Hacker News Reader

Comment by stormbeard | Hacker News Reader

There's not much context here that explains why this is noteworthy. Arena allocation by itself isn't anything to write home about- is anyone able to explain why this is special?

View on HN

The standard that you're describing is not really the standard that most HN'ers hold submissions to. The guidelines[1] suggest the following:

> On-Topic: Anything that good hackers would find interesting. That includes more than hacking and startups. If you had to reduce it to a sentence, the answer might be: anything that gratifies one's intellectual curiosity.

1 = https://news.ycombinator.com/newsguidelines.html

someone might find a boots-on-the-ground, in-context treatment of arena allocation particularly interesting, moreso than just reading about it in the abstract on wikipedia

someone might have never heard of anything of the sort

someone might be interested to see lispers present a use case where they care about reaching down to the finer points of memory allocation in a language that typically benefits from being above that; see also the wealth of info on the garbage collector in the ocaml documentation

the number of users on this website who might get some sort of kick out of this link is nonzero which is the generally agreed-upon criterion for being worth posting

Based on the number of posts I’ve seen about arena allocators on here the last month or so, I’d say the audience is much higher than nonzero.

I’m curious, what’s higher than nonzero? Since any positive, finite number is nonzero, are we looking at an ℵ₀ audience?

I knew there'd be at least one pedant! And I almost altered what I wrote to pre-satisfy you!

Begone pedant, and feast thine eyes on the surreal numbers [0]!

> In mathematics, the surreal number system is a totally ordered proper class containing not only the real numbers but also infinite and infinitesimal numbers, respectively larger or smaller in absolute value than any positive real number.

(Also, go look at the `fil` unit in TeX for a good practical application of this ;).)

[0] https://en.wikipedia.org/wiki/Surreal_number

Region-based allocation by itself is definitely something to write about.

https://en.m.wikipedia.org/wiki/Region-based_memory_manageme...

The SBCL implementation is a very self-documented example, which is helpful if you wish to make arena-based allocation available in your language.

That Wikipedia article is confusing to me, because I’m really used to “region” being used to refer to type-system-level allocation tracking, Tofte&Talpin style (even if the T&T hope of inferring everything didn’t really work out, and the Rust tradition decided to rename the concept to “lifetimes”). “Arena” seems to be a more common term for a purely runtime construct.

i'm not sure how it's going to work in CL application code, but if you can re-use fixed memory and avoid runtime allocation with it like you can in C, it seems like it would be really nice for lisp game development, where avoiding GC pauses can be unintuitive or annoying

re-using memory is old trick in Common Lisp application code.

This is more "create a separate arena where we can allocate a ton of stuff and then drop it without actually going through collector".

In fact, this is very similar to features provided in at least Symbolics Genera, which provided developer-created arenas as well.

there have been other posts related to Arena based allocation on hn recently and I was grateful because it wasn’t something I was aware of. I found this interesting - I see this post is part of a trend.

Right, how is this top of HN? Something is off, as there is nothing newsworthy about this. If I submitted to HN a random documentation page from LWN or Linux kernel, it would surely not be #1.

There's nothing edifying about yucking other people's yums. If many others are upvoting something, and you don't see why it's interesting, maybe it's OK for the situation to be as simple as that: it's interesting to others even though it isn't interesting to you.

Also, there is absolutely no requirement that HN submissions all qualify as newsworthy. As others have pointed out in this thread, the gist of what's considered appropriate is quite simple and broad: "Anything that good hackers would find interesting."

> yucking other people's yums

that is gross please do not refer to sbcl memory management features as 'yums'

> as there is nothing newsworthy about this

Since when do HN submission have to be "newsworthy"? People regularly post old articles.

> If I submitted to HN a random documentation page from LWN or Linux kernel, it would surely not be #1.

If that page is particularly interesting and you're lucky, why not?

LWN write-ups on fairly niche subjects are regularly featured on HN front page.

Reply on news.ycombinator.com