I have seen quite a few research publications where the literature review section doesn't compare past work to the author's work. It's like:
 did this using X technique.  used Y technology. (etc.)
This type of literature reviews are easier to write as I don't have to read similar works in extreme detail to create a list of things that we are doing and they are not. I can understand their technique by reading their abstract but it won't matter too much as long as I am using a completely different technology.
I have also seen some good publications which compare each and every work in literature review section to their own work and say "They did not do this", "Their work couldn't do that". While, of course, this looks a lot better but it's a lot harder to write.
Let's say I using queuing theory for a similar work which has been done using some other technique. I can easily state that they used this technique in their work instead of trying to find drawbacks in their work.
So which approach is better when writing the literature review section?