> I got a LOT of motivation out of this post

Wow, thanks, I was hoping it was more encouragement than crticism!

I've started quite a few games (even lenthier projects) and never got finished because I took too much care of details before I started adding actual content (in this case I wanted to do some playable and enjoyable maps ASAP).

Good for you! This is one of the most important things an indie developer can do to ensure their game is actually finished. And I love the idea of getting playable maps.

Here's a system that may be helpful: if you have a nice long "to do list", mark each entry as either red, yellow or green. Red entries are super critical 'can't proceed without them' things, like major bugs you've noticed, or very important features. Yellow entries are minor bugs (an object graphic blips when a level loads, or text that doesn't fit on a line prints out fully before being moved to the next line) that do not affect the ability of players or testers to play the game from start to finish. They include such ideas as 'more frames in walking animation' and 'level intro swishes'. Green entries are the types of things that even 'real' console type games get released with, that would be nice to fix in a perfect world, but which may only bother the creators of the game. I imagine these things are mostly aesthetic, and are very very minor. Once you have your list categorised and rearranged, be sure to finish red items before yellow, before green (unless you need a tiny thing to do - indies need motivation sometimes, since they don't get money!).

...I'm just getting used to pixeling, so a portrait might be some work for me.

Ah, well, you coulda fooled me  . I guess since I was the one who brought it up, I should be willing to do the work of an edit to show what I meant - if that's okay with you and setz.

I like your new trident guy. Might I suggest "Ink Guard" as a possible name? (oh dear, I guess that means I'm really getting into the enjoyment of a project if I start thinking of names for enemies... )

What would be useful would be a tool in which I can modify the pallette of a picture easily and compare the same picture with different pallettes, recommendations anyone?

Personally, I use GIMP for all my needs, because its a tool I'm used to using. But it is a little tricky to optimise it for pixelling. How about a left field suggestion from the world of ROM hacking? I have experience with tile viewing programs that display binary art as graphics in different formats (SNES, GBA, Mega Drive, &c.) and can import custom palettes, which you switch between while the art is displayed. Palettes can be built manually, or easily put together with an 'eyedropper' tool I use that a friend of mine made. The sprite-editing abilities are somewhat minimal, but it's useful for trial and error, maybe. If this kind of thing appeals to you, I can provide more detailed information.

I'm still thinking about adding colors to the pallette, namely some greens and some shades of the red and violet.

There are palette wizards here and on PJ who might be able to offer suggestions. There is also, I believe, some consensus on the practise of using pre-existing palettes for inspiration as long as credit is given. Searching for topics that mention palettes usually proves informative.

I've been trying to learn and adapt painters' palette techniques to pixelling, and here are some links that might be useful from the great art blog of Dinotopia creator James Gurney:
Post about the colour scheme designer website.
Post about an online gamut masking tool.
A seven-post mini series with easy introductions to colour theory and the colour wheel, with info on custom palettes.

It might also be useful to keep in mind that most games of, say, the NES or SNES era used a different palette for each world or level, so it may not be necessary to create a giant meta-palette, but rather a collection of smaller palettes that are complimentary to themselves.

I'm thinking about some tiny and big (shadows of) fishes, coral rifs or sunken ships, maybe all just in one dark color that gets faded against the background, but I don't really know where to start...

Those are all great ideas, and the more you wildly free associate, the more chances you have of coming up with even more cool ideas. I love the look of multiple silhouette background layers (not just because silhouettes are easier to make...) so I heartily encourage you to try out all your ideas with different colours for different layers, and play with the player's expectations of mass (e.g. if the level seems to have a distant horizon, a sudden, nearby, hulking shipwreck can create a surprising dissonance that most people interpret as a pleasant thrill).

Hmm to be honest I specifically designed the font for this game to give it a retro-ish look and I like it so far (a few letters and numbers aside).

The font is quite attractive, actually (though I think the "S" should have the smaller curve at the top rather than the bottom) so if you think it would look good at the resolution the game is homaging, then there's no problem, though my guess is the lines are a bit thin. I don't know the pixel dimensions of your font, but is it possible to bulk up the lines a bit, perhaps with a text effect like a drop shadow or coloured border?

I'm thinking about changing the size of the text-box to contain only two lines but maybe that's not enough for the text to be fluently readable...

Two or three lines is pretty standard. For me, the smallest amount of text you can show on the screen at a time will increase the readability, as well as making the writing seem punchier. Here's an example textbox from one of my all-time favourites, Phantasy Star 4, and another textbox from Zelda, Link to the Past, and a third textbox from Zelda, Minish Cap, that show some really good proportions (though, of course, the GBA one has to squeeze a lot into a small package!). Playing games like this helped to convince me that menus (which text boxes are a subset of, I'd argue) should always float like a popup, so that the illusion is maintained that the player still knows what is "under" them (they become less "intrusive" that way).

But that's one of the things I want to test on a map that is fun to play for me and some testers...

Great idea! I volunteer, selfishly, to be one of those testers!

The pearls (I'm now gonna call them like that  ) are a motivational thing.

Heh, I wasn't trying to be persuasive with subtle hint words like "marine"...honest! 

(which will be changed to +X when you collect more pearls in a row btw.)

Ah, well, that's a...whole new kettle of fish! That sounds like an excellent idea, and watching the numbers build up in chains sounds like great fun. Please make the lag time before the chain is broken be somewhat forgiving so that this enjoyment can be experienced to the full, since it sounds like it could be a major source of fun in your game design!

I'm going to do this, plus I wanted to add a little screen shaking when they stomp near you.

Cool, though I would think to use shaking sparingly, as the type of thing saved for boss battles, etc.

I'm thinking about a rotating star fish.

Might I also suggest a subtype of your spike ball that is on a chain, such as this one in Sonic & Knuckles?


Well, I've gone on a mile again - hope it helps!