Because we cannot yet identify dyscalculia based on brain function, we have to diagnose it based on its effect, i.e. difficulty with maths. This is difficult because "there are many reasons for being bad at maths!". Reasons other than dyscalculia include inadequate instruction, lack of motivation, attentional disorders, anxiety disorders, or across the board academic difficulties.
For this reason, most methods of diagnosis include not only 1) identification of a difficulty in maths that affects academic or everyday life, but also 2) an attempt to rule out some of these other factors. [1, 2]
One problem with identifying a difficulty in maths is that there are many different tests to use. Which test should be chosen, and where should the cut-off be? There is no general consensus on this, and tests vary widely in their content. The implication for interpreting research is that it can be difficult to compare studies because of the different criteria used. 
What this means for teachers and parents is that diagnosis should be done by an expert - a qualified school psychologist or special education worker. This person will talk to parents and teachers, and run a variety of tests, including for maths, but also for IQ, and possibly for reading or attention. Most importantly, they will then use their professional judgement.
An important outcome of research over the next few years should be to vastly improve identification. Once we know more about the symptoms of dyscalculia, how it develops over childhood, and its brain bases, we will be able to develop better tests and use tests much earlier.
In general, however, what happens is that by the time parents and teachers realise that a child has a serious problem with maths and find out how to do something about it, the child is already 9 or 10 years old and is 3 years behind in school. It would be wonderful to be able to test children's behaviour and brain patterns in kindergarten, and pick out those at risk for dyscalculia so that they were already receiving extra monitoring and tuition. There is even the hope that dyscalculia could be "prevented" in this way.
 R. S. Shalev and V. Gross-Tsur, "Developmental dyscalculia," Pediatric Neurology, vol. 24, pp. 337-42, 2001.
 M. l. M. M. Mazzocco and G. F. Myers, "Complexities in Identifying and Defining Mathematics Learning Disability in the Primary School-Age Years," Annals of Dyslexia, vol. 53, pp. 218, 2003.
 B. Butterworth, "Developmental dyscalculia," in Handbook of Mathematical Cognition, J. Campbell, Ed. New York: Psychology Press, 2005.
 B. Butterworth, "Dyscalculia Screener." London: nferNelson, 2003.
Reprinted with the permission of Dr. Anna Wilson. © 2007-2008 Dr. Anna Wilson.