I'm building a similar corpus in my work on system requirements documents. I
started off extracting an IT corpus from the BNC. It's 1.7M words although about
60% are news stories. Does that match your search for 'computer terms'? I have
word and POS tag frequency lists already produced.
Paul.