That got a bit boring – every computer manufacturer seemed to have built their MSDOS a little differently – so we ended up doing some genetic engineering. For every patch we needed to make, we identified a pattern of assembler code near that patch. Now, when the user ran our program, our program would search for each of the patterns within the MSDOS code, and when it had made sure that it had found them all, this meant it knew where to insert all the patches. So our program would work on versions of MSDOS that we’d never even seen.

This is exactly what happens in DNA engineering, where a “restriction enzyme” looks for a pattern of genetic code and splits the DNA molecule at that point.

This sort of technique might save you some work (unless you’re planning to tell us, in the next part, that that’s what you’re doing). If Turbine changed the C++ compiler then you’d be stuck, because the assembler code emitted by one compiler will be quite different from that emitted by another; but it is very, very rare for a software manufacturer to change compilers. Even if your existing compiler has bugs, at least you know (by now) what the bugs are. If you changed to a new compiler, those old bugs would probably be gone – but how long would it take you to find the new ones?!

Thank you.

P.S. Looked for meaning of “Retraction” in Wikepedia, Google, various others. Please clarify that you didn’t mean something like the following: “Retraction of the foreskin alone is a simple and effective alternative to circumcision in managing most boys with a symptomatic, non-retractable prepuce.”