But he's not parsing html. The parent node never even mentioned parsing html. All he wants to do is match a certain substring and replace it with another one. The minor fact that this string happens to contain data that superficially resembles html has absolutely no bearing on this.