Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
hpen
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
17 ms
·
181.
▲
by
hpen
5y ago
Hey have you guys considered licensing an API?
182.
▲
by
hpen
5y ago
Hey bro you should credit Bjarne with your little quote there
183.
▲
by
hpen
5y ago
Totally agree. But it DID happen in Java. So, I was thinking some might be turned off from bad PR.
184.
▲
by
hpen
5y ago
Nah this is a genuine question from me. My friend (a Java programmer) brought up the idea, and I was curious. Also I read something on HN about it being the most wide spread and critical exploit of the decade.
185.
▲
React is building a compiler for auto memoization
(youtube.com)
4 points
by
hpen
5y ago
|
0 comments
186.
▲
by
hpen
5y ago
Very strong argument. That's sarcasm.
187.
▲
Ask HN: Will Log4j exploit be the end of Java?
1 points
by
hpen
5y ago
|
9 comments
188.
▲
by
hpen
5y ago
Is he a bigot because he doesn't share the same opinions on LGBTQ and same-sex marriage? Or is there more to the story?
189.
▲
by
hpen
5y ago
Is this trolling?
190.
▲
by
hpen
5y ago
This exactly.
191.
▲
by
hpen
5y ago
That being said, I think it is rewarding well payed work.
192.
▲
by
hpen
5y ago
Swift is great. I am actually making a pivot to Swift & iOS/MacOS now. I would say it is more niche than you might imagine and you will be expected to know a lot about the ecosystem.
193.
▲
by
hpen
5y ago
Lack of an interface mechanism, with multi inheritance instead, is messy. This is my biggest complaint that doesn't apply to pretty much all dynamic languages.
194.
▲
by
hpen
5y ago
I like async. It's is easy to reason about. In what paradigm are race conditions avoided completely? I know Rust avoids "Data Races" but not race conditions as far as I'm aware.
195.
▲
by
hpen
5y ago
I have done JS for several years and TS more recently. I think TS really fixes a lot of the issues that arise while building large systems in NodejS. But I think people are right to hate on the dependency nightmare that is npm. I could comp
196.
▲
by
hpen
5y ago
Definitely our home boy Lex Fridman
197.
▲
by
hpen
5y ago
I have also wondered this question, as I have relatives who always pay their tithe to the church but are middle class and have over $100k in debt at retirement age. It's seams unnecessary to require such a thing of a "saint"
198.
▲
by
hpen
5y ago
Well I agree marketing is effective. But what's the argument for housing as a human right? You say "It doesn't really matter if housing as a human right doesn't fit some narrow implied definitions of 18th-century philoso
199.
▲
by
hpen
5y ago
I mean open source pushes things forward but lack of IP? I don't think so. Linux is a great example of how open source brings great things, it's also a great example of how IP has helped open source.
200.
▲
by
hpen
5y ago
< Goals can be avoided or punted. By calling housing a "right", we elevate it as an imperative, not just something to be shrugged off. It is explicitly the purpose to make the issue more difficult to avoid. I think your more in
201.
▲
by
hpen
5y ago
It's interesting because even FANG companies can't agree on testing. I've heard Facebook and Google are very different in this regard.
202.
▲
by
hpen
5y ago
Good point about scalability. The first company I worked for was tiny, and this really hits home :)
203.
▲
by
hpen
5y ago
My original reply to you was actually me agreeing with you and slightly modifying your use case for me. How is that self selecting or self defining? > I'm not sure what scale of current, existing plurality we need to start ignoring
204.
▲
by
hpen
5y ago
Generally I believe in engineering tradeoffs but I think some tests are always important. Even with MVPs. Writing test might double your time to market, but they will allow you to move faster and make less regressions once in production. Th
205.
▲
by
hpen
5y ago
Meh not sure what the big deal is with "self selecting" and "self defined circumstances". Perhaps you should make more arguments for it so I don't have to but you do you. That circumstance I presented isn't act
206.
▲
by
hpen
5y ago
Sounds reasonable. Especially for a ML group. I just know that many times the prototype turns into the production product.
207.
▲
by
hpen
5y ago
To me Python is fast just like not writing tests is fast.
208.
▲
by
hpen
5y ago
Well I did come up with one use case in my last comment. So i wouldn't say I ruled out all potential uses for myself :)
209.
▲
by
hpen
5y ago
Python really sucks for large projects. Also, I hate the GIL
210.
▲
by
hpen
5y ago
I love Python for many things: scripts, data science, prototypes, etc. But I would never use it to build a large backend system. I just can't handle all the warts. Yes it has been used by large companies, but not without their problems
More ›