There are many types of barcode, in a range of sizes and employing various methods of error checking. Some barcodes can encode letters as well as numbers, e.g. Code 3 of 9, Extended code 3 of 9. Code 93 can encode punctuation characters too. Most barcodes consist of vertical black and white bars (1dimensional), but some are an array of black and white tiles (2dimensional). '2D' barcodes are easy to distinguish (see DataMatrix 2D and PDF417 2D), but to identify 1D barcodes, look out for these features:

Vertical bars of uniform width, e.g. BPO four state code, POSTNET and USPS Intelligent mail 4.
Vertical bars – both black and white – of several different widths e.g. CODABAR, Code 3of9, UPC.
Some barcodes encode characters in pairs: one in black bars, a second in the white bars separating the black bars of the first. E.g. Interleaved 2 of 5. The bars may be placed between horizontal 'bearer bars', ensuring that there is a distinctive border around the bars encoding the data: see ANA interleaved 2 of 5.
Barcodes may include 'guard bars' to signal the start and end of a barcode, e.g. UPC-E. Some include a guard element in the middle too: see UPC-8. A barcode may include supplementary data in extra 'add-on' bars, making them wider: see UPC-13.
Many barcodes are built up of vertical bands. There is usually a top band and a bottom band, plus a number of mid-section bands: the more mid-bands you insert in a particular barcode, the taller it will be – e.g. Telepen, Code 128, MSI.  Numbers at the foot of a barcode enable Optical Character Recognition. These are contained in the bottom band, but they are optional and may be omitted: see UPC barcodes.