HNHacker News
TopNewBestAskShowJobs

Goog_l_Dude_l

4 karma · joined March 31, 2021

submissionscomments
Goog_l_Dude_l··on Israel may have achieved herd immunity against Covid-19
If you look at age distribution, the numbers are actually much higher for adults, but Israel has a very large proportion of children under 18. For adults over 40 the number is around 85%, for adults under 40 it's around 75%. This is the Israeli health ministry's dashboard in Hebrew, but you can use Google translate. https://datadashboard.health.gov.il/COVID-19/general?tileNam...
Goog_l_Dude_l··on Isamu Akasaki, inventor of first efficient blue LED, has died
For anyone wondering (like I was): Why A Blue LED Is Worth A Nobel Prize (2014) https://www.popsci.com/article/technology/why-blue-led-worth...
Goog_l_Dude_l··on Modern Code Review: A Case Study at Google (2018) [pdf]
The things that always blows my mind is how every individual piece of code you read has good, clean, well documented, well tested code. But overall it's just a big mess.

Not everything, though. Infrastructure code and libraries are usually very high quality and very nice to work with.

Goog_l_Dude_l··on Modern Code Review: A Case Study at Google (2018) [pdf]
It was certainly valuable, especially in C++ which includes a lot of internal standards and libraries, although at times I felt like a lot of these conventions can be enforced by a smarter linter, but I am writing better C++ code after this process despite having years of experience pre-Google.

I didn't feel the same for Python though.

Goog_l_Dude_l··on Modern Code Review: A Case Study at Google (2018) [pdf]
The review experience can really vary between teams, and depending on whether you have readability or not and whether you're changing code you own or not. This can make the difference between a couple of hours to a week to submit the same change. Readability in general is a process that while valuable, but can be really streamlined. I was working in a fast moving team using Python where everyone had readability, and I just didn't have time to delay most non trivial CLs, so even after submitting hundreds of CLs I did not have readability, having sent only a handful for review. Another negative effect I saw was people opting not to use a less popular language that was the right choice for a project, just because not many people in their office have readability in that language.

A much easier process would go something like this: Suppose my teammate has readability in language X, They can tag review comments as being readability related or not. Once someone has N CLs under their belt with readability related comments under some low percentage in the past M CLs, they are granted readability automatically, or have to submit just 1-2 CLs for a speedy readability process.