HNHacker News
TopNewBestAskShowJobs

johnday

1,394 karma · joined March 22, 2014

submissionscomments
johnday··on Looking at the experience of black Britons through an American lens
It doesn't quite fall under your "10/20/30", but a lot of Irish and Northern Irish people were killed in The Troubles by British soldiers - a major event which many remember vividly.
johnday··on Judge temporarily blocks U.S. ban on TikTok downloads from U.S. app stores
The quote you pulled spells it out:

- Social priming does not replicate;

- Social psychology has a replicability rate of 25%.

johnday··on Haskell's Children
Interesting article. I think it's worth explicitly calling out that rather than proclaiming the death of Haskell, this article is talking about the children of Haskell and what they are capable of doing from a clean slate and with a clear goal.

IMO Haskell remains as the go-to language for teaching functional programming and for experimenting with PL concepts. Long may it remain.

johnday··on Haskell RecordDotSyntax language extension proposal (Accepted)
It is very cumbersome to write nested updates using the old syntax, and overlapping names between records was quite painful to deal with.
johnday··on Advocates want to name heatwaves like hurricanes to raise awareness of dangers
Sounds like a good idea. It might be harder to delineate when a heatwave is worthy of being named, but I'm sure the Met office (or the equivalent across the Atlantic) can manage it.

That said, I'm not convinced that giving it a name is enough to make people consider it dangerous.

johnday··on Apple: Apps should not require users to opt into tracking to access content
What would you actually be buying, if you bought the software without being legally capable of using it?
johnday··on Hermann Hauser: ‘It’s in Nvidia’s Interests to Destroy Arm’
Taxes on income do as you say. There are other forms of taxation which are gaining political ground over time, including a land value tax and (to a lesser extent) a wealth tax.
johnday··on Facebook isn’t happy about Apple’s upcoming ad tracking restrictions
I know that targeted advertising is more effective and I am not disputing that. I did not dispute it in my comment.

I am disputing your random assertion that, if advertising became less effective (say by removing targeting), the quantity of advertising that people see would go up significantly.

The amount of adverts that people see currently is not based solely on their economic value - it is based on their value relative to how much people are willing to put up with being advertised to. If advertising becomes less valuable, that doesn't mean people will suddenly put up with more of it.

Indeed, if the value of adverts went down, I would expect to see fewer advertisements, not more, since the value of showing them to people is now less relative to their willingness to put up with them.

johnday··on Facebook isn’t happy about Apple’s upcoming ad tracking restrictions
> The economic value of a well targeted ad is probably 10x otherwise.

> Ergo - unless you are willing to pay for apps - you might now see 10x more ads appearing on pages than otherwise.

I'm afraid that this is nonsense. If the economic value of selling bread went down by 10x, do you think people would start offering you 10x as much bread?

johnday··on ReMarkable MicroSD (2019)
People have written programs for the rM for doing this, yes. It's not a first-party use case though.
johnday··on ReMarkable MicroSD (2019)
In the case of ReMarkable 1, there are almost no issues with vectorised PDFs - when changing pages the screen update is usually the limiting factor, and that's relatively quick. For rasters, though, it does take a while to render out (>1 second for particularly complex pages). I think this is mostly a software limitation - some time is given over to drawing and saving the small thumbnail.
johnday··on Haskell In New Clothes
I addressed this point precisely in the edit! :)
johnday··on A philosophical difference between Haskell and Lisp (2015)
Oh good lord, that's horrible!
johnday··on Haskell In New Clothes
The fairly popular `LambdaCase` extension gives

  fib = \case
    0 -> 0
    [...]
johnday··on Haskell In New Clothes
Standard, base Haskell admits the following definition:

  let f lst = lst 
              & map normalize
              & filter (\x -> x > 0)
              & average
Which is barely different from your second presentation. (EDIT: just in case you think I'm ass-pulling here, this is fairly idiomatic code that I'd certainly write in prod! See another comment I posted earlier this week...)
johnday··on A philosophical difference between Haskell and Lisp (2015)
For maximum readability I'd be tempted to write the lambdas explicitly.

  xs =
    5
      & (\n -> take n [1, 2, 3])
      & filter (/= 1)
      & head
      & (\n -> take n [1, 2, 3])
johnday··on Capitalism Drives Cancel Culture
Movements are built off the back of small victories. It's not a surprise or a revelation that what you paint as 'cancel culture', a concept without personification, is exercising its (limited) power in the ways you describe. It seems likely to me that, as victories accumulate, so will the socio-political power of the group, until they are capable of making serious change at the level of nation-states.
johnday··on Capitalism Drives Cancel Culture
> Theoretically if it was known that obeying lead to better outcomes than anything you could do on your own people would follow them without coercion.

Anti-mask-wearers and anti-vaccinators have conclusively shown that this cannot happen in practise.

johnday··on Rethinking the Dictionary
As a person in a marginalised community (LGBT, so admittedly less marginalised than many in the West), I feel that UrbanDictionary provides an essential service to marginalised people. It is impossible to find out precisely what someone is trying to say about you without turning to UD, or diving into considerably more upsetting parts of the internet.

Indeed, I think the ideal mode would be to introduce a 'protected' flag, similar to Wikipedia's, and for which particular care is taken when constructing and assessing definitions. This will allow UD to act as a reference for the topics where it is most relevant, and not yet another vector for abuse.

johnday··on Flutter for Linux
It's also super laggy for me on Chrome.
johnday··on Web Dark Ages
A touch over 65,000.

[1] https://www.wolframalpha.com/input/?i=10+furlongs+in+attopar...

johnday··on New Facebook tool allows employers to suppress “unionize” in workplace chat
You would expect that the surplus of space in many US metropolitan areas would make it considerably easier to install and operate metro transit lines, but this seems not to be the case. I suspect the systemic bias towards road users is much more of a problem than geographical scale.
johnday··on Pain Points of Haskell
Absolutely. It's worth noting that every single effect system implementation that currently exists also leverages monads at the user level, even if they're doing exciting type-theoretic things underneath.
johnday··on Pain Points of Haskell
Sure, I'm outlining a worst-case. But realistically, even having two different hie versions is enough to make switching environments a Hard Problem.
johnday··on Pain Points of Haskell
Absolutely.

I'm only saying in relation to the idea of "reading Haskell code", it is possible to understand how they work (at the shallowest level) by analogy to `forEach` and to shell syntax.

Shell syntax also doesn't explain important constructs like currying, partial application, etc. But it's good enough that someone can read "boring" haskell code without being intimidated and turned off.

johnday··on Pain Points of Haskell
Oh, sure! In that case, mission accomplished with flying colours :)
johnday··on Pain Points of Haskell
As a PhD student, I don't have enough experience working with Haskell in an industrial context to comment on the engineering concerns. I believe you when you say that these things are all problems. However, I'm aware that at least some solutions are underway. For example, the shake library [1] is a nice way to set up competent, reproducible build systems.

[1] https://hackage.haskell.org/package/shake

johnday··on Pain Points of Haskell
I promise, you don't need to be a mathematician to use Haskell.

If you've written shell scripts before, you understand how Haskell functions calls are written.

If you've used `forEach()` in imperative languages then you understand functors.

These two things alone are half of the groundwork you need to read basic Haskell code! Of course, the fundamental underpinnings of the language will not appear just from a surface reading of the code, but if you're a person who learns by example then the above two are more than good enough.

If you're more of a linear learner then Learn You A Haskell [1] is an excellent, if earnest, introduction to the language. Depending on your confidence level you may be able to skip a decent way in.

[1]: http://learnyouahaskell.com/starting-out#ready-set-go

johnday··on Pain Points of Haskell
Yeah, sorry, my website got nuked from orbit. Hopefully it'll be back up tomorrow.
johnday··on Pain Points of Haskell
> Nowadays, the Haskell IDE engine is good enough for general use.

Sadly I am unconvinced of this. If it were true, there would not be an immediate and significant push to make `haskell-language-server`.

Having tried many times to get `hie` working, I can say that it's a pain in the bum.

Suppose I have 12 projects, one made every month for the last year. Each of these 12 will be using a different stack resolver, possibly with different ghc versions. As a result, I will need to compile and use up to twelve different versions of hie, and make sure that the correct one is on my $PATH depending on which project I have open.

← PreviousPage 9 of 17Next →