Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
josephlord
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
11 ms
·
61.
▲
by
josephlord
12y ago
Number 2 could probably be abused. Get several people to respond to somebody and when they reply block them. Number 3 needs to be applied carefully as wrongfully banned people (it does happen) will be harmed.
62.
▲
by
josephlord
12y ago
I think anonymous accounts should be expensive (in terms of time) to set up. Anonymity and pseudonyms are important for those who need to whistleblow or are from oppressed groups so the possibility absolutely needs to be preserved. But what
63.
▲
by
josephlord
12y ago
Fair point the need to Box things is a bit of a nuisance. Also staying pure and immutable gets tricky as soon as you hit the Cocoa APIs but it is definitely a step in the correct direction and I still like it.
64.
▲
by
josephlord
12y ago
Swift passes the test in my view provided you can cope with using the word struct instead of object. (Objects based on classes are reference types in Swift but structs, enums and the default collections are value types). You can write prett
65.
▲
by
josephlord
12y ago
It might well be that it was a target price but there is actually a fair chance that the company may have achieved it by tightening their margins rather than expanding them. Source working for a TV manufacturer and knowing how desperate the
66.
▲
by
josephlord
12y ago
Land can make money purely by rising value over time so people can speculate by owning land. There may be little incentive to actually use the land. The ROI on a forest takes years whether or not the land value is taxed. You also said earli
67.
▲
by
josephlord
12y ago
> Land Value tax makes the rich pay a lot but hits the poorest really hard since the threshold to owning land (and creating wealth) is substantially raised. That seems a bizarre statement. Firstly I don't see any connection between
68.
▲
by
josephlord
12y ago
It is still personal data and theoretically regulated in Europe at least. It is hard for me to see how the information about non-users is legitimate from my understanding of UK data protection law but maybe they just stay inside the Irish l
69.
▲
by
josephlord
12y ago
They seem to trying suck up a significant proportion of Android and iOS devs in the London area. They are producing lots of interesting libraries too, shame they are Facebook.
70.
▲
by
josephlord
12y ago
What do you want? Do you want the investment instead? Do you just want to warn the investors that they might be being tricked? Do you want them to start the company and employ you (not a good idea if you can't trust them)? I suspect yo
71.
▲
The Importance of Being final (Responds to “Swift Resistance”)
(blog.human-friendly.com)
6 points
by
josephlord
12y ago
|
0 comments
72.
▲
by
josephlord
12y ago
I've posted this response: http://blog.human-friendly.com/the-importance-of-being-final Separately I got excellent optimised performance with a final class and this renderWeirdGraphics method which outperforms the Obje
73.
▲
by
josephlord
12y ago
The only thing that doesn't work in iOS7 (except for the new iOS8 APIs obviously) is frameworks. This can cause an issue on larger projects because the Swift compile can take some time but there are workarounds for during development.
74.
▲
by
josephlord
12y ago
This looks quite cool. Kind of like a lighterweight less invasive ReactiveCocoa. Obviously less powerful but with a lot less complexity and need to make it core to your app.
75.
▲
by
josephlord
12y ago
This PR puts the Swift array on terms I think: https://github.com/owensd/handmade-swift/pull/1 The final keyword means that at least one indirection is avoided for each access to the buffer. The speedup is si
76.
▲
by
josephlord
12y ago
Pull requests sent with both suggested changes. Both have a similar speedup effect (making the buffer final is if anything the bigger effect).
77.
▲
by
josephlord
12y ago
I've just tried making the buffer final and it results in a speedup of 8 times (not rigourously measured). From ~0.02 to between 0.0045 and 0.0025 on my computer.
78.
▲
by
josephlord
12y ago
One significant thing slowing this down is all the access to the the shared mutable buffer. I'm confident it would go quicker if renderWeirdGradient was changed to be a pure function that either returned a new populated buffer or had t
79.
▲
by
josephlord
12y ago
Mixed but the Swift team mostly yes.
80.
▲
by
josephlord
12y ago
> I guess it really comes down to who can 'reconstruct' my life. If it's the police or security services I believe that's absolutely fine, whoever you are (journalists, lawyers etc. included) Well I'm don't.
81.
▲
by
josephlord
12y ago
Others are worse. I don't think anybody would deny that. But that in itself does not make it a good state. But when did the US and UK start comparing themselves to the most oppressive rather than the freest in the world?
82.
▲
by
josephlord
12y ago
It doesn't make a difference until someone trawls the databases and can reconstruct much of your life. Also even if it doesn't matter to you don't you want journalists and lawyers free from communication (and who they contact
83.
▲
by
josephlord
12y ago
Ads?
84.
▲
by
josephlord
12y ago
There is an email address where you can request an expedited review. I'm sure they would accept for security issues (assuming you only make occasional requests).
85.
▲
by
josephlord
12y ago
Don't know if they have shipped anything major but the WWDC app was meant to be in Swift I think. You could probably tell if you get it in the debugger.
86.
▲
by
josephlord
12y ago
If you can record and identify the sound as you emit it and can identify the echo you shouldn't need to worry about jitter in output just the sampling frequency and any jitter there. I think you are right about the likely cluttered res
87.
▲
by
josephlord
12y ago
Thanks, not exactly encouraging for genuine reports though. It also doesn't say tht they will pass the report to all those groups just that you consent that they can. I would still hope for judgement to be applied. Even with that I wou
88.
▲
by
josephlord
12y ago
Surely best practice for abuse reports that you want to pass on would be to anonymize the complaints and act as conduit if appropriate. If the complainants wanted to complain to the site directly (rather than the proxy/CDN) they could
89.
▲
Cloudflare passes details of abuse reporters to 8chan
(mobile.twitter.com)
3 points
by
josephlord
12y ago
|
4 comments
90.
▲
by
josephlord
12y ago
Why wouldn't companies compete at least as much as now? They might charge more for higher tiers but I don't see any, justification behind your assertions.
More ›