Hacker Newsnew | past | comments | ask | show | jobs | submit | emddudley's commentslogin

Pyrefly vs. ty: Comparing Python’s Two New Rust-Based Type Checkers (2025-05-27) https://blog.edward-li.com/tech/comparing-pyrefly-vs-ty/

HN discussion of above: https://news.ycombinator.com/item?id=44107655

How Well Do New Python Type Checkers Conform? A Deep Dive into Ty, Pyrefly, and Zuban (2025-08-29) https://sinon.github.io/future-python-type-checkers/


amazing reply. Thanks!


A note - the second link talks mostly about conformance with a standard suite of tests, only briefly touching on real-world use.

I would very much like to understand how good Zuban is today compared to the competition.


There's an old XML one called Data Format Description Language (DFDL).


Short form video is tailored for the fast, instinctive, emotional brain. Not the slower, deliberative, logical brain.


Ironically, PFAS levels have been found to be higher in wealthy people. People with money own more furniture and clothing with stain resistant treatments, for example.


Also brand new items versus used items. When you buy a used item, someone else has already absorbed the PFAS, and the depreciation for that matter.


Not related to PURLs (Persistent URLs) administered by the Internet Archive.

https://purl.archive.org/


Or PURLs in general, the concept for which was developed in 1995, per https://en.m.wikipedia.org/wiki/Persistent_uniform_resource_...


Nor to the Purl programming language: https://esolangs.org/wiki/Purl

I wonder if Yarn will support PURLs. ;)


I remember when purl.org namespace URIs where the thing for RSS 1.0 modules. 25 years ago,


Where can I read more about this?


We maintain the spec at https://github.com/package-url/purl-spec

And the new thing, working towards making it a real standard with Ecma https://tc54.org/purl/ ... :)


Not at all related. Just nicknamed the same.


Similar to the cited "Do you Need Blockchain" there is NISTIR 8202: Blockchain Technology Overview, page 42.

1. Do you need a shared, consistent data store?

2. Does more than one entity need to contribute data?

3. Data records, once written, are never updated or deleted?

4. Sensitive identifiers WILL NOT be written to the data store?

5. Are the entities with write access having a hard time deciding who should be in control of the data store?

6. Do you want a tamperproof log of all writes to the data store?

https://nvlpubs.nist.gov/nistpubs/ir/2018/NIST.IR.8202.pdf


I think the interesting interplay with this is in "need" vs. "want".

I share the conventional wisdom that nothing, or very nearly nothing, needs these properties. But I don't think it's unreasonable that people have been looking around for things that want them.


Why would you voluntarily want a data store that can't store sensitive values?


I don't know, but I also think it's reasonable for people (who aren't me) to see if they can figure out use cases where they do want that, and actually make them work. I'm not particularly bullish on any of these projects, I just also think that thinking outside the box in this way is one of the ways that interesting things come about.

I'm personally pretty happy to be conventional and milquetoast, but I think it's fairly self defeating to be stuck in the box of "only projects that work on traditional database architecture could ever make sense", so I'm glad people are out there trying different things, even when I think they're pretty likely to fail.


I can think of two reasons:

1) The data store offers some other interesting property that necessitates this tradeoff.

2) You want your data store to be transparent/auditable by anyone.


> 3. Data records, once written, are never updated or deleted?

This is not a requirement for any blockchain with smart contracts.


PURL (https://purl.archive.org/) is a similar permanent URL service but you choose the URL.

It used to be hosted at purl.org and run by the OCLC but in 2016 it was transferred to the Internet Archive.

https://web.archive.org/web/20161002094639/https://www.oclc....


PURL is in the same space as w3id.org, not perma.cc. Purl and w3id work by creating stable URLs thar can redirect to a (potentially changing) origin, perma.cc/archive.org/archivebox create WARC archives or the content at a given instant.


You've been fooled.


I'm embarrassed to admit it but I found out this game has an ending.


Yahoo (Verizon) did eventually sell Flickr to SmugMug in 2018 for an undisclosed sum.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: