Journals in Pure Mathematics have very low impact factors generally below 1 while journals of Applied Mathematics/Applied Sciences have more impact factors, then how the work in different disciplined be measured/compared by impact factors?
The Impact Factor of a journal reflects also the size of the scientific community in the specific field targeted by this journal. There have been propositions for alternative metrics better able to compare journals in differents scientific fields, eg the SNIP index proposed by the Elsevier's database Scopus (see attached link).
The Impact Factor has to do with how much the publication impacts on us. The higher, the better. It might be the case, disagreeing a bit with Philippe, that we have people from other areas interested in what we do.
I am fully agree with Dr. George Stoica. I think there should be some better index than Impact Factor which can judge the quality of any journal from any area of research. There are many journals in electrical engineering which are having low impact factor (4) but it is easier to get a publication in such journals. Although, these journals are having higher impact factor but do not have good reputation in the area.
The impact factor of a journal rather indicates the "eaverage impact" (in terms of number of citations) to be expected from a paper. To estimate the actual impact of a specific paper, it is much more informative to look at its number of citations rather than at the impact factor of the journal it is published in. To make a long story short, a paper published in Nature but never cited has certainly much less impact that a paper published in lesser known journal which has received thousands of citations!