HTML5 Bones: The template that goes back to basics(html5bones.com)
html5bones.com
HTML5 Bones: The template that goes back to basics
http://html5bones.com/
16 comments
Edit: I've made my first ever pull request on Github. I hope I'm doing it right:
https://github.com/iandevlin/html5bones/pull/2
---------
Looking at all the stuff that index.html includes I think the name is misleading. Google Analytics? I know HTML5 Boilerplate has it, but HTML5 BP is maintained by Google people.
Actually, HTML5 BP has an index page that's smaller than HTML5 Bones: https://github.com/h5bp/html5-boilerplate/blob/master/index....
I think that the index page HTML5 Bones has is rather bloated in comparison: https://github.com/iandevlin/html5bones/blob/master/index.ht...
And it seems like a bad template to start from, since I'd be erasing most of it, like the <aside role="complementary"> and abbr tags. In fact, most of that page is just explaining the page, which really doesn't belong in a template.
The project does, however, showcase how to properly use a lot of the new HTML5 tags, and I think that's commendable in its own right. I just don't think it makes the template as useful as it might otherwise be, especially if its purporting to be (bare) bones.
I'm probably just being way too harsh though.
---------
For the record, since I might as well be the first to pedantry, this is the smallest validating HTML5 page:
https://github.com/iandevlin/html5bones/pull/2
---------
Looking at all the stuff that index.html includes I think the name is misleading. Google Analytics? I know HTML5 Boilerplate has it, but HTML5 BP is maintained by Google people.
Actually, HTML5 BP has an index page that's smaller than HTML5 Bones: https://github.com/h5bp/html5-boilerplate/blob/master/index....
I think that the index page HTML5 Bones has is rather bloated in comparison: https://github.com/iandevlin/html5bones/blob/master/index.ht...
And it seems like a bad template to start from, since I'd be erasing most of it, like the <aside role="complementary"> and abbr tags. In fact, most of that page is just explaining the page, which really doesn't belong in a template.
The project does, however, showcase how to properly use a lot of the new HTML5 tags, and I think that's commendable in its own right. I just don't think it makes the template as useful as it might otherwise be, especially if its purporting to be (bare) bones.
I'm probably just being way too harsh though.
---------
For the record, since I might as well be the first to pedantry, this is the smallest validating HTML5 page:
<!DOCTYPE html>
<meta charset="utf-8" />
<title></title>
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8" />
<title></title>
</head>
<body>
</body>
</html>I'm not sure why people feel the need to capitalize DOCTYPE - the spec allows <!doctype html>. (Here's the exact text from the spec: '1. A string that is an ASCII case-insensitive match for the string "<!DOCTYPE"....')
While we're on the subject of pedantry:
1) Given the unclear use of "smallest", !doctype is smaller, because lowercase letters are smaller than uppercase letters.
2) Your newlines unnecessarily add two extra characters to the page (arguably three if you count the newline after the title string)
3) It's unclear what you mean by "html5 page". Assuming you mean "html document which conforms to the syntax of the most recent html5 draft published by the W3C", you're not actually allowed to leave out the html element.
...of course, I'm not actually sure what you mean by "validating" either, but you shouldn't use any particular validator on the internet, as it introduces a layer of indirection. Just use the syntax given by the spec itself: http://www.w3.org/TR/html5/syntax.html#syntax
As a sidenote, you could also use the syntax given by the WHATWG's living spec, but the shortest conforming html document turns out to be the same regardless. WHATWG syntax: http://www.whatwg.org/specs/web-apps/current-work/multipage/...
And finally, since I despise people who criticize without offering alternatives, here's my attempt at the shortest conforming HTML document: "<!doctype html><html><title></title><body></body></html>"
As a sidenote: I'm not sure, but it might be conforming to re-write empty tags as <tagname /> instead of <tagname></tagname>, in which case the smallest conforming document would be: "<!doctype html><html><title /><body /></html>"
While we're on the subject of pedantry:
1) Given the unclear use of "smallest", !doctype is smaller, because lowercase letters are smaller than uppercase letters.
2) Your newlines unnecessarily add two extra characters to the page (arguably three if you count the newline after the title string)
3) It's unclear what you mean by "html5 page". Assuming you mean "html document which conforms to the syntax of the most recent html5 draft published by the W3C", you're not actually allowed to leave out the html element.
...of course, I'm not actually sure what you mean by "validating" either, but you shouldn't use any particular validator on the internet, as it introduces a layer of indirection. Just use the syntax given by the spec itself: http://www.w3.org/TR/html5/syntax.html#syntax
As a sidenote, you could also use the syntax given by the WHATWG's living spec, but the shortest conforming html document turns out to be the same regardless. WHATWG syntax: http://www.whatwg.org/specs/web-apps/current-work/multipage/...
And finally, since I despise people who criticize without offering alternatives, here's my attempt at the shortest conforming HTML document: "<!doctype html><html><title></title><body></body></html>"
As a sidenote: I'm not sure, but it might be conforming to re-write empty tags as <tagname /> instead of <tagname></tagname>, in which case the smallest conforming document would be: "<!doctype html><html><title /><body /></html>"
Probably way too harsh - BP gives the minimum that you can try and plug your own code into.
html5bones is clearly taking a different route, by making everything and every tag as clear as possible. Everything is commented to give someone starting out a good explanation of why all this stuff is there.
eg. in HTML5 BP:
html5bones:
html5bones is clearly taking a different route, by making everything and every tag as clear as possible. Everything is commented to give someone starting out a good explanation of why all this stuff is there.
eg. in HTML5 BP:
<script>window.jQuery || document.write('<script src="js/vendor/jquery-1.8.3.min.js"><\/script>')</script>
html5bones:
<!-- An <aside> is used to enclose content that is additional to the
main content but not essential. If it were removed, the meaning of
the main content should not be lost, but the content of the <aside>
also retains its meaning. -->
<!-- ARIA: the landmark role "complementary" is added here as it
contains supprting information for the main content that remains
meaningful even when separated from it -->
<aside role="complementary">
Hi, I wrote HTML5 Bones and to be fair Simon has a good idea (detailed in the GitHub pull request) in that there should be two versions, a file with all the comments and one without - and so catering for both types of users.
I definitely don't want to remove the comments as they are integral to what I want the "template" to help people do.
Thanks for your comments so far!
I definitely don't want to remove the comments as they are integral to what I want the "template" to help people do.
Thanks for your comments so far!
Since you adding ARIA related info a small example for a table can be useful. Thanks for publishing it.
For the record, the smallest validating document is actually:
<!doctype html><title></title>
Although I can't right now see the part of the spec that allows an empty title element, which is either a bug in my reading, a validator bug or a spec bug.
Of course, declaring a charset is generally a good thing to do (arguably more useful than a <title> if you expect the document will be in an <iframe> for example) but even if that were required you could get rid of the entirely optional trailing " /" and save two characters. And strip the quotes around the attribute value for two more characters.
Although I can't right now see the part of the spec that allows an empty title element, which is either a bug in my reading, a validator bug or a spec bug.
Of course, declaring a charset is generally a good thing to do (arguably more useful than a <title> if you expect the document will be in an <iframe> for example) but even if that were required you could get rid of the entirely optional trailing " /" and save two characters. And strip the quotes around the attribute value for two more characters.
that is correct, this is the real html5 bones
web devs dont give a "beep" about tags, if they did , they would have supported XHMLT... they give even less "beep" about validation... And they are so lazy they cant even take the time to write their own CSS files...
I'm sort of lost on the idea of this and Boiler Plate. Is HTML really this difficult to use? I don't understand why these are less confusing than the very well-written WHATWG documentation. [1]
I'm not really sure who the target is for these HTML 5 frameworks: brand new HTML users with no coding experience? This seems like a target that is completely lost anyways, as you are now enforcing some idea that they really won't understand anyways. Yeah, I get the idea, but I won't use it because writing correct and valid HTML5 is not hard at all.
Twitter Bootstrap has a very simple mission statement: give those who can't design the ability to create some minimally nice design. Why? Because design is very hard. HTML isn't.
[1] http://www.whatwg.org/specs/web-apps/current-work/multipage/...
I'm not really sure who the target is for these HTML 5 frameworks: brand new HTML users with no coding experience? This seems like a target that is completely lost anyways, as you are now enforcing some idea that they really won't understand anyways. Yeah, I get the idea, but I won't use it because writing correct and valid HTML5 is not hard at all.
Twitter Bootstrap has a very simple mission statement: give those who can't design the ability to create some minimally nice design. Why? Because design is very hard. HTML isn't.
[1] http://www.whatwg.org/specs/web-apps/current-work/multipage/...
In a word, yes.
There is a lot of knowledge about optimisation, handling different devices, responsive design buried in these boilerplates.
There is a lot of knowledge about optimisation, handling different devices, responsive design buried in these boilerplates.
I'm frameworked out.
If only there were a framework for keeping track of all these frameworks!
This isn't really a framework though. It's more like a boiler plate or template that just gives you a good place to start. Maybe you are boilerplated out...
2012: The year of frameworks.
I love what you've done here. It's indeed a great starting template and unlike that -other- html5 template it does a good job explaining why you would or wouldn't use some of the elements.
I also disagree with some of the other commenters on the inclusion of GA code, yeah it could do without but i have never put a site online that didn't have that tag in it while i have released many that did not need Modernizr, jquery or some of the other fancy pants stuff in boilerplate.
Also i love the plug-and-play video template. Well done sir!
I also disagree with some of the other commenters on the inclusion of GA code, yeah it could do without but i have never put a site online that didn't have that tag in it while i have released many that did not need Modernizr, jquery or some of the other fancy pants stuff in boilerplate.
Also i love the plug-and-play video template. Well done sir!
It's not simple enough to be bare bones and not useful enough to bookmark, in my opinion. The idea is good, and the domain also.
Not quite the simple template I expected this to be, so I created my own: https://github.com/Vheissu/Actually-Barebones-HTML5-Template — Most of these HTML5 templates I use I strip 99% of the crap out anyway and only need the doctype, head and body tags.
A minimal HTML5 boilerplate of my choice: http://csswizardry.com/2011/01/the-real-html5-boilerplate/
I never really got why would anyone need things like image replacement or Google Analytics in a HTML5 template. The KISS principle comes to mind.
I never really got why would anyone need things like image replacement or Google Analytics in a HTML5 template. The KISS principle comes to mind.
Styling bug on homepage http://cl.ly/image/1s3G0w2K2T0Q
This might be a Chrome Canary Bug. Get it on a lot of pages where elements are floated. The n+1th element always has a margin on top. Couldn't figure out what the real issue is though.
If you are wanting to properly support IE7 (as mentioned in the README), you may want to downgrade normalize.css to 1.0.2 as normalize.css 2.x only supports IE8+.
There are only a few edge cases that are affected but they can bite if you were expecting full compatibility.
There are only a few edge cases that are affected but they can bite if you were expecting full compatibility.
If you are looking for a slim html5 boilerplate, Skeleton is a very good choice I highly recommend: http://www.getskeleton.com
This is nice, HTML5 Boilerplate was too overbloated
If it's really barebones, it should be small enough to copy and paste it from the site, I would think.
Right?
Right?
As this featured project is, in some way, defining itself relative to what it believes HTML5 Boilerplate is and should be, I'm going to chime in with some thoughts.
If someone wants to create a new project after their concerns have not been shared by others maintaining a large, existing, community project (e.g., how Lo-Dash came to be) then that's fair enough. But I'm always disappointing to see people do so before raising their concerns and making their arguments. It would avoid the repetition of errors and discussions that have been addressed previously. We could also work together to adapt HTML5 Boilerplate if there is a good case to do so.
Once-upon-a-time, I didn't appreciate the work and community knowledge that went into HTML5 Boilerplate. However, after I started to open issues against the repo and contribute to the project, I realised that it has always been very open to contributions and changes of direction. I had significantly underestimated how much thought and combined knowledge was going into the project, and how much it was constantly evolving.
It's hard to design a useful starting template that's clear and friendly enough for newcomers, documented enough for explorers, and svelte enough for experienced developers. We're always trying to figure out the best way to present the value of the project while not scaring away too many people. Hopefully, we're doing a decent job of balancing those concerns and making people feel welcome in the discussions.
I hope that developers will continue to work with -- and look to join -- the people developing HTML5 Boilerplate at any given time. If you think something is missing or should be removed, make the case for it. We listen and learn; we're constantly evolving the project.
Thanks
------
> HTML5 BP is maintained by Google people. http://news.ycombinator.com/item?id=4949018
That's not true. Only Paul Irish works at Google.
> "Slim HTML5 Boilerplate" would be a better title http://news.ycombinator.com/item?id=4949635
If that's what you're looking for, just delete jQuery and Modernizr from HTML5 Boilerplate :)