You are right: the range of the NDVI is limited to values between -1 and 1 due to its' normalization properties.
The negative limit of -1 will be reached if you encounter maximal reflectance (1) in the red wavelength region and zero reflectance in the NIR. The positive limit will be reached by maximal reflectance in the NIR region and zero reflectance in the read wavelength region.
uses coefficients to correct for canopy background () and aerosol influences ( , ) (see e.g. Huete et al. (2002) for details.
Especially for artificial surfaces you can encounter values out of the range between -1 and 1. For example, blue rooftops, which show high reflectance (e.g. above 0.6) in the blue wavelength region will generate EVI values outside the "typical" range. This is caused by the use of the blue band to correct for aerosol influences in the red band.
Consequently, you can encounter values greater 1 or below -1 for the EVI but for the NDVI the values will always range between -1 and 1.
Regarding your second question: you should try different vegetation indices and analysis techniques according to your problem at hand. If you want to estimate vegetation abundance in urban areas you could for example use spectral mixture analysis.