So I'm going to try and comment on the classes, but do it in a way where I'm not simply rehashing discussions from the tier list threads.



Beguiler

Beguiler feels like a magical rogue with a focused spell list and no Sneak Attack. In terms of party role, they have extremely similar weapon and armor proficiencies, both have lots of skill points and huge skill lists to take on all sorts of non-combat roles and both can do trapfinding. Main difference is that Beguiler is going to need other people to finish the job.

I had always thought the class was kinda weak against mindless creatures, but it’s not as bad as I had thought prior to reading up on the class. Still, I’d want to pick something else if I was in a campaign predominantly based around, say, undead. Having a solid chunk of your spell list not be usable wouldn't be fun.

Not my thing (I'm more of a blasting man) but a solid class.



Dragon Shaman

One interesting thing I would like to note about Dragon Shaman is that it only existed for four months before we got Dragonfire Adept (PHB II was released May 2006, Dragon Magic was released September 2006). Given how similar their design goals were, it makes me want to speculate that they weren't happy overall with Dragon Shaman, which leads me to wonder what the design and testing process was for the classes in the book, and if Dragon Shaman got through but was later deemed "not good enough".

Power Aura makes me want to do a multiattack + sorcerer build.

I was originally going to complain about breath weapon doing 10d6 at level 20 being more flavour than usable, and would be really hard to optimize. However, while doing some background research for the thread, I found this old article, discussing high-level combat, and at least one of the people involved were involved in design and development of 3e. (here). Now, we can critique how the game was run and how the Balor was used all we want, but if you're playing the sort of game where a Wizard feels the best use of their time is using Quickened True Strike + Manyshot, doing 10d6 damage in an area probably isn't as bad an option as we think it is now.

Last observation, it's kinda weird that True Neutral characters cannot be Dragon Shamans. Makes sense given the alignments of the various MM dragons, but by the end of 3e, we had a bunch of Neutral dragons (Amethyst, Astral, Concordant, Ethereal, Lung, Mist, and Sand).



Duskblade

If I had a newbie, or even someone who wasn't deeply familiar with the rules who wanted to mix up spellcasting and swordfighting, I'd just give them a build with this class. There are stronger options but this gets the job done without undue complexity most gish builds have.

Twilight enchantment is weird to highlight. You can wear light and medium armor and up to heavy shields with no trouble but heavy armor has a minimum 30% penalty. You probably wouldn't want the mobility restriction of heavy armor with your build anyway, never mind a 20% ASF.



Knight

There seems to be a thought that heavier armor means more AC with the devs of 3e. However, the total of max dex bonus plus AC from armor class nearly always comes out to around 8 - you'll only really get higher than that with a race with a really high dex bonus, or if you're rocking some exotic heavy armor like Mountain Plate (at least, before we start book-delving).

It seems like the Knight was designed for tactical 2D combat with attempts to restrict enemy movement, with features like area control, making terrain difficult terrain, and so forth, while combat in 3e actually moves towards **** like flight, tactical teleportation, and so forth. Again, I'm wondering at what sort of playtesting happened during the development of 3e.

A lot of the discussion for the class is "Code of Conduct sucks and makes a weak class worse". That's true, but I look at it as WOTC wanting people to roleplay the whole Knights as Lawful with their Chivalric Code. If you're on-board for that, you want to adhere to the rules and the downsides of the Code of Conduct aren't really relevant.

I like the overall concept of the Knight, even though I think it needs a lot of work to mechanically support its intended use.