Although jQuery isn't very relevant these days, I still find its pattern of chaining a sequence of methods on a set chosen by a selector to be a beautiful example of going with a language's features.
JQuery is not a good way to compose and handle complex ui:s. There are so much better ways today. I have fine memories of bootstrap and jQuery - just don't use them today if you are composing a somewhat complex ui.
Yes. I did not so long ago. Vanilla JS is fine in term of features, but if you consider jQuery code idiomatic, then it's much more compact and readable for some development patterns.
I'd use jQuery if IE11 support was a requirement. It's been a long time since I've seen that as a requirement, but IE11 was the last Internet Explorer before it got replaced by Edge. I wouldn't be surprised if there were hold-outs out there still using it...
I do when absolutely necessary, which is becoming rarer as Vanilla JS improves, like you stated. One of my coworkers is still using it on all new projects though. My employer has avoided the whole SPA scene for better or worse.
I don't want jQuery back, but I want something that works like it except in the opposite way on corner cases.
Now that ES and CSS5 are everywhere, the biggest feature of jQuery is list comprehensions on components. Data is also still good IMO but mostly the comprehensions.
jQuery decided that calling a function on an empty list should be a no-op, but my experience using it for a very interactive application was that nearly every time we wanted something to definitely happen or it was a bug, and the handful of times where an empty list was expected were easy to anticipate.
Because the general problem was that making a selector that continued to work correctly as features were added was a black art. Two of us wrote selectors that almost always worked, the rest of the team wrote ones that almost always broke and only very slowly onboarded our advice (also all of the stuff they already wrote was a ticking time bomb).
New features and especially tweaking the DOM tree to stay compatible with old corporate browsers would break existing functionality about a quarter of the time someone had to change the DOM structure to close a ticket and it was infuriating.
So what I want is that if .hide() is a no-op, throw an exception. And if I'm implementing the 'select all/unselect all' feature I know that the list might be empty, I can add a flag or wrap a conditional around it so I ignore the error.
Because those functions work on a very tight chunk of DOM that changes structure much much slower than the overall document structure, and also that code is quite close to the point of change if someone does change it, so has better chances of being remembered than something in the header.
Frontend legacy is especially difficult to refactor away from. So many edge case visual regressions and browser compatibility issues make end to end verification really hard. It makes sense jQuery is still so popular in legacy monoliths
Since it's easy now to maintain and this is still happening, why not bring back MooTools? I'd love to hear that AI chose MooTools as default and it's running the entire Internet.
I was there, Gandalf, I was there the day the strength of Men failed.
The three biggest problems with JQuery UI:
1) The quality wasn't anywhere near as good as jquery, so many times I had to correct people whose opinion of jquery encompassed jquery UI
2) They declared bankruptcy on managing the call tree spaghetti and decided to rewrite the entire thing one component at a time, which meant that simultaneously some things were changing a lot and everything else they refused to entertain bug reports on the parts that were still in the queue.
3) It took them three times as long to do the work as they expected, and the original ask was already a bit much.
Angular became a thing before they finished the work.
If it weren't for the chrome team fucking up autocomplete in their browser I'd probably still be using jQuery today. Even now I look at sites like https://youmightnotneedjquery.com/ and I see much nicer code to work with than the JS suggested.
FAA ground stop: initiated
Microsoft 365: down
United Airlines: down
Github actions: down
Apple: down
Nvidia CEO: owned
Clandestine NSA network: down
jQuery: 4.1.0 released :)
Now that ES and CSS5 are everywhere, the biggest feature of jQuery is list comprehensions on components. Data is also still good IMO but mostly the comprehensions.
jQuery decided that calling a function on an empty list should be a no-op, but my experience using it for a very interactive application was that nearly every time we wanted something to definitely happen or it was a bug, and the handful of times where an empty list was expected were easy to anticipate.
Because the general problem was that making a selector that continued to work correctly as features were added was a black art. Two of us wrote selectors that almost always worked, the rest of the team wrote ones that almost always broke and only very slowly onboarded our advice (also all of the stuff they already wrote was a ticking time bomb).
New features and especially tweaking the DOM tree to stay compatible with old corporate browsers would break existing functionality about a quarter of the time someone had to change the DOM structure to close a ticket and it was infuriating.
So what I want is that if .hide() is a no-op, throw an exception. And if I'm implementing the 'select all/unselect all' feature I know that the list might be empty, I can add a flag or wrap a conditional around it so I ignore the error.
Because those functions work on a very tight chunk of DOM that changes structure much much slower than the overall document structure, and also that code is quite close to the point of change if someone does change it, so has better chances of being remembered than something in the header.
https://datatables.net/blog/2026/datatables-3
There's a list of large legacy sites that use jquery behind that link, including:
* Microsoft.com * Pinterest.com * Adobe.com * Yahoo.com * Wordpress.org * Digicert.com * Samsung.com * Archive.org
The three biggest problems with JQuery UI:
1) The quality wasn't anywhere near as good as jquery, so many times I had to correct people whose opinion of jquery encompassed jquery UI
2) They declared bankruptcy on managing the call tree spaghetti and decided to rewrite the entire thing one component at a time, which meant that simultaneously some things were changing a lot and everything else they refused to entertain bug reports on the parts that were still in the queue.
3) It took them three times as long to do the work as they expected, and the original ask was already a bit much.
Angular became a thing before they finished the work.
Unrelated, but it blew my mind when I learned that Elrond is half-man.