a bad idea, but then value considerations enter the picture. If these are just optional rules, then they can't be well supported, adding a conversion "cost" to the use of any official material. If one uses only 1/4 of the optional rules (which is FAR more than it sounds like), then did one get one's money's worth?
Did you get $35 worth of fun through these rules variants?
These are completely subjective questions.
I certainly admire the courage WotC showed in releasing this. It is definitely a risk, chiefly because the question many will ask themselves is: "Why should I pay for something I can/did create myself?"
I do think it succeeds at what it attempts, and adds a lot of possibilities to the campaign. There are a lot of ideas here that are worthwhile, many that are worth giving shot that might not otherwise occur to a DM or player. This is probably the Handbook's greatest use: as a source of ideas for spicing up a game that you feel might be getting a bit dull.