The history and endurance of vi is impressive. I never thought I would be using the same editor today that I started using in the mid 90s because it was more l33t.
The comments about LLM contributed code seems like a specific axe to grind that otherwise detracts from a nice history lesson.
yes, but that could've been part of the explanation in vim-classic. Other attributes and reasons for improving/forking are explained in the entry of the fork.
I started learning vi around the same time, but in my case (since I was expecting to work on Unix systems for decades, which has proven true) it was "because it'll always be there." I.e. if you're SSHing into a server to fix a problem, it's possible that /usr/bin/emacs won't be there (perhaps the problem you're fixing is that /usr isn't mounting), but you can nearly always count on /bin/vi being available: if you can access the server at all, you will be able to access vi, so at least learn its basic keystrokes, our prof told us.
That advice was not entirely accurate (sometimes vi is in /usr/bin/vi, for example), and the merging of /bin with /usr/bin has made it kind of a moot point. (EDIT to add: Though the fact that busybox includes a basic vi implementation has kind of un-mooted the point, actually). But I first started learning vi because I figured I would need it professionally, and when the modal-editing workflow "clicked" for me, I figured out that I had just learned the editor I would want to stick with for years.
And although vim replaced vi and nvim replaced vim in my finger-macros, that has remained true to this day.
Never really learned emacs so didn't know TRAMP existed. When was it created? I was given that advice ("vi will always be available on the server") in the late 90's so I'm curious to know if TRAMP was an option my prof didn't know about (or didn't mention), or whether it was developed later and the advice was good at the time.
EDIT: Found http://www.fifi.org/doc/tramp/tramp-emacs.html which mentions that TRAMP started development in November 1998. I would have been getting that advice in late 1997 or early 1998, given when I started my Unix class at college. So the answer appears to be that the advice was actually correct at the time, but superseded sooner than I thought it was.
Well, TRAMP was preceded by ange-ftp, which let you edit files remotely over ftp. I was using that in 1995, and I didn't get the impression then that it was brand new, so it had probably been around for a while already.
Of course, if your problem is "/usr won't mount", then it's likely that the ftp server isn't running either, so the advice still makes sense.
Yes, I can use TRAMP but as I ssh to the server anyway to run commands, I'm editing the files with vi there. Furthermore I'm sure I don't inadvertently edit the local version of the file instead of the remote one, or that I forget to kill the buffer with the remote file and edit it instead of the local one after a few days. What's on the server stays on the server.
I often use vim despite knowing only a few basic commands because I'm already running various commands on the remote machine in the terminal and don't want the context switch. And because connection / file path completion in TRAMP is annoyingly slow if the server is far enough.
As an extreme example, today I needed to combine parts of two files into one and decided that
cat foo bar > foobar && mv foobar bar && vim bar
of all things will better keep me in the flow than either googling how to insert one file into another in vim or starting up TRAMP.
The inclusion of comments about LLM generated code don't bother me and will probably be quite revealing (for better or worse) when people read this post in the future.
Also, I have not been following the d2d development of vim closely after Bram's passing but I can't help but wonder what he'd have thought about this approach to development of vim.
Yes, it almost feels like inserting politics to an otherwise apolitical discussion. Just unnecessarily charges the discussion and makes one not want to engage with the article.
The appeal of GraphQL is that it eliminates the need for a BFF and easily solves service meshing. Over fetching is more of a component design problem than a performance issue.
Right it felt pretty bad. It chugs tons of tokens just to be like "I need to scroll up!". Then 5 seconds later it scrolls up, chugs more tokens. "I need to scroll up more!"
Google would control everything if they could, but this won’t achieve that and they know that so it’s not the specific intention of this. Even if you’re feeling doomerish about it.
My theory is that Google wants to bake Gemini into Chrome to preempt a future antitrust ruling ordering them to spin the browser out, for the same reason Microsoft made IE an integral part of Windows 98.
It’s how a lot of code is being generated by owners of small startups with minimal engineering engagement. PMs are producing full walking skeletons of designs that used to take a week to polish to that fidelity.
Sure--but it's also the most widely used IDE for integrated AI assistance to normal software engineers. It's a "vibe coding" app in the same way that a washing machine is a "sock cleaning machine." I mean, yes, it does that, but that's a small part of its designed and in-practice usage.
They used an internal fork delivered via MDM. There are no guarantees that Signal can make about the software running on those phones and per the reports it’s a lot of phones.
Even the oral surgeon seemed to only regretfully remove them when I handed over the forms.