Profile picture
Dan Luu
@danluu
, 5 tweets, 2 min read Read on Twitter
Ken Thompson on languages and safety

(from @peterseibel 's book, Coders at Work)
Hillel Wayne's response to Uncle Bob also seems to work as a response to Ken Thompson

hillelwayne.com/post/uncle-bob/
I often wonder if programmers are unique in the level of stockholm syndrome we display. Are other professions like this?
Interesting to see Swift's justification for adding isEven and isOdd to the language (because it would help people) in contrast to old school attitude attitudes (people who don't know everything or make mistakes deserve punishment).

github.com/apple/swift-ev…
One thing I've always found funny about the "programmers deserve problems if they make mistakes" attitude is that most of the cost falls onto users, not programmers.

Do people think users deserve problems because of original sin or something?
Missing some Tweet in this thread?
You can try to force a refresh.

Like this thread? Get email updates or save it to PDF!

Subscribe to Dan Luu
Profile picture

Get real-time email alerts when new unrolls are available from this author!

This content may be removed anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

More from @danluu see all

Profile picture
Dan Luu
@danluu
An empirical study on the effect of open offices: "the volume of face-to-face interactions decreased significantly (approx. 70%)"

rstb.royalsocietypublishing.org/content/royptb…
Unfortunately, it's hard to tell what's going on from the text of the study.

In "Study 1", they got 52 people from a floor with people in "technology, sales and pricing, human resources (HR), finance, and product development, as well as the top leadership"
These are really different things and one might expect different results. perhaps even directionally different results, depending on the composition, but the actual composition isn't mentioned at all (nor is the idea that it might matter).
Read 5 tweets
Profile picture
Dan Luu
@danluu
Google has a new API that tells you how "toxic" text is, h/t @taoeffect

perspectiveapi.com
It's 🤣 to poke around and see what biases the system picked up from the training data. 😰 to think about actual applications, though.
Maybe it's not fair to try single words when the model is designed for use on comments, so let's try things that could be entire comments 😇
Read 4 tweets
Profile picture
Dan Luu
@danluu
Developer hiring and the market for lemons

danluu.com/hiring-lemons/
I've thought about this @holman tweet at least once a week since last July.

Even when I've worked on things that are considered relatively hard (e.g., an algorithm that won best paper award at a top conf), the work hasn't been interview-style hard.

To the extent that things are hard, it's been stuff like:

Your TL thinks version control and automated tests are a waste of time. How do you convince them otherwise?

Your team has 75% turnover in well under a year for... reasons. How do you stay productive?
Read 4 tweets

Related threads

Profile picture
Hillel
@Hillelogram
My stance was "doing up-front design radically improved our product", and Ron asked why we weren't designing before then. This is a complicated question. The short answer is: we thought we were doing good design, and we were wrong. It was mainstream, but not good.
We were developing a product continually, iteratively, designing and implementing in short sprints. We had continual customer feedback and agile adaptation. We used TDD. This is the way that people recommend designing, see, ronjeffries.com/articles/019-0…, , etc.
This worked... poorly. For two years. And this wasn't a junior project! People with tons of experience were building this. We were "designing" things _right_, as engineers are _supposed_ to design. And it didn't work, and the fires kept burning. It was too essentially complex.
Read 10 tweets
Profile picture
Hillel
@Hillelogram
First day back at work? Procrastinate just a little bit longer by reading my 8,000 word postmortem of the npm event-stream incident. I promise it's interesting even if you don't care about JavaScript. You'll learn how to do accident analysis! hillelwayne.com/post/stamping-…
This was a really intense one so I'm extra grateful to the people who gave feedback. I want to find a better way to thank them than just "thanks to" at the end, so until I find that better way I'll just write twitter plugs for all of them below.
@bellmar is a senior software engineer at the US Digital Service and an expert on govtech bureaucracy. No seriously, she's like writing a book on it.

She also writes about COBOL and mainframes! And she _might_ have convinced me to donate blood.
Read 5 tweets
Profile picture
Hillel
@Hillelogram
_Why don't_ these accomplishments make it into the mainstream? Two easy answers are "they aren't that great" or "people don't know what's good for them". I see both used all the time.

A harder answer: we are REALLY bad at persuading people.
No matter how good an idea is, people won't naturally accept it. You have to convince them, which is hard. Most people are skeptical of new ideas in programming, which I think is usually a good thing. For every idea like Quickcheck there are a dozen like Tonal Bitcoin.
Persuading people is a skill! It's one a LOT of us are bad at. We often advocate for things we can't imagine working without... so we can't empathize with people who DO work without it and so struggle to find common ground.
Read 12 tweets
Profile picture
Hillel
@Hillelogram
People often say formal methods are just an academia or "NASA" thing. They're not. Here are just a few examples of "industry" problems you can tackle with formal specification. I've included examples if they're publicly available, otherwise I've seen/done it in private code.
Business code: Does my system accurately follow all the requirements of the customer? medium.com/espark-enginee…
Service architectures: if we have several independent communicating services, do we ever have dangerous race conditions?
Read 12 tweets
Profile picture
Hillel
@Hillelogram
Microservices are an operations tradeoff: scalability and fault tolerance for the price of managing a distributed system. If you're using microservices as a development tradeoff you're not modularizing your code properly (or your language does not support proper modules.)
To be "proper" modules you need to be able to define the public interface of the module AND be able to put whatever you want in the modules implementation, including other modules. Most languages only have a "biggest box with a proper interface" that can't be nested, eg classes.
Using microservices here _sort of_ gives you a bigger box: you can only interface by the open routes, such as by http calls. But even then you can't draw a bigger box around microservices, which is why people accidentally create microliths.
Read 7 tweets
Profile picture
Hillel
@Hillelogram
This is a claim I see a lot, so I'd like to share my thoughts on it.

There's no good evidence that static typing leads to more correct software. Most papers on the topic are either flawed or find no connection. There may be one, but we shouldn't assume there is one without data.
First some quick premises. Software correctness is a spectrum. Safety-critical systems need 100% correctness. Everything else has some leeway: even Google or hospitals can prob get away with 99.9% correctness. Most of us would be stellar with 90% or even 85%.
Second, there are different costs to deal with correctness. 100% is much more expensive than 99%.

Third, there are different severities of software correctness. "This doesn't work the first click so you have to click twice" is less severe than "This returns the wrong data".
Read 20 tweets

Trending hashtags

#pulitizer #UnitedShades #IOF #trumphotel #arabic #WestBank #Mnuchin #dc #maga #thistown #sciencenotpolitics #InternetBillofRights #JordanValley #NationalWalkoutDay #WERKforConsent #MeepMeep #KhashoggiMurder #america #hospitality #Uber #RecordStoreDay #RogerWaters #Brexit #TBT #Israel
Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just three indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!