|
Yaron Minsky
@
yminsky
|
|
Occasional OCaml programmer
|
|
|
5.452
Tweetovi
|
165
Pratim
|
6.561
Osobe koje vas prate
|
| Tweetovi |
| Yaron Minsky proslijedio/la je tweet | ||
|
Richard Godbee
@areuugee
|
2. velj |
|
It's the year 2020, and we're still finding buffer overflows in programs written in C.
Fortunately, those programs are only esoteric, rarely installed programs like...
<checks notes>
sudo.
|
||
|
||
|
|
Yaron Minsky
@yminsky
|
21 h |
|
A mix. Some of them are projects with my kids, which aren't proprietary, but I don't want them to be public either...
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
2. velj |
|
A big issue with git is that its design prevents you from abstracting over the implementation. The store is the API, which means that your only shot for doing things that scale better is to use a VFS that exposes a faked-out store. And even that has limitations.
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
2. velj |
|
You're obviously right about the usage trends. I think hg's UI is way better than git's, and it is more hospitable to scaling. Google's narrow cloning is a big deal in that department.
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
2. velj |
|
Sigh. Just deleted all my repos from bitbucket, migrating them to git and github. The dropping of Mercurial support is quite sad.
|
||
|
|
||
| Yaron Minsky proslijedio/la je tweet | ||
|
ShriramKrishnamurthi
@ShriramKMurthi
|
1. velj |
|
First they came for Meta Language
And I did not speak out
Because I used OCaml... twitter.com/laurence_diver…
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
31. sij |
|
Thanks! Nice summary.
And that's an interesting solution to the "software engineering courses are horrible boring" problem.
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
30. sij |
|
This is totally unreasonable to ask for on Twitter, but can you explain it here? I'd love to know.
|
||
|
|
||
| Yaron Minsky proslijedio/la je tweet | ||
|
Andrey Mokhov
@andreymokhov
|
29. sij |
|
This year the @PLDI conference hosts the first Build Automation and Programming Languages workshop: pldi20.sigplan.org/home/bapl-2020
If you are coming to PLDI or just happen to be in London on Saturday (20 June) morning, please attend and/or submit a talk!
|
||
|
|
||
| Yaron Minsky proslijedio/la je tweet | ||
|
natefoster
@natefoster
|
25. sij |
|
I had to miss the first lecture of cs3110.org, so I got some friends to tell my students why they love functional programming. Enjoy!
youtu.be/Fi83KOAqWb4
|
||
|
|
||
| Yaron Minsky proslijedio/la je tweet | ||
|
|
natefoster
@natefoster
|
22. sij |
|
Wednesday afternoon hipster topic session at @poplconf pic.twitter.com/kbpF52iB2i
|
||
|
|
||
| Yaron Minsky proslijedio/la je tweet | ||
|
Edwin Brady
@edwinbrady
|
21. sij |
|
I enjoy using dynamically typed languages (especially Scheme) from time to time. I feel like I take a type driven approach when I do, but there's some kinds of exploration when it just feels like a nice way to work. twitter.com/lexi_lambda/st…
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
19. sij |
|
Yeah, I coined the phrase for this talk. It was "make illegal states unrepresentable", which I liked because it sounded vaguely like a political slogan.
blog.janestreet.com/effective-ml-v…
|
||
|
|
||
| Yaron Minsky proslijedio/la je tweet | ||
|
Patrick Walton
@pcwalton
|
10. sij |
|
I think the biggest GC-related performance benefit of value types isn’t that you can allocate them on the stack, since a generational GC will easily eat up those allocations. It’s so that more *long-lived* data can become pointer-free and not scanned at all.
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
11. sij |
|
I'm not sure if this amounts to a project suggestion, but there's a lot of PL and networking work to be done there.
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
11. sij |
|
Now that you have all these tools for managing complexity, you can add heterogeneity and structure to your network layout, including cleverly placed middleboxes, accelerators, L1 switches, you name it!
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
11. sij |
|
The next step is to reduce duplication and increase structure by using metaprogramming techniques to generate your network config generation.
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
11. sij |
|
The next step is to invert control; make changes to the VCS, and push them out to the switches from there. Now you can use static analysis (and code review!) to prevent mistakes.
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
11. sij |
|
You can now start doing static analysis on your network configs! Which can notify you of mistakes, but only after you make them.
|
||
|
|
||
|
|
Yaron Minsky
@yminsky
|
11. sij |
|
By which I mean: assume your starting state is the traditional network, with network admins manually configuring their standard switches. The first step is to build a system that captures all of those configurations and stores them in a VCS.
|
||
|
|
||