You are going to have to help me out and at least try to make an argument instead of linking to others. That wikipedia page doesn't suggest anything in your supporting as far as I can tell besides 'when talking about animal rights, the ability to feel pain is a sufficient argument' and backs that statement up with 'higher functioning is not required because no matter how crippled, humans are still sentient'. This is an emotional argument and I am forced to say that ratiponallly no, humans that are sufficiently disabled are not sentient.
The lobster article just starts with the premise that they are sentient, dismisses the complete absence of brain structures that, if we lacked, we would likely be considered braindead, and say 'but they don't like being boiled!' Terribly unconvincing.
The first sentence of that Wikipedia page clarifies that animals are sentient? "Sentience is the ability to feel, perceive or be conscious, or to have subjective experiences."
The 'Consider the Lobster' article simply investigates the philosophical idea that the ability to suffer is the test for sentience and applies it to lobsters, noting how furiously they attempt to escape pain.
Sentience does not imply that living creatures must have the same biological structures to 'feel, perceive, or be conscious' so your point that they are absent brain structures we have seems irrelevant when considering the definition of the word.
Well the simplest answer: because a roomba is not a living thing.
Regardless of how sophisticated Roomba Mk #9999 is in aping the behaviour of living things, it itself is not living and so is not sentient: "Sentience is the ability to feel, perceive or be conscious, or to have subjective experiences."
I'm partial to Roger Penrose' position on AI, and his interpretation of Gödel's incompleteness theorem and its implications for Strong AI.
It is impossible, and therefore futile, to create consciousness through construction of algorithms, no matter how complex, as the mind is non-algorithmic.
I know there was recent HN exasperation on the apparent 'misapplication' of Gödel's theorem to the mind / matter problem, but I"d dispute the implication that it is misapplied through ignorance. It is simply educated disagreement between two opposing philosophical positions.
Many of the issues around materialism / physicalism / computationalism and the mind / matter problem are covered here http://rational-buddhism.blogspot.com. Yes, it's from a Buddhist perspective, but you'll quickly find a lot of articles seem to cover the issues from a computer science perspective.
Penrose's position boils down to him being uncomfortable with the idea that Gödel's theorems might apply to himself. His books have received criticism from the majority of the academic community and his suggestion for where the non-algorithmic nature of the mind may come from has not seen empirical confirmation. Simply put, it is just wishful thinking.
Regardless, Penrose himself dismisses the role of any supernatural factors; this leaves us open to the possibility that a non-algorithic but nevertheless synthetic intelligence could be created (and in theory wedged into my roomba). Orch-OR does not require 'meat'.
