I'll toss $20-50 your way to bump up the priority on writing that knowledge down, only strings are it has to actually get done and be publicly available
This is an oft-repeated misunderstanding. DDR5 memory uses error correcting codes internally to correct on-die errors, but this does not defend against errors on busses between the DIMM and memory controller. For that the old scheme of extra chips to store additional ECC data is still the only way.
> It actually does protect against errors on access, but support is optional.
Can you explain how exactly the on-die ECC capability can help protect data in transit? What is the optional functionality you're referring to, if not the traditional sideband ECC achieved by adding another chip's worth of data lines to every channel?
Not really. ECC memory will have an extra ram chip, and store an extra bit per byte or so, for that error detection/correction. DDR5 only has error-correction bits added to the bus, regular DDR5 doesn't have extra chips/bits for error correction of the data while it is stored.
But also, what you really want is ECC that reports all the way up to the OS the corrected and un-corrected bits. This is how you know if it's on the edge, becoming a real problem. Otherwise, it works fine until it doesn't shrug which is the same as regular normal memory.
I think the ECC added to the DDR5 bus is kinda just enough to get the higher data-rate signaling to be as reliable as DDR4. It's nice for marketing to put ECC on the DDR5 box but it's not more robust than DDR4.
The best thing that could come out of jujitsu is git itself adopting the change-id system (which I believe I read somewhere is being considered). If you actually take time to learn your tools and how they're intended to be used, there's really not reason to learn jj IMO
git is both a (bad) UI and a protocol. Jujutsu is a UI on top of git (the protocol).
There's nothing wrong with taking the time to learn how to use a bad UI, especially if there's no other option. But don't mistake your personal mastery of git for evidence that it's better than jj.
In all likelihood, the git proposal you allude to would not extend further than adding a bit of persistent metadata that follows commits after "destructive" changes. And even then, it'd be imperatively backing into the change-as-commit-graph data model rather than coming by it honestly.
> If you actually take time to learn your tools and how they're intended to be used, there's really not reason to learn jj IMO
This is like saying if people take the time to learn curl, there's really no reason to learn Firefox.
And it doesn't suggest to me that you're all that familiar with jj!
- automatic rebasing! goodbye to N+1 rebases forever
- first-class conflict resolution that doesn't force you to stop the world and fix
- the revset/template languages: incredibly expressive; nothing like it in git
- undo literally any jj action; restore the repo to any previous state. try that with the reflog...
No amount of learning git nets you any of these things.
Yes - but only a bit! And personally I still like it as a reference point when I tell people about it - because in the end, it mainly affects what they see of the repo and how they interact with it, and the result is still a valid git repo (if they've colocated).
True. People define the speech of others as violence because they think it makes a violent response into self defense. It isn't true and never has been. If you're responding to speech with violence... you're the baddie.
I think packetlost knows that. I think the argument being put forward was that "if you are the kind of person that thinks that speech is violence, then you would believe that allowing someone a platform..."
What is this for kind of logic. So if you have a store, and you happen to have a client that smashes someone else's head in your shop, then that means in your logic that you as a shop owner are guilty of violence.
But I guess you say you should know the reputation of every person coming into your shop, and if their reputation is deemed inappriopriate by a certain group, they should not be allowed into the shop to prevent them from harassing any other customers.
But how are you gonna regulate that? Who is gonna decide who is inappropriate and who isn't?
I think we already have a fairly well organized system for that: law & order. If someone breaks laws, they are punished for it. So if someone is violent, whether it's inside or outside a shop, they can be punished for it.
And you as a shop owner don't have to also individually take the effort to investigate and punish the individual. Although if you want to you have the freedom to; it's your shop in the end.
Speech is not violence, guilt by association is undemocratic, and this hypothesis of de-platforming as a tactic to limit uncouth ideas was thoroughly tested over the last ~15 years and demonstrably shown to be false: Trump, Alex Jones, and many others were banned across platforms. One of these people now sits in the White House, in part because of backlash to the deplatforming of him and others with similar politics.
It cost $44B to get him unbanned so I think that's actually pretty good evidence it worked.
> One of these people now sits in the White House, in part because of backlash to the deplatforming of him and others with similar politics.
It's not because of anything. Cause and effect doesn't apply to the brain of the median American voter - they live in a world of pure imagination. You could say they thought prices would go back down to 2016 levels, but that makes too much sense. If you look up what they actually think it's like "I voted for Trump because I want to protect abortion".
> It cost $44B to get him unbanned so I think that's actually pretty good evidence it worked.
Good evidence that it worked to do what? Limit his influence and popularity? This is false. His unbanning had little effect besides the right wing giving Musk brownie points, but the initial ban fueled grievance politics and became a huge rallying cry for the right. It was an extraordinary backfire.
> It's not because of anything. Cause and effect doesn't apply to the brain of the median American voter - they live in a world of pure imagination.
I flatly disagree with this. Human beings are endlessly deep and complex. The extremes of the internet cause us to group people together and create 1-dimensional strawmen of them, but if you talk to any American voter -- offline and 1-on-1 -- you will find complexity, nuance, and surprise in their opinions. At least, that has been my experience, with a pretty decent sample size.
Edit: I've been loosely watching the score on these comments, and it's interesting to see how rapidly it fluctuates up and down. For those that disagree, please leave a comment. IMO what I wrote is pretty common sense and moderate, so I'm interested in hearing disagreements.
> and it's interesting to see how rapidly it fluctuates up and down.
I've noticed there are usually wild swings depending on the active timezone. It would be interesting to try to extract a rough sentiment of each longitude, by looking at the timing.
There's a workaround for this case (relevant issue has a link at the top), which is cool, but it uses an "internal" function to solve it, which is not.
Because it seems that microsoft could not shitify windows experience anymore.
I like windows, Its a great system specially for being productive, but the godamn start menu using react and edge and the online requirements are a pain in the ass.
Sometimes it just hangs while you click the windows key. All I want is to open notepad++...
This is the story of Windows since 7 (and even earlier if you used the crappified Windows typically included with hardware): "The default experience is dogshit, but with enough work you can fix it." With every release the work required to make Windows bearable increases.
Microsoft to push pwa outlook as the default client also is terrible. Why would I want a e-amil client that occupies 3 times the memory of the default outlook client?
WSL... filesystem? Either way, I firmly disagree, there are not many cases where I've been unable to do dev work on WSL. Only when I need particularly weird / specific networking or hardware (ie. GPU, which might work now) have I had significant problems.
File IO from windows into the wsl disk and vice-versa is significantly slower so it's not great to, for example, use wsl git on a project living under your windows user directory or visual studio on a directory under your wsl home.
I think they're just using FUSE to make it work but don't quote me on that part.
Last I heard, using native git on Windows was slow anyways? Something about how NT handles files. So even with git it was already best to keep things inside WSL and just use ex. VS Code with a remote in WSL, and at that point it doesn't matter if you use radicle in WSL instead of git.
It's used as the index for a simple KV store I did as an interview problem awhile back, it pretty handily does 500k inserts/s and 5m reads/s and it's nothing special (basic write coalescing, append-only log):
reply