This question already has answers here:
Using custom HTML Tags
(10 answers)
Closed 9 years ago.
I've learnt HTML and the basics of CSS only recently and am enjoying the liberty of finally being web literate. One thing puzzles me, with the usefulness (literally and semantically) of elements in HTML5 such as
<footer>
etc why dont we just create and define our own elements such as
<price>
or
<icon>
instead of giving standard ones classes or defining divs all the time which seems confusing in comparison? It seems CSS has no problem styling these (in my tests anyway) like anything else providing I state whether display is inline or block etc.
Is this a totally standard thing to do I have just stupidly realised or am I missing a major downfall with this here?
Because:
Semantics only do any good if the user agent (browser, search engine, etc) understands them
If you make up your own elements and a set of semantics, then someone else (including the W3C) may introduce their own element with the same name but with different semantics and user agents would interpret your element "incorrectly".
Related
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 4 years ago.
Improve this question
The inverse question has been asked multiple times. However I could neither find any cons nor convincing proof of their complete absence.
I am all for progress and semantic documents. In fact I have started to use new semantic tags in my markups already a while ago (it feels nice!). Still I wonder if there are any (possibly obscure) caveats one should know about.
Will the rendering break (more) on that office machine which rolls prehistoric versions of IE?
backwards compatibility of css selectors like article?
strange side effects to the DOM?
...(anything)?
Some older browsers don't recognize the html5 elements. At best this means you have to define default style rules for those elements in your CSS.
In some cases you might need to apply a shim, or in this case, a shiv:
https://css-tricks.com/snippets/javascript/make-html5-elements-work-in-old-ie/
Again, old browsers (I'm looking at you, IE) won't even recognize them as elements, so yes, it would likely change the structure of your DOM. Also, I've seen old browsers just output the html5 tags as plain text, so you see the <article> tag right there in your content.
If you happen to be in a situation where you can't support the new html5 elements, one handy workaround is to use divs and give them classes that match the desired element names. So for example you'd use:
<div class="article">
instead of
<article>
This approach has the benefit of sticking to the new "mental model", you can still think of your content in terms of articles and asides or whatnot. This approach also has the benefit of making it relatively straightforward to upgrade your code to full html5 spec at some point in the future, as all you have to do is search for the classes and replace the divs as appropriate.
This question already has answers here:
What is the difference between id and class in CSS, and when should I use them? [duplicate]
(15 answers)
Closed 7 years ago.
So I often use a website LiveWeave.com to test HTML, CSS, and JavaScript code that I've written. It has a syntax checker, and whenever I use an ID as a selector in the CSS section, it says that it is improper to use an ID as a selector.
I have demonstrated it in this Weaver. To the right of line three in the CSS window is a yellow icon, which, when hovered over, says that it is improper to use IDs as a selector. I was under the impression that that is specifically for the purpose of being used as a selector for a single DOM element, as opposed to classes, which are designed to be applied to multiple DOM elements.
Am I wrong? IS it improper to use an ID as a selector?
The only other instance I can think of an ID being used is for JavaScript document.getElementById(), and similar functions. What is the proper use of an ID?
Note that I am NOT asking the difference between an ID and a Class, but rather whether it is proper to use an ID as a selector.
Using an ID is the most efficient way of selecting a DOM node in both CSS and Javascript. I personally like to use classes for all repeated items and ids for unique items, or unique configurations of repeated modules. There are many CSS patterns, I use a style called BEM (Block, Element, Modifier as seen here) which is a class based naming convention. Look at your favorite websites, right click or inspect. You will find that there is no one right answer to your question, only many right answers.
May I also say that both exist in the standard for a reason and serve a purpose depending on your applications needs.
Below is the order of efficiency for selectors. IDs are the most efficient and pseudo classes and pseudo elements are the least efficient.
id (#myid)
class (.myclass)
tag (div, h1, p)
adjacent sibling (h1 + p)
child (ul > li)
descendent (li a)
universal (*)
attribute (a[rel=”external”])
pseudo-class and pseudo element (a:hover, li:first)
See here...
It is not improper to use ID's as selectors, so long as the ID being used corresponds to only one element in the DOM (document object model).
If you'd like a selector that is multi-purpose, and able to be applied to multiple elements in the DOM, use a class. Although I'm sure you knew that.
The main reason ID's are frowned upon by some CSS developers, and full stack designers, is simply because they aren't as versatile and they have a higher specificity than classes, which can either help or hinder development (based on CSS knowledge).
For more information on CSS specificity, read here: https://css-tricks.com/specifics-on-css-specificity/
It's valid, it's just considered bad practice by some developers because it can make it difficult to maintain your CSS if you're not disciplined about it. I'm no expert on CSS but I'm pretty sure it's all to do with #'s having a really high specificity rating and if you have them dotted around your CSS files it makes it difficult to manage the cascade i.e. inheritance of style rules. So it's considered best by some to use IDs only for referencing elements in your JavaScript.
I've actually heard this argument before.
Some people push the idea of using solely classes for pure css stuff and keeping id for javascript and other id specific functionality.
It would seem that website follows that ideology, so they are trying to get their users to adopt it. I'm not sure if it is yet best practice to keep id out of css
You can decide for yourself whether an id is worth using, when you could just use a class instead.
If you used an ID as a selector and your using it in your Javascript too then you could make situation where if you decide to rename it then you've created a dependency that wouldn't be there if you had used a class name in your CSS.
Also, though using the ID is faster, it isn't faster if you then use #text a - since CSS reads right to left and has to check all the anchor elements first and then find the one with the ID of #text.
This also means the style isn't reusable and you can't use multiple classes either.
So I think the answer really is, based on all the pros and cons of using an ID as the selector, the best practice to keep you out of possible future problems is to not do it. Of course, this all really depends on how you code, the scope of the project and how many other people are working in the project. It's not against the rules, just not really best practice due to possible issues you might be building in that could bite you later.
On top of what has already been mentioned, even in CSS, ID's can be useful depending on what is the structural design.
For example; if every page in your website requires a header and a footer, I don't see why it would not be useful to make it an id.
What is wrong with doing:
#header {}
#footer {}
If you know for sure that your page has only one header and one footer, I don't see the point in using a class.
Mentioning the id is very specific and the page structure is undubious in this case.
Moreover, I also don't see what is wrong by doing something for example like:
.menu{}
#header .menu li{}
#footer .menu li{}
To add specific styling depending on the page segment. Seems very legit to me.
Ultimately, I even think that using ID's to indicate page sections might be more beneficial by ´knowing´ that they are unique (although they might be recurrent across different pages).
Reading an id in a CSS file should give the CSS designer the benefit of immediately knowing what page segment the following css rules are referring to.
A sheet with only classes would in that case seem less clear than using ID's imo.
Closed. This question is opinion-based. It is not currently accepting answers.
Want to improve this question? Update the question so it can be answered with facts and citations by editing this post.
Closed 9 years ago.
Improve this question
I read an interesting post on using a css classname instead of the id attribute for identifying modules or widgets. The css classname could be prefixed with a hash or underscore, to indicate that the classname is used as an id. The reason for this being, that ids can only be used once per webpage whereas a module or widget could appear multiple times.
Simple Example
Instead of using ids such as
<div id="currencyConverter1" class="foo">EUR/USD</div>
<div id="currencyConverter2" class="foo">GB/USD</div>
prefixed classnames are used
<div class="#currencyConverter foo">EUR/USD</div>
<div class="#currencyConverter foo">GB/USD</div>
In the article it suggests, that an underscore could be used instead of a hash, because hashes need to be escaped.
I quite like this idea, but I'm not sure if this is good practice or has any drawbacks. What is the general opinion on this?
It doesn't get simpler than this: does it uniquely identify an element within a document tree now and forever? Use an ID. If not, use a class name.
Remember that IDs and classes are properties pertaining to each element, not a group of elements collectively. Using "identifier" to mean "identifying a group of elements" is futile; that's why you have class names instead, to classify those elements as being related in some manner.
If you're "identifying" elements that are members of a specific widget or module, you have free reign to add a class name to each element representing that widget or module, and select it in addition to the element's class:
<div class="my-module foo">.my-module.foo</div>
<div class="my-module bar">.my-module.bar</div>
Or, if specificity is such a huge issue that you can't even afford to stack another class selector to your rules, prefix the element's class name with the name of that module and select it.
<div class="my-module-foo">.my-module-foo</div>
<div class="my-module-bar">.my-module-bar</div>
If that's the question, then yes, it's perfectly legitimate — as I've mentioned, that's the whole point of class names and selectors.
But there are no legitimate benefits to making a class selector "look like" an ID selector without functioning like one. What it does cause, on the other hand, is needless confusion, especially for other authors who may not know better. If you want to avoid ID selectors like the plague, fine, leave them alone; nobody's forcing you to use them. But if you want to uniquely identify singular elements, then you may want to remember that CSS already provides a feature for that to complement the id attribute in HTML, known as the ID selector. There's no need to hack other parts of the selector syntax to emulate some other feature that's already readily available to you and has been since the very beginning.
Incidentally, if you're running into specificity issues with your CSS rules, then it's your CSS rules that need refactoring, not your markup. Hacking your markup to accommodate your style rules only leads to further trouble, at least in my experience. (I understand the irony in saying this given that hashes are allowed in identifiers starting with HTML5.)
Its mostly driven by your own, personal taste. there are a lot of opinions and articles on this topic, even complete books were written.
I suggest the following:
SMACSS
OOCSS
MVCSS
All of them are mentioning a more-or-less similiar naming convention for CSS, while also saying that these are just rules of thumb, not dogmas.
personally, i follow that approach:
.modName // module
.modName__sub // an object or sub-module
.modName__sub--modifier // a modifier
A similiar structure is used for example by InuitCSS
If you want to use classnames as unique identifiers, than just do it, there is nothing wrong about that. Further, it is future proof in case that you wish to use it as a "standard" class. However, i would shirk that hash for obvious reasons.
Not sure what your modules or widgets are for (wordpress?) but the methodology I choose to use when coding is this:
1: If it is DOM element that has a specific function that I know will only appear once on the page, then I use an ID (things like #main_navigation, #global_header).
2: The DOM element is used for styling purposes (CSS) then I use class names. I keep class names as descriptive of what the DOM element is doing as I can. I don't use vauge names like .blue_text (explained below).
3: I need to attach some sort of information to the DOM element that is kind of awkward and doesn't fit this scheme, then I use custom HTML data attributes. For example, if I create a backend for a site and a user can pick a background color for a div, instead of putting .user_selected_background_color_class as a class in the div I will instead write data="user_selected_color". This is a kind of lame example, but I just built something where a user can select a bunch of images to be in gallery mode or slideshow mode and I used data attributes to determine how the container div should be styled.
This question already has answers here:
Closed 10 years ago.
Possible Duplicate:
Stop CSS styles to be applied in particular sections of the code
HTML code inserted into WordPress with many plugins (each of them probably has it's own style) inherit global styles automatically.
An approach is to find unwanted styles and override them. Sometimes it's difficult to find it as in my case)
How can I define an HTML object to be clean from base and don't inherit any style from the page ?
According to the browser, it's the built in style: the border-spacing is set to 2px.
I think you meant to clear this with cellspacing=0, but the page is an HTML5 page, which doesn't support this attribute.
That said, there's some advice for you. Using tables for mark-up is outdated. You may get away with it, but if you are creating an html 5 page, you should obey the rules that come with it. :) It turns out that it isn't any of the WordPress CSS that is bothering you. It was just the built-in style, which could not be overruled using outdated attributes. :)
For the future, in most browsers you can inspect the elements and their properties. In Chrome this is built in, in FireFox I think you still need to install the FireBug plugin. With that, you can right-click -> inspect element and then check the element or its parent to see what is going on:
This question already has answers here:
Closed 12 years ago.
Possible Duplicate:
Why do /**/ comments work in stylesheets but // comments don't?
In CSS there is only one way to create comments: /* I'm a comment */
The problem is that it isn't nestable.
Why don't we have an alternative comment syntax like //?
I think the real answer is that CSS treats newlines like any other whitespace, so it wouldn't make sense to have comments that are terminated by a newline. This is from the CSS1 spec: http://www.w3.org/TR/REC-CSS1
A CSS style sheet, for any version of CSS, consists of a list of statements. There are two kinds of statements: at-rules and rulesets. There may be whitespace (spaces, tabs, newlines) around the statements.
Of course, this also makes a lot of sense in the context of minification, as mentioned here: Why do /**/ comments work in stylesheets but // comments don't?.
It's not in the specification, and because CSS is widely used and supported, adding it in is virtually impossible. You can't just publish a new specification and expect all browsers to magically support it.
Internet Explorer 6, a browser more than 10 years old, is still widely used, so you can safely assume that even if this addition to the specification were made, it'd take another 10 years to be supported enough to bother. The problem with //-style comments is that they don't scale - unlike new HTML tags, which can be safely ignored as long as the rest of the document makes sense, adding a //-comment will break unaware user agents.
So the short answer is, we don't have it because we don't.
If it really means that much to you, write a script or macro that converts //-comments into /* */-comments, and apply it before running your web application.
There is a way to have // comments in CSS. If you use Sass/Compass.
I really like using Compass, because it gives me everything I miss about CSS, like functions, variables and so on...
Here is the Compass home page and the underlying Sass-language.
Compass is very nice, because you just have a program running in the background that compiles your Sass-code into real CSS, so your workflow is exactly as normal, but in other files (SCSS or Sass) and with very extended functionality!