If I just want to link to a paper, blog post, presentation etc. that I think might be relevant, I post it as a comment, not as an answer. Although pointing people to external resources can be helpful, that in itself is not an answer.
A good answer can certainly refer to external links, but it should at least offer a paragraph or two of explanation in the answerer's own words of the technique and why/how it's applicable to the specific question being asked. This isn't just about ensuring that the answer remains useful even if the link evaporates (although that is important); it's also a way to demonstrate that the answerer has actually thought about it and not just typed some keywords into Google. A good example is John Calsbeek's answer on spherical noise.
A good answer can also be a "literature survey" that has several links, covering a range of different possible techniques or algorithms. In this case the answerer can't explain everything in detail as the answer would be huge, but they should still try to put things in context and compare/contrast the different approaches, so the reader can understand how all the links relate to each other, and make decisions what to investigate further. Krzysztof Narkowicz's answer on LOD techniques is a great example of this.
The answer referred to in this question doesn't provide any of the answerer's own expertise and moreover does not even clearly cite the pasted text as being from somewhere else. The link to the actual paper only appears in a comment! It is simply lazy, one step above outright plagiarism, and should be deleted. I have voted to do so.