In an ideal world, our photospectrometer would be a simple instrument: light source, sample, detector. The detector would measure absorbance as

, where I is the intensity at the detector and I
0 is the intensity at the source. Naturally, as light is absorbed, I decreases, I/I0 decreases and absorbance is greater. In the ideal world, A = 0 for a blank sample as intensity at detector would be the same as the source.
In real life however, many things (diffusion, scattering, absorption by glass) can reduce the intensity at the detector. This means as intensity at the detector is always less than the source. Also due to power fluctuations in the source, we cannot know for sure what I
0 is. Therefore usually at least two measurements are conducted at any given time, one measuring a blank and another measuring the sample (i.e. light travel down two paths in a UV vis machine, see your textbook). This means I
0 is really arbitrary, and absorbance can be anything since it is relative, the computer inside the spectrometer take care of that for you (and you can set the reference absorbance to any arbitrary number). For simplicity sake we generally calibrate the the blank to be approximately zero. I have worked with both negative and positive absorbances, it does't make a difference.