I should very much like to add some things here during the next day (just in a hurry). Though I generally do not contradict to cr0sh's sound opinions, I will in this case:
I'm no sage...
- I have hardly ever used a resistor outside the E3 series, this is around what you already have on your pick list. Buy 100 of each of them. That might sound ridiculous, but they will cost the same as 50, and 20 are definitely not enough.
Part of this does depend on what you are planning on doing; if you don't know what you plan to do, then buying a wide range might be a good deal (though it is expensive; I did it and spent quite the amount to outfit my resistor kit - 100 pcs each of E-24 adds up!) - but I wanted to cover my bases. A smaller E-3 kit is fine if all you plan on doing is general purpose digital and LED type work, but it likely won't cover it if you plan on say, building custom ultrasonic ranging devices, or RF, or amplifiers, or oscillators, etc.
- Don't buy any 78xx there are fine replacements, and you dont need 100 but 5 pcs. will do for the start
Well - if you can find a bunch of 7805s for 10 cents a piece, it might be a great bargain! There are better options, true, but the 780x series are workhorses as long as you keep the voltage in/out diff low (they aren't very efficient when the difference is large).
I shall come out with a shopping list soon (opamp, ...) You might form some ideas about LEDs in the meantime yourself (1.8mm? 3mm? 5mm? RGB? Dual color? with internal resitors?)
I still say stay basic with the LEDs - a standard 3 or 5 mm r/g/y set, perhaps 8-16 of each color. Maybe a few dual color if you think they might be handy. RGB only if that is something you want to play with or have a project in mind for such a device. Stay away from ones with internal resistors unless you know you need that for a project. For most experimentation, if you don't care about color, cheapo reds will do fine. Leave the fancier ones for when you know what your project(s) will ultimately be about...
First investments feel expensive... How much $ do you plan to spend? $100?, $200? $300? If even more waits in your purse, than please. please think of purchasing an oscilloscop - it is an eye-opener, it will give you the best of information what is happening, and it will save you hundreds of hours of frustrating debugging.
Only if you know how to choose the right one and use one; otherwise it can be an exercise in frustration (and if you pick the wrong one, you might end up buying more than you ever need). An o-scope is one of those things that you should possibly wait for buying until you need it, unless you know you'll be doing projects ahead of time where you'll know one might come in handy. You don't want to buy an expensive paperweight you won't use.
What tools do you have already?
I would say a good quality DMM with LC ranges is a better thing to have on the bench; even a cheap DMM is better than nothing - next would be a logic tester. You might want to invest in a small analog meter (some things can be tested better with an analog meter than a digital one). For certain things, a function generator and a frequency counter are nice - but not necessary (typically, you'll want those when you buy an o-scope anyhow).
A better investment might be a good dual output (or triple if you can swing it) adjustable regulated bench power supply...perhaps 0-15 VDC and 0-5 amps...
