I am not sure how many bps is variant within of a splicing junction can it be called splicing sites? I read some exome squencing papers, they say they found some splice site, but non of them present this number.
The sequence ontology definition is reasonable but this definition is not consistently used. In defining "splice site mutations" some groups may use a simpler definition. For purposes of mutation annotation, it is common to consider the 'splice sites' to be only the two intronic bases closest to each exon. The adjacent bases in the exons may change amino acid sequence even if the splicing pattern is unaffected, so these will often be classified as coding mutations instead of splicing. In reality it is of course an oversimplification. However, the 'two bases' rule is often justified by observing that these two intronic bases at each end of the intron are the most highly conserved. For example in human, a large majority of introns have a 'GT' donor and 'AG' acceptor. The remaining bases around the splice site are considerably less conserved.
A typical human Donor/Acceptor layout looks like this:
5' - Exon1 - GT .. Intron1 ... branch site ... AG - Exon2 - 3'
In the Sequence Ontology, a splice_region_variant is defined as "A sequence variant in which a change has occurred within the region of the splice site, either within 1-3 bases of the exon or 3-8 bases of the intron."