@RyanEls4 And just like code, we go step by step. Label it, trace it, find its purpose, find a better place, shape, colour for it. Step back and look for the patterns.
And if all else fails, "the scream test" gets results fast.
🧵Next week, you will see people using something called Benford's Law to try to prove election fraud.
⛔️These people are wrong⛔️
I am a scientist who has published on Benford's Law. Let me tell you what it is and why what they are doing is mathematically incorrect. 1/
@TotherAlistair If you're in civilised parts of the UK it's pronounced Henderson's; and yup got 2 bottles in the house right now because one is getting low. Should see us thru to the middle of next year.
@allenholub@AgileJebrim This doesn't mean that we all should go down to Assembler. As I said, that abstraction often makes us more productive.
But there are things where the efficiency and direct control of machine programming in assembly language are necessary.
Do you have a system of delivery or a Rube Goldberg machine?
System of delivery:
git commit && git push ⇒ runs tests, if they pass, deploys to pro
Everything signitifantly more complicated than that is possibly a Rube Goldberg machine.
#xp#ContinuousDelivery
Happy 23rd birthday Wayback Machine!
On October 24, 2001, The Internet Archive organization launched a free digital archive of websites for the general public called the Wayback Machine.
The oldest pages stored in the archive date back to 1996.
#InternetHistory
@allenholub Confident dev: It's just a CSS change, it can't break the site, so we don't need to test it
me: .html { display: none; }
dev: that doesn't break anything
me: Yeah, I think we're gonna test that change
@EzProgramming Uses the same kind of logic as the infamous:
All programs contain at least one bug.
All programs can be shorted by at least one line.
So all programs can be reduced to one broken line of code.
Someone actually said this in a discussion and believes it:
"You should be able to plan a sprint that you can complete in time. If you can plan a sprint that you can complete in a two week period, then it naturally follows you can plan 2 sprints, or a quarter, with the same results."
I am 99% sure they have never written a line of code for an enterprise before.
Claude just shipped “Computer Use” which allows an LLM to control your computer and do arbitrary things like browse websites, download and run files, and more.
A slew of startups doing browser agents feel obsolete overnight.
Seriously cool stuff.
Cookie banners.
Just visited a US website, from the US, that ships a 457.11kB (minified!) JS bundle of a GDPR "banner SDK" from a "trust" provider. Over 3x the size of React. It ships its very own version of jQuery inside.
Going to the "trust" provider website yields a 5.3s LCP (i.e.: 5s+ to load the first screen completely). 53% of visitors experience a similarly terrible loading experience. They couldn't care less.
This is the stuff that's silently destroying the web.
• For most users, these providers have gamified beyond belief your ability to actually block cookies. The primary buttons are always "Accept All", even when you "customize your preferences". More often than not, they're cheating you into accepting everything.
• The amount of JS they ship, downloading from a 3rd party website, is destroying the web's performance
• The aesthetics and function of websites is massively compromised. In the spirit of "privacy and trust", they're eroding the open web in favor of proprietary platforms.
• They undermine the product engineering teams' efforts. I've met so many amazing design engineers who spend countless hours honing experiences, only to have them destroyed by cookie junk.
Cookie banner slop has to stop.
@allenholub See also "it's not a global, it's a 'singleton' pattern"
and "it's not a global it's a 'config' object"
Globals aren't inherently bad, they're useful in the right place and for the right reasons. Just like databases.
EJB1 and EJB2 and AWT and Applets and JavaBeans and early RMI. Those still occasionally rise like Lovecraftian horrors in my head.
Almost as bad as Oracle 7 and Websphere.
Just one interesting insight from this one is that early f$ck-ups can affect the image of a technology pretty much forever. Several people mentioned J2EE, while it's nearly 20 years now since this has been a thing.