Most of the universe fills with magnetized plasma. When twisted or sheared, the magnetic field may reconnect rapidly, converting magnetic energy into heat and kinetic energy. Because these phenomena often occur in environments of large electric conductivity, the energy conversion is usually confined to a small local region [an X-type neutral point, a current sheet (CS), or a quasi-separatix layer]. The catastrophe theory of the solar eruption suggests the formation of a CS following the onset of the eruption due to the catastrophic loss of equilibrium in the system. The CS is suggested to be long because the time-scale of diffusion is long compared to that of the catastrophe. On the other hand, the CS is traditionally expected too thin to be observable since its thickness is believed to be roughly the proton Larmor radius, which is about tens of meters in the coronal environment. However, the first direct observation of the CME/flare CS showed that it could be as thick as a few times 10$^{4}$~km! A set of follow-ups by different instruments, and in different wavelengths consistently brought its thickness to the range from about 10$^{4}$~km to about 10$^{5}$~km! 
This talk will weave through the prediction of the catastrophe model for the long CS in the major eruptive process, the first identification of such structures, direct measurements of its thickness, discussions of the huge difference in the thickness between observations and theoretical expectations, possible impacts that may affect measurements, physical causes leading to a thick CS in which reconnection can still undergo at a reasonably fast rate, the numerical results displaying complex structures inside the sheet, as well as the results of the most recent laboratory experiments of reconnection driven by the energetic laser beams. Finally, conclusions based on the facts that could be collected so far are drawn.