Bootstrap's 'one-line-of-code-and-go' simplicity enabled me to make my first websites with no IT education on library computer in rural Australia [we didn't have one at home]. Today, I write code for a SF-based unicorn and whilst I don't use Bootstrap at work, I attribute my career to that CSS framework.
'Easier' tools enable more people to make use mankind's collective knowledge and change their lives for the better; I don't give a fuck if you're happy about it.
You look and see - "oh, this requires sass. sass requires some libs, or npm, or something else. Now I need to know how to go install those (assuming I'm on a machine I can install software on right now), then make sure it's all downloaded and installed correctly (because there are multiple ways to install it 'wrong'), then learn how to use those tools, then learn how to use them well enough to make something decent, the learn how to bundle/deploy that to my website."
You're not saving anywhere near "20 seconds" using bootstrap CDN vs the alternatives (and frankly, not even using bootstrap itself, if you pull it down and configure and compile all its options as well). You're saving hours or days of learning up front, and focusing on getting something basic up and running, then determining what style/ui changes are required later.
But keep making the false comparison between "downloading" vs "learning portion of multiple ecosystem and technologies" - you're coming across as quite learned here.
When the airline industry made a major effort to improve safety elitism was something they explicitly challenged.
Just something to think about.
The entire reason why I got into programming was to eliminate the need for "minor chores". I want to tell the computer how to do it and then never think about it again (mostly). Bootstrap isn't perfect, but it's convenient and easy enough to use that I haven't dug too deeply into alternatives.
The way I think about it: any extra effort that has no direct impact in making my life "as a programmer" easier, is a bug that needs to be fixed. Same thing applies when designing for people (i.e non technical users)
Here's an inspiring interview from Alan Key that illustrates the issue: http://www.drdobbs.com/architecture-and-design/interview-wit...
And 20 years later I am able to differentiate between "minor chores" and real "real work savers".
And downloading something is definitely a minor chore, one that can be "fixed" by a few lines of code in your build pipeline. After having read a few discussion here on HN I totally get that this is "too much" for some (just revisit the discussion about the npm left-pad disaster where the consensus on HN seems to be that it "is ok to download code for that because writing your own routine for left-padding a string might take an hour"). When I was learning, I'd spend that our in my spare time because I'd be ashamed to have somebody pay me for this. Now that I am older, I'd just fire whoever utters such crap on the spot. To be perfectly clear: "I don't have to download the libs!" is NOT a valid reasoning for a technical choice. Not at all.
In the last 20 years of professional software development I've never ever encountered a situation where the time "downloading libraries" was worth more than a rounding error in a weekly budget.
PS: man alias. trust me, just a single line of config in your shell will save you hours, well minutes, well, maybe just seconds over your lifetime. But you can spend twice as much time to make the non-problem go away - and you qualify for a free shirt "for every solution, I have the problem!".
EDIT: sp4ke, I know this sounds harsh, but it is not, in any way, meant as a personal attack. Instead, please treat it as the ramblings of an old, misanthropic programmer ;)
"one that can be "fixed" by a few lines of code in your build pipeline"
Why do I even need a "build pipeline" for some projects?
If you actually think much of this npm/gulp/sass/build stuff will age well, and you'll be able to go back to a "build pipeline" for a project 7 years from now.... it's very likely not going to work.
"In the last 20 years of professional software development I've never ever encountered a situation where the time "downloading libraries" was worth more than a rounding error in a weekly budget."
Hrm... just last week - "npm install" on a project took 45 minutes. Then... welp, gotta do it again on the server too. 345 meg of stuff, multiple dozens of which were complete duplicates of each other. That's 90 minutes waiting, 45 of which not even sure how much time is left (2 mins? 5 seconds? an hour?) while the machine slowed to a crawl.
"I don't have to download the source libraries, then make changes, then recompile, and build and publish and push just to get a half-decent default when I can just reference these over here from a CDN" - that certainly is a valid technical choice. You can spend hours downloading, configuring, learning something that wasn't bootstrap and delivering less than "way better than bootstrap" UI/UX... that's a bad technical choice.
You're right about aliasing commands, building stuff to save yourself time, but avoiding prebuilt stuff that is perfectly suitable for a project in pursuit of pixel-level control when it's not necessary (for most projects) is wasting time.
I'd hope that people with misguided optimization agendas that effectively cost me as an employer money will be weeded out before 20 years pass.
The QED is the rest of your post; I don't care about your problem with NPM, I neither mentioned it nor do I used it beyond amusement purposes.
The commenter who you said had "by far the worst line of reasoning I've ever encountered" was annoyed at having to use NPM to install a simple CSS library. Not only did you flat out insult the person, you did so without taking seriously the tech stack that caused the headache! How do you expect to contribute to a discussion of the pros and cons of new tech when you yourself aren't using it to do real work?
Of course it's not right for every project, but it's perfect for a lot of them.
This is why yak shaving is a thing we had to name. In isolation none of these things is hard or even a big ask. In combination, you end up diving into how PATH is set in your shell because you added npm's bin dir to it but you've actually added the wrong one because you didn't realize you accidentally installed it through both curlpipe and homebrew and (add ten more steps of complications here) just to tinker around with a freaking static HTML template.
I need a minimally responsive site that doesn't look like ass, where I can possibly change a couple colors, but things basically look good to go out of the box.
Bootstrap was a fucking godsend in 2011 compared to everything else out there at the time. I can have something up that fits a client's requirements (something that doesn't look ugly) and spend more time on their actual functional requirements vs spending bullshit time downloading 200,000 NPM files and fucking with dependency hell just to have grunt and gulp and sass and less and all that other bullshit work.
I delivered line of business apps that looked decent, and frankly still look pretty decent (BS2 has held up OK, and BS3 is still decent even without any tweaking).
And guess what? For many many many types of web software that's delivered, Bootstrap is GOOD ENOUGH. It's a 'minimum viable css grid', where "minimum" and "viable" include "not looking like crap that I have to spend hours tweaking".
Any of that reasoning encounter you any better?
Furthermore more, Bootstrap made "mobile first" easy to embrace. It certainly helped to usher in RWD.
They're not trivial things if they're adding to the complexity already present there.
For example, you may not in fact use node for anything else in the project (eg. it's a Django or Rails project). If you now need to depend on node just for this one thing, the marginal cost/utility is non-trivial.
These things add up, and being aware (and aversive) to this is not whining.
All they're saying is that in terms of popularity, having lower barriers to entry helps.