Clans, I agree that feeding someone mana could be useful, but why bother feeding someone mana if you can heal them and yourself to full life for 40 more? Then who cares what spells they can cast, you can just run through/over everything. Yoshi's suggestion makes Disable even more powerful, but with the "Rally" a more useful spell, swapping them isn't as big of a deal. I just can't see someone spending 80 mana to save 3 seconds of stuntime unless they're about to die. I also can't see one spell doing so many things (Rally, re: medics->marines) so here's a slightly modified bit of what I posted above:
L1: Medic summon: If you have Medics nearby, they are turned into Marines. A medic still spawns either way, though.
L2: Rally: Remove stuns/Negative effects from self and allies, +50 mana for nearby allies.
L3: Disable: 30 Second Disable, remove long term benefit spells, set to 0 Energy (Okay, the stun was probably a bit much. =p)
L4: Heal: Medic Full, Nearby allies to level +5 civs.
An all around powerful L2 and a situationally powerful L3 would definitely increase the balance of the medic. In response to Iceman, if it's an added effect that is offensive in nature but doesn't
remove a support spell, the medic isn't getting any less supportive. That's why I kept the prior spells both in play.
Clans I can agree most of what you said about my suggestions is at least a valid complaint, but these two I don't:
I would never touch this summoner with a ten foot pole. You want to add depth to him by just giving him another melee zerg summon? REALLY? This broodling idea takes the cake for your worst idea ever faz-. The fact is that even with just Lv1's the summoner is micro-intensive, fun, unique, and successful. Your solution to diversity is to take away the most popular build, when what you should be doing is giving the player better reasons and options to break away from the standard build.
You want to make every summoner play exactly the same game every time, and yet you remove the option of even a non-standard build. The L2/L3 upgrades are decent solutions, but that just gives the summoner NO extra options. He can't even go out of base reliant on L3, anymore. Just because it's fun and balanced with just an L1+mana spam doesn't mean it can't be improved upon. I am not opposed to combining the two solutions, however. L2 can give lings speed AND make Broodlings. That will ensure that the summoner levels up spells, anyway.
Also, you said the point of the LM is that it's slow and difficult to move around. Well, if it wasn't for abusing reaver splash and storm to teamkill, because of that weakness this character would be absolutely HORRIBLE in comparison to others. It is by far the slowest and has a tiny ranged attack, so it can't even hit under swarm. PARTIALLY removing the weakness of speed by giving it teleport allows it to leave cannons and play a useful role in a Temple Siege game before the 20 minute mark. That's my goal with the character, to make it similar to the others so it can be useful, rather than just ignored or avoided for the entire game. Nothing can face a late-game LM except a Medic, but early game it's toast if it moves into the map without a teammate. I think the teleport->L1 combo WILL be enough to make players level ground weapons, and the fact that Reavers are at L3 will make them less leveled. There's the balance. A godly Wall of Light + Reaver transform as L4 would increase that balance. That would be a LM that I would want to play, because you could take chances and do something besides kill spawns.
As for the Mech getting Vulture Speed at L3, that's just dumb. Then I'll be upgrading to my tank to be able to use my Vulture better...? I really haven't seen any solutions that even hint at balancing this character besides the Machine Shop or putting Vulture at L3. And I'd really prefer that the Vulture not be L3, because I die enough early game as a slow Goliath trying to be useful. Basically, if I get to Vulture and 7 damage upgrades without dying, I have a HUGE advantage. Mech is very weak to 2v1's early game and weak if you can stun a Vulture without mana. That's it, so that was my fix.
Archer: Well, I envisioned that this character be a damage one, but I disagree that having damage spells would be an issue. The L4 is mostly a building killer or someone who is stunlocked, so that's fine. The L3 I pictured being used if you were sniping someone, if someone was stunned, etc. Making it a machine-gun cast also means you can't target anything with it, and it loses that effect. The Archer tag-along game is just dull, so by removing the huge power of L1 and giving it some decent late-game PK options, that would have indirectly been fixed. I still think 27+3 is OP for the L1 mutas, I'm working on getting semi-decent as an Archer right now so I can prove it to people.
Moose's (UU's) solution as a precast for Mutant L3 should be perfect, so no need for another nerf there.
As for the Warrior/Volt comparison, I can see it. The difference is that the Warrior is more offensive and the Volt is defensive support, due to their speed and ease of escape. The Volt has more options if it goes huge mana and L4, simply by nature of the L3->L4 combo. Maybe simply increasing the Warrior's L2 damage plus his shields would be fair as an HP nerf. Either way, I think everyone so far has agreed the Warrior has too much damned life.
With the Drone->Guardian change, my suggestion for player owned detection is to allow Drone purchasing for
vespene. At 25 mana per drone, free hatchery, 25 for a sunken/spore, you have base detection and swarm protection available if you want it. Adding expensive (100 mana) nyduses as well would be a REALLY interesting feature that could potentially pan out. If your opponents are sitting in their base, that's one way to increase the effects of having map control.
On the whole, I'm surprised with what was agreed with and what hasn't been. I personally thought the most "duh" feature was the Machine Shop and that the Medic changes would be ridiculed, but the opposite has been true. I'm excited for the next version, even just with what has found mostly consensus.
None.