Update : I dropped it in my project followed by s/bootstrap-3.1.1/bootstrap-3.2.0/g . Nothing broken. That's a delight !
Some of them were affecting me, so I am very happy that this release fixes them.
Whenever I see default Bootstrap styles out in the wild though, I get pretty disgusted.
Never trust a company that uses default Bootstrap styles in their production apps. It just shouts "I don't care" or "I don't know any better."
I've seen a bunch of stuff submitted here to HN that was nothing more than a landing page with a sign up form. I've rarely (if not never) signed up.
However, give me a page with a working proof-of-concept, I'll likely sign up regardless if it's built on default Bootstrap styling.
Reminds me of the days when people's biggest complaints about Reddit was that it looked bad.
Sounds like a pretty positive indicator to me.
Whenever a web app doesn't look different, complaints flood in.
Consistent styling on disparate web pages is a good thing, in my optinion.
The number of times I want to be 'surprised and delighted' by an interface can be counted on the fingers of one hand. Most of the time I'm trying to achieve something other than appreciate the genius of the designer involved.
For example - I think many of the templates here: https://wrapbootstrap.com/ are rather lovely. Do they elicit your disgust?
Reductio ad absurdum...
You can write a grid system in less than 10 lines of code with Sass. You can also do a typographic mixin in 5 lines of code then customise it as you need. You can also have a icon web font set only for <i> tag in 5 lines of code. What else is there? Forms. You can reset the forms in another 5 lines of code then add stuff for [type=submit] and so on.
Why do you need thousands of lines of code when you can have a pretty nice CSS framework in less than 50 lines of code? This costs you time, money and your site will take more time to render by slow browsers.
So your question can be reformulated as "Bootstrap contains functionality that I don't think is needed".
If you agree that's a reasonable restatement then you can start a much more fruitful discussion if you could give a few examples and your reasoning behind why each one isn't needed.
What are you claiming the extra "thousands - 20" lines of code are for? I'm curious to know whether I need them or not.
H5BP is still available and there's a neat little packager, http://www.initializr.com/, that lets you include the parts you want and preconfigure it.
One reason to use a template system (with version control/plugins so you can modify locally and still merge in upstream changes and such) is that you can spend an awful lot of time keeping track of browser changes, standards changes, security issues, accessibility best practice, wide cross-browser testing and all the other niceties that go in to even a default template.
The idea that using Bootstrap cost me time is objectively wrong, as is that of it costing me money. Currently, my minified CSS and javascript from bootstrap is not affecting the page load time in any significant way. For me, using Bootstrap for MVPs has proved supremely useful, time and time again. I would not be as bold as you and claim my experience is applicable to everyone, but it is obviously very beneficial for me.
Something not often considered by the haters is that most of the time, the customers that the MVP is directed at have no clue about Bootstrap. Unless I'm building something for developers, it's usually not something they have seen before. I've actually gotten many compliments on design, when it was largely just Bootstrap's CSS.
Further, putting some structure in for SASS typography isn't the same as creating good typography. I've found it extremely difficult to create readable, custom typography in the past and I'm quite sure I am not alone.
Your "framework" gives me some kind of grid system, some structure that I can go out and create my own typography system inside of and some kind of functionality to "reset" forms. And then I can add "stuff for [type=submit]". GREAT.
Things I used from Bootstrap 3 in my MVP created in one night:
1. Responsive grid system. 2. Typography 3. Colors 4. Notifications 5. Forms 6. Buttons 7. Modals 8. Panels 9. Wells 10. Image helper classes 11. Navigation Bar 12. Pagination 13. Popovers 14. Media Objects
Your idea for replacing Bootstrap falls so short of the mark it is laughable.
It's also not an all or nothing proposition, you can simply include the bits and pieces you want to keep the size down.
e.g., @import '../../bower_components/bootstrap/less/grid';
The most popular project on Github is : https://github.com/search?q=stars%3a%3E1&s=stars&type=Reposi...
Clearly there has to be something worth more than 50 lines of CSS.
Sure you can make your own hammer, and you can even use a non-hammer you find sitting right next to you. But you'll quickly find out that some non-hammer won't accomplish everything you think a hammer ought to accomplish. Bootstrap isn't just an invention/innovation — it is a framework.
Your question honestly is no different from "Why do we need frameworks?" And we just need to stop wasting out time asking these bloated philosophical questions that get us nowhere.
"Why do we need dictionaries?" Well, because if everyone went about their own way, they'd all re-invent the wheel a thousand times over, but we'd never see any cars.