HN user

cpdean

564 karma

I like doing computers

Posts30
Comments144
View on HN
engineering.squarespace.com 8y ago

Building on Solid Ground: Getting Postgres Foundations Right with Pgbedrock

cpdean
4pts0
news.ycombinator.com 8y ago

Ask HN: Laptop reviews, but specifically for Linux/BSD hardware support?

cpdean
4pts1
www.nytimes.com 9y ago

Lyft Gets $500M in New Funding as Its Rival Uber Wobbles

cpdean
1pts0
www.sitepoint.com 9y ago

Life After JavaScript: The Benefits of Learning a 2nd Language

cpdean
2pts0
www.latimes.com 9y ago

The question that continues to loom over Uber and Lyft – and vex lawyers

cpdean
1pts0
www.forbes.com 9y ago

iPhone 8 Leak Reveals Apple's Expensive Secrets

cpdean
2pts1
www.glamour.com 9y ago

Work Stress Is Killing Us, but What Can We Do About It?

cpdean
2pts1
nypost.com 9y ago

Stress, isolation blamed for suicide cluster rocking Ivy League campus

cpdean
2pts0
fortune.com 9y ago

Why Amazon Alexa Will Take Over Your Phone and Your Office Next

cpdean
2pts0
www.wired.com 9y ago

The Secret Uber Data That Could Fix Your Commute

cpdean
3pts0
www.forbes.com 9y ago

Mindfulness Meditation May Help Treat Anxiety Disorders

cpdean
3pts0
arstechnica.com 9y ago

Cyanogen Inc. shuts down CyanogenMod in Christmas bloodbath

cpdean
2pts1
talkincloud.com 9y ago

Dropbox Unveils Infrastructure Updates

cpdean
2pts0
www.businessinsider.com 9y ago

Analysts say Airbnb is hurting hotels more than predicted

cpdean
2pts0
fortune.com 9y ago

Mark Zuckerberg Continues to Miss the Point on Facebook and Media

cpdean
11pts0
www.forbes.com 9y ago

Fake Facebook News Is a Disaster for Political Discourse

cpdean
16pts1
www.bloomberg.com 9y ago

The Spy Who Added Me on LinkedIn

cpdean
3pts0
www.pcmag.com 9y ago

Hands on with Snapchat Spectacles

cpdean
2pts0
qz.com 9y ago

Can psychopaths be changed by love?

cpdean
2pts0
www.huffingtonpost.com 9y ago

Childhood Origins of Narcissism

cpdean
1pts0
deadline.com 9y ago

LAPD Prepping for “Riot Mode” in Some Neighborhoods If Donald Trump Wins

cpdean
13pts0
www.wired.com 9y ago

FAQ: Analyzing Social Data to Understand the US Electorate

cpdean
1pts0
www.seattletimes.com 9y ago

How to survive this ‘stress trigger’ of an election

cpdean
1pts0
blog.conradpdean.com 9y ago

You're Writing the Wrong Tests

cpdean
1pts0
blog.cloudera.com 10y ago

How-To: Prepare Your Apache Hadoop Cluster for PySpark Jobs

cpdean
1pts0
blog.cloudera.com 10y ago

How-To: Translate from MapReduce to Apache Spark

cpdean
2pts0
www.theguardian.com 11y ago

Malcolm Gladwell: how I ruined my best friend's wedding

cpdean
1pts0
archive.org 12y ago

The Voynich Manuscript

cpdean
1pts0
www.reasonwell.com 13y ago

Will software save the economy?

cpdean
1pts0
www.scientificamerican.com 13y ago

Wisdom from Psychopaths?

cpdean
110pts58
Clolog 1 year ago

I absolutely love the aesthetic of a repo having a giant README.md

Fair Cake-Cutting 2 years ago

There's a really interesting boardgame based on the cake-cutting problem https://boardgamegeek.com/boardgame/173648/booty .

You play many rounds of trying to fairly divvy up piles of loot. Different pieces combine in ways to increase their overall value, as well as each piece having a base level amount of value. As a result the actual act of dividing it fairly becomes complicated, and players will try to influence the outcome in order to maximize their score at the end of the game to win.

I like SQL and all but I really don't care to follow ISO releases. They're hundreds of dollars and nobody actually implements the whole thing. I get way more excited about database releases.

Does anyone else find value in what's in an ISO standard?

Even if a project is not hosted on GitHub, other people have the legal right (depending on the license) to redistribute the source code. It means that they have the right to share the code of others on GitHub, as long as they respect the terms of license. This is totally legal. But then, Copilot will be able to analyze the code and violates the license terms, which isn’t.

While encouraging people to not distribute code via Github may mitigate the issue some, the actual issue is how Github has mass-automated the process of violating open source licenses. Github should pay a fine for every suggestion Copilot produces that violates a software license, plain and simple. Don't blame the people that unknowingly upload code to the training dataset.

Fascination of Awk 3 years ago

I'm a huge fan of awk but the "Python vs awk" page this links to [1] shows python code that's almost deliberately atrocious.

Take this function the author wrote for converting a list of integers (or strings) into floats

    def ints2float(integerlist):
        for n in range(0,len(integerlist)):
            integerlist[n]=float(integerlist[n])
        return integerlist
Using `range(0,len(integerlist))` immediately betrays how the author doesn't understand python. The first arg in `range` is entirely redundant. Mutating the input list like this is also just bad design. If someone has used python for longer than a month, you'd write this with just `[float(i) for i in integerlist]`.

Further down in the function `format_captured` you see this attempt at obfuscation:

    freqs=ints2float(filter(None,captured[n].split(' '))[2:5])
Why bother with a `filter`? Who hurt you?
    freqs = ints2float(captured[n].split(' ')[2:5])
That said, the author's implementation in awk does look pretty clean. I'm just peeved that they straw-manned the other language.

[1] https://pmitev.github.io/to-awk-or-not/Python_vs_awk/

i can’t for the life of me reason why the changes to auto correct were ever seen as useful. that it changes a word you’ve typed two words ago (so it’s outside of your attention by the time you’re focused on getting the most recent word correct) is baffling.

Additionally, by designing it this way you’re guaranteed to only notice when this feature fucks up. If you typed the wrong word and moved on, you never notice. If you typed the right word and moved on, you see auto correct has fucked up after you send the message, or as it’s fucking up because you’ve learned to not trust the keyboard anymore so now you’re constantly rereading the last few words as you’re typing.

Really a shame they put effort into a feature that’s designed to frustrate the user in this way.

HN is up again 4 years ago

Its weird how I saw this headline, thought to myself "Oh good! It's back!" before realizing that I was using HN to see this very headline...

For me the issue was less about the bugs in pipenv, and more about how Kenneth Reitz leveraged his commit rights to other projects' documentation to update them to claim that they all recommended pipenv when pipenv was in its infancy.

If they get banned, people like me will lose a part of their hobby and noisemakers will find another way to be obnoxious (burnouts? screechy braking? loud music?)

Like most things in New York, it comes down to what NYPD officers actually care about enforcing. Per the article:

But the law doesn’t doesn’t compel law enforcement to issue tickets, nor does it require police cars to be equipped with decibel readers.

tldr: these laws won't do shit then. Officers won't enforce what they aren't required to, this just gets added to the list of offenses they can use to harass someone or tack onto a speeding ticket.

You are correct, ereaders have supported viewing the web for years, but it has been a miserable experience since the beginning.

    * Inline links everywhere tempting you to click away to a new page, both fracturing your attention as well as being a multiple-second delay between pages
    * Images are everywhere, needing to be interpreted by the rendering engine of the ereader either with a simple monochromatic threshold to figure out what parts of the fully color image are black and white, or if you're lucky it has some way to re-process the image through dithering so that it's possible to see what it is
    * web layouts always wanting you to have a screen larger than an ereader can support, so you're stuck scrolling back and forth on a tiny screen that already has a miserable refresh rate of once or twice a second.
    * the web is an application platform now, the vast majority of its features just not fitting within what an ereader can support
You're all correct, you can easily implement what Gemini does if you make one website that is only text. In fact, you can probably do it better since you don't have to be constrained to gemtext, or constrained to its handful of weekend-project servers for hosting the site, all of which have their own issues.

But the thing that keeps drawing me back to is is that while you can easily implement Gemini with HTML and HTTP, it is impossible to implement HTML and HTTP in Gemini. The fact that you can't implement a webapp in Gemini means that if you're browsing pages in Gemini, you get a consistent experience and every site is clean and respectful of the experience (whether the author wanted to respect it or not, they have no choice).

Perhaps the real counter-argument to Gemini is not "why don't website authors just make their websites simpler?" but instead "why don't modern browsers run in 'reader mode' by default?".

The problem I've noticed is you can't get enough people to respect the minimalism. Yes you could implement Gemini in HTTP by having everyone play by the rules, but how many of the links even on HN go to pages with ads, autoplay videos, and "sign up for updates" modal popups?

When I browse Gemini with Lagrange ( https://gmi.skyjake.fi/lagrange/ ) if you avoid orange links, then you're guaranteed to have a quiet, text-only experience.

what you cite as a bug feels like a refreshing feature to me. Looking at how the Lagrange [1] client interprets Gemini shows how rich and enjoyable Gemini can be, and it leaves me optimistic that we could browse it in the future on something like an ereader. The constraints that it imposes makes it so the author is encouraged to write text, not a website.

This doesn't have to reproduce the entire suite of features that a modern web application offers.

[1] https://gmi.skyjake.fi/lagrange/

Something that is still a difficult hurdle for me to overcome is how Prolog starts out feeling declarative but then very quickly you need to be aware of all the backtracking rules to correctly guide Prolog to the solution you want. I have tried to get through Clocksin's Programming in Prolog a couple times now and it still hits me how quickly you need to get into the invisible details nobody blogs about in order to be even base-level productive in it.