Well dual wield will work only with smaller weapons. Even if I just thought about a perks called "Big Hands" that let's you dual wield with small machineguns rifles
I understand what you mean, not having a skill that can make a Class particularly powerful, however remember that I can also balance that with the items. If dual wield works only with pistols for example, they could do less damage in general than a single combat rifle or laser phasor or whatever 2 handed weapon I'll design. But yes having Dexterity influence other stuff beside DW is a good idea and I plan to follow it.
Well, again, the issue is that non-DWers would have, essentially, 5 different ways to improve their attack ability (the 4 combat skills plus the base advantage from dexterity) while DWers would have 6 (the same 5 different ways as non-DWers plus the added advantage for DWing provided by dexterity; essentially double-dipping from dexterity). You can always balance for a single point in the game regardless of this, but what this implies is that DWing scales faster than non-DWers; that in turn will usually mean the either DW is very weak early or is too powerful later. This is not necessarily a problem (you might actually want that to be how DW works) but it does change the difficulty curve of the game depending on player choices.
It's true you can use item scaling to provide another form of scaling which can compensate (if 2hd weapon damage scaled much better than pistol/one handed damage). The main issue with that approach is that it very tightly couples the 2hd/1hd item scaling by making all 1hd weapons scale as if they would be DW'd (this doesn't matter so much if you assume everyone is either 2hd or DW; it matters a lot if you expect there to be some benefit to 1hd non-DWers). If skills were balanced in their scaling without relying on item scaling, it decouples the design space (so instead of having to balance skill + weapon vs skill + weapon, you could balance skill vs skill and weapon vs weapon, greatly reducing the total number of combinations). Decoupling the design space between skills and items also makes things like the 'SMG DWing' talents much easier to balance, too, since you can balance SMGs without having to choose to balance as a DW or non-DW weapon.
Again, my suggestion is to have dexterity do something to improve something that DWers have to give up to DW (another way to accomplish this would be for 2hd users to get double the accuracy bonus from dexterity compared to DWers). Without that 'trade off' inherently within dexterity, it's going to be difficult to make dexterity not a no-brainer stat (either too good, too poor, or only good if DW and never good otherwise).
Basically the idea is that dexterity (like all other stats, ideally) should improve the same number of aspects of a character regardless of the character (the actual aspects can, and probably should, be different between character builds). Right now what it looks like is:
DW: Dexterity improves accuracy, DW effectiveness (2 different ways)
Non-DW: Dexterity improves accuracy (1 way)
So non-DW either needs something more or DW should lose something. 'Something more' can be more of the same (improves accuracy + improves accuracy (2hd)) or something unique (improves accuracy + improves initiative (2hd) or improves accuracy + improves shields (are there shields?) or improves accuracy + improves items assuming items aren't usable in DW). Fundamentally it can be really anything which helps non-DWers , though thematically it seems like the justification should be something your offhand helps with (psionic power ability when not DWing is mechanically viable if probably not thematically suitable). Making the secondary effect be related to your direct attack also decouples dexterity from everything but direct attack balancing, which again makes balancing easier (note improves item usage or shield usage or improved initiative does not follow this guideline; decoupling makes balance easier but it can also make the mechanic too straightforward).
Another way to look at it is that there's way to make DWing better, but there isn't a way to make 2hd better. You could e.g. have strength be the '2hd' stat (strength decreases delay on 2hd weapons or something) but then you run into the same issue as the weapon skills - no reason (or relatively little reason) to put points anywhere outside of your specialty. Again, the key observation is whether or not a stat controls a similar number of things between different builds/classes.
If a stat governs different number of things for one build versus another, it makes it hard to make that stat relevant for both builds. This is not necessarily a problem, it's just that it removes a choice from the user (instead of choosing a build, then choosing a stat, choosing a build is effectively choosing a stat). Also, more technically, while I refer to the 'number of things' as an intuitive quantitative measure, somethings do count for more than other things; this doesn't break the analysis, it's just that some items are effectively multiple things (e.g. doing damage + doing damage + doing damage + noncombat check).
Lastly I will mention that I focus on scaling and balance not because they are the most important things, but because they can be more easily quantified and measured. Having a breakable mechanical system is not necessarily a bad thing (trying to break game mechanics is often part of the fun); however the worst result is probably if choices turn out to not really be choices because one is just better than another.
Appendix: Things that might make a good secondary effects to dexterity for non-DW (alternatives, not altogether...)
- Accuracy with 2hd weapons (if non-2hd/DW is viable, requires some other compensation, e.g. this could become accuracy while non-DW instead of accuracy while 2hd)
- Speed with 2hd weapons (same as above)
- Evasion bonus when not DWing/nothing in offhand/using 2hd - note tanking stats can be a weak choice given the aggro mechanic since there's builds outside of DWer or tank
- Effectiveness with shield (if there are shields) - this choice requires some benefit for nonDW/nonShield however; see note about evasion
- Effectiveness with items (if DWing prevents item use) * note that this is still a weaker effect since the DWer could swap items into offhand presumably at the cost of some time, but it's still a bigger benefit to someone who doesn't have to swap
- Effectiveness with burst fire mode if not DWing (this and next are again somewhat weaker options since non-DW doesn't imply burst fire specialization)
- Effectiveness with aim fire mode if not DWing (as above; note though _together_ it works pretty well)
- Better crit damage while not DWing
- Psionic effectiveness/speed with free offhand (sort of a somatic components idea) - would require added benefit for 2hd or shield types, or would apply to 2hd as well (doesn't quite make as much sense there unless DWing takes more concentration or something, though even then it'd be unclear how _dexterity_ helps when both hands are full; maybe makes you really good at moving hand off weapon and back on?)
Initiative is probably bad since the player could keep a 2hd weapon to win initiative and then switch to DWing (exact benefit depends on how costly swaps are). If swaps are pretty expensive initiative might be viable. Damage effects feel thematically inappropriate but mechanically could probably be made to work.