Profile avatar
julianhyde.bsky.social
On a mission to tame data. (ASF & Google.)
134 posts 372 followers 64 following
Regular Contributor
Active Commenter
Posts 19 Comments 31

How to create data frames in a statically typed language? Morel's answer is progressive typing. www.youtube.com/watch?v=uybU...

submitted 11 hours ago • 1 comment

For some reason we all stopped doing meetups for a few years. Let's fix that. See you in Santa Clara on Thursday. www.meetup.com/futureofdata...

submitted 6 days ago • 0 comments

This excellent post identifies some huge flaws in a "study" by the MRC which doesn't interrogate why right-wing sources like OANN and Newsmax are deprecated (not banned) as sources on Wikipedia, and ignored right-leaning sources that are considered reliable and are widely used.

submitted 11 days ago • 5 comments

"In the future, terminals will be so wide that we won't have to worry about line length" We'll have flying cars before that one comes true. In the mean time, let's start teaching coders how to split lines for legibility.

submitted 13 days ago • 0 comments

I’m spending a lot of time reading about overloading. If Morel queries are to accept both lists and bags, and simple queries with only lists are ordered, type derivation gets complicated. The solution may be an internal “join” function overloaded for lists and bags. dl.acm.org/doi/pdf/10.1...

submitted 15 days ago • 0 comments

I just spent a couple of days rebasing an incomplete PR that I wrote in a couple of hours last year.

submitted 19 days ago • 0 comments

Introducing the “Cheese Jar”. Now I can keep my Taleggio or Gorgonzola in the fridge, and let it sit on the counter for an hour before lunch coming to room temperature, without repulsing the family. Consider it my contribution to western civilization.

submitted 19 days ago • 0 comments

Number one design goal of a cereal bowl: 16 of them (family of 4 for 4 days) fit easily in the upper rack of a dishwasher. Also important: holds breakfast cereal.

submitted 19 days ago • 0 comments

Outlier AI. You are doing it wrong. Hiring people to post completely nonsenese or copy&pasted issues in reputable open-source repositories - and make maintainers train your AI on it ? not good. There are 50 such issues in last few days in @airflow repo [1] and counting. More details in [2] […]

submitted 29 days ago • 0 comments

Wordle 1,316 3/6* ⬜🟨⬜⬜🟨 🟩🟩⬜🟩⬜ 🟩🟩🟩🟩🟩 Today’s word is an English food, yesterday’s was a French food, and both derive from the same Latin word.

submitted 30 days ago • 0 comments

Spotify just hit me with a 'sponsored recommendation'. They used to call that 'payola'. If my monthly fee isn't good enough for them maybe I should go elsewhere.

submitted 31 days ago • 0 comments

I've been using git since 2012 and I still don't understand what the index is or why it is called 'index'.

submitted 33 days ago • 3 comments

"Truly open source"

submitted 34 days ago • 0 comments

It’s fun to write your own implementation of a classic algorithm. Over the weekend I implemented union-find (packaging it as a Java Map whose keys can be merged into equivalence sets). github.com/hydromatic/m...

submitted 42 days ago • 0 comments

#WashingtonPost #Bezos #JeffBezos #Amazon #Cartoonist #AnnTelnaes #Zuckerberg #MarkZuckerberg #Meta #Facebook #Altman #SamAltman #OpenAI #OkloInc #HelionEnergy #Soon-Shiong #SoonShiong #LATimes #Oncologist Democracy can't function without a free press. Abridged (shortened) article thread ⏬ 3 min

submitted 51 days ago • 2 comments

Working on lists, relations and ordering in Morel has got me thinking about the information content of collections. We generally assume that data structures for various types of collections (sets, ordered sets, multisets, lists) use similar amounts of memory (plus a little overhead for hash /

submitted 58 days ago • 1 comment

Did Quincy Jones just say “Sing it Acapulco”?

submitted 58 days ago • 0 comments

Jeez. Wordle 1,279 5/6* 🟨🟨🟨⬜🟨 🟩🟩🟨🟨⬜ 🟩🟩🟩🟩⬜ 🟩🟩🟩🟩⬜ 🟩🟩🟩🟩🟩

submitted 67 days ago • 0 comments

I’m building a utility to help test Calcite’s many SQL dialects. Run your queries in ‘record’ mode against a database (say Snowflake), and it generates a file. Then you can run offline in ‘play’ mode, getting the results from the file. github.com/hydromatic/q...

submitted 67 days ago • 1 comment