First of all: I really hope this works out for Saxonica -- I am well aware of challenges, but hope they can be resolved. Such change(s) might make it easier to sell Saxon; I hope I can convince people I work with to fork out cash (so far have been unsuccesful). Wrt Intel exitting: I would have mixed feelings if it was due to free components (since I might be guilty to some small degree :) ), but I suspect much of it has to do with both xml processing tools becoming commodity, and tendency of commercial vendors to basically lie regarding performance boost one would get. Actually, same goes for many free software authors -- not Michael btw, never seen a false claim here, which is noteworthy and lends a lot to credibility of Saxon in my opinion. There are way too many snake oil salesmen, and too many sheep to fleece. From what I have seen, Intel has not been shy at making outrageous claims. But maybe it is understandable given prevalence of this practice within industry; and lack of freely available objective performance test results. You can't really test Intel's products in this field, even if you want to. And this tendency to exaggerate benefits is problematic because it (a) raises expectations to unreasonable levels (no, not parser will give you 10x speed increase over comparable best-of-breed alternatives) and (b) risks future of one's own sales, when product doesn't deliver up to claims.