Using Popper as an excuse for not engaging with those you find intolerant is lazy. The paradox is that unlimited - not any - tolerance is self-defeating. Here's the man himself on the subject:
> I do not imply, for instance, that we should always suppress the utterance of intolerant philosophies; as long as we can counter them by rational argument and keep them in check by public opinion, suppression would certainly be unwise. But we should claim the right to suppress them if necessary even by force; for it may easily turn out that they are not prepared to meet us on the level of rational argument, but begin by denouncing all argument; they may forbid their followers to listen to rational argument, because it is deceptive, and teach them to answer arguments by the use of their fists or pistols. We should therefore claim, in the name of tolerance, the right not to tolerate the intolerant. -- The Open Society and Its Enemies
Popper doesn't conclude that one must absolutely intolerate intolerance, but that one must reserve the right to not tolerate intolerance when the expression of that intolerance "answers arguments by the use of their fists or pistols." One might even argue that by refusing to "meet them on the level of rational argument, but beginning by denouncing all argument" you are in fact that intolerant person that Popper argues must be met with intolerance.
This is a great example of what I keep trying to get people to recognize.
The trending method of improving diversity and inclusivity is through exclusivity.
This will not work. People and groups who don't talk to each other grow further apart and trust each other less.
It is a short step from being willing to dismiss opinions with a simple "that's racist" to being unable to discuss reality. See https://news.ycombinator.com/item?id=16908414 for an example of a conversation that is hard to have when you dismiss other people's views with "that's racist".
To take an obvious example, I'm sure his friends thought he was a swell guy, but I don't think we need to revisit Hitler's ideas about the Jews. I don't think we need to give much serious thought to theories about moon landing hoaxes or theories about how the Earth is flat either, no matter how nice the people promulgating them may be. We are only given so much time on this Earth; I can think of better ways to spend it than seriously engaging with obvious claptrap.
Everyone I know (and know of) espouses obvious claptrap from time to time. If that is the sieve with which you denote undesirables to remove from your personal sphere, your circle either has a radius of 1 or you delude yourself about your detection abilities.
However at this point "racists" organized and wound up controlling Congress, the Senate, the Presidency and the Supreme Court. The people that you'd like to ignore are currently running the place. Ignoring them might not be the best idea...
It's pretty unlikely that any dyed-in-the-wool Republican is going to hear your rational argument and be so astonished at your doctrine that he changes his mind.
I don't think this kind of exercise is that difficult, and I get plenty of practice on HN talking to people who have very right-wing ideas about the economy. But come on; it's not a way of working miracles. You might persuade your uncle who loves Donald Trump to moderate a couple of positions; you aren't going to turn him away from the Republican Party.
This is the main reason that election campaigns have moved on and focus much more heavily on turning out their own supporters than persuading "undecided" voters, who are a statistical nullity.
Is that what's at issue here? I think very few people -- nearly none -- are actually open to being rationally persuaded to completely flip their political beliefs. People with completely different political commitments usually start from different axioms.
What is a racist? Is this a fixed idea or does it change year to year and depend on the context? Shouldn't we be careful about writing off people based on fluctuating social mores? What about fascists, Marxists, communists, and anarchists? And, say, Trump supporters? Do they deserve basic human respect? Is there any reason not to dehumanize these groups too? Or others who fail our moral tests?
I lived with my sisters kid who was 22 at the time. We decided to do Christmas at our apartment. So about 30 people crammed into our two bedroom place. Coreys dad came and and was wearing a MAGA hat and shirt. Then he said Trump will win and my beaner girlfriend will be deported. He is a cop in Vancouver Washington at this very moment.
So I just grab a book and six pack and head to the roof of the the Jack in the Box a few blocks away. Easy to get up there and I can read and drink in peace.
Fast forward a year.
Now I live alone. Christmas at my apartment. Corey is in San Diego. The Trump supporter who has no real connection to any family that will be at my apartment asks for directions to the C-mas party. I tell him he is not invited. He is not happy. I do not care.
Seriously. Life is short. I am not going to try to get along with openly racist people. They can die in a fire.
That's pretty misleading. All systems I am aware of essentially try to compile query-specific code and avoid re-compiling runtime code that doesn't vary between queries (e.g. if you look at the HyPeR paper, that's exactly what they describe).
Compiling everything is questionable. There's not much point re-compiling runtime code or code outside of the hot path, it's expensive and doesn't bring any benefit.
E.g. things that aren't beneficial to compile per query include:
* Loops over a column of the same datatype, with no query specific branches (e.g. decoding a column of integers)
* Other static code, e.g. some hash table operations
* Outer loops that don't execute frequently
* Rarely executed code, e.g. error handling.
There are two general designs that let you compile only the necessary things. 1) the runtime calls into compiled code for hot loops vs 2) the compiled code drives the query and calls into the runtime. A lot of the systems you mentioned do the second, but Impala does the first, which seems to be the source of some misunderstanding. Also there were some cases where hot loops in earlier versions of Impala weren't compiled, but that's changed - generally we try to ensure that all hot loops are compiled.
I think generally the optimal design wouldn't be complete query compilation, but rather something more like a traditional JIT that selectively compiles parts of the query.
AFAIK, SQL Server's Hekaton engine only compiles the queries inside of stored procedures and UDFs. They do not compile every query that comes in from the client:
Which "HyperDB" are you talking about? I already listed Tableau's HyPer. Here are the other "hyper" systems that I know about and they don't do LLVM compilation:
In what way was it interesting? It was a document DBMS that supported MVCC.