Let's start with something "easy". I found some tiny but interesting-looking crystals in a piece of coarse-grained granite one day.  Luckily, there were one or two that were well-formed enough for me to make this drawing (using a microscope). They appear to be tetragonal system... or is it orthorhombic? Note from the drawing that the crystal's cross section would not be square - it would be rectangular...

 Let's skip this, uh, "evidence" for a moment. Back to the things we are sure of: they're brown, subtranslucent to opaque, have a vitreous to resinous luster, a nearly colorless streak, and a hardness greater than 5 1/2. A couple of the broken crystals appeared to have cleaved along the prism faces, in two directions. The sum of this evidence suggests zircon. Perhaps. Some of the crystals do appear to have the requisite square cross-section, but then there are others that don't. What is going on here?  

I'll now answer the mystery for you:  All this time I've known, with near 100% certainty, that the crystals were zircon.  

Distorted or atypical crystals that don't seem to "fit" their crystal system are actually not uncommon in nature.  There are "wire" pyrites, pseudocubic calcites, and many more.   Pyrite is of the isometric system, leading us to the false expectation that pyrite must always be cubic-- or at least equant (e.g., regular-dodecahedral).  

Zircon is of the tetragonal system, leading us to the equally false expectation that zircon crystals must always exhibit a square cross section.  Usually, they do.  It's just that sometimes one encounters a rectangular one instead.  If the crystals experienced uneven pressures during their growth or plastic deformation afterward, you might not even see rectangular ones.  The cross-sections could present uneven sides, non-perpendicular axes, and so forth.

Let's suppose you're collecting in a place such as Franklin, New Jersey, where there are a couple of hundred mineral species. Of those couple hundred, you will probably encounter at least 50, especially if you're a micromounter. Of those 50 minerals or so, there will be plenty of groups of 4 or 5 (or 10) that look exactly alike. Worse, there is plenty of overlap between the groups. Massive gahnite is often confused with diopside here (chemical tests can distinguish them without too much difficulty). It's a bit frustrating when you don't have access to a quantitative analysis lab or expensive test equipment.   There are companies that offer instrumental analysis services for "unknown" mineral specimens, but this can get expensive when one has a whole shoebox full of unknowns.

That brown to black pyroxene mineral that pops up so often at Franklin, Sterling Hill, and elsewhere: what is it? The clinopyroxenes are fairly common here, but the problem is that they all look pretty much alike. The local names "schefferite" and "zinc-schefferite" refer to something in the clinopyroxene group which appears to be a zinc / manganese-bearing form of aegirine, diopside, augite or hedenbergite. The pyroxene-group minerals are well nigh impossible to distinguish from each other by mere sight, at least at Franklin-Sterling Hill (a clue, however, is assemblages;  if yours is the same as a piece that has been analyzed, well...)  You could just label your unknown pyroxene as a "pyroxene" or "pyroxene-group mineral" if you're not sure. When I don't know exactly which pyroxene it is, it usually gets labeled "diopside" until I find out otherwise.

How can you be sure what you've found? In most cases you can't, unless you get one of the "easy" ones like fluorite, calcite, sphalerite, or willemite.  Even franklinite has thrown sight-ID people;  a friend of mine collected a specimen that was dull green (like gahnite) to black (franklinite), but turned out to be neither.  It proved to be magnetite, a fact that lay unknown until he decided to put a magnet to it several months later.

Just remember that locality information is much more important than specimen I.D.  Feel free to put a label on it that says "UNIDENTIFIED", as long as you make sure to specify the place you found it (and the date, preferably). That's good enough for me.  Mineral identity can be established with tests; locality can't be recovered if you failed to write it down.  Trust me, your memory will fail you on that.
It is useful to do some basic tests to narrow the identity of a mineral. For silicate-group minerals, though, streak and hardness tests are all but useless. (Actually, that's a bit of an exaggeration; they do have marginal usefulness, even with silicates). As for crystal form... What crystals??