CSS Areas Thought-about Dangerous – A Record Aside

Article Continues Beneath

Håkon Wium Lie is the daddy of CSS, the CTO of Opera, and a pioneer advocate for internet requirements. His final article on this journal led on to actual fonts on the net. When Håkon speaks, whether or not we at all times agree or not, we pay attention. In at the moment’s put up, Håkon shares his opinion on CSS Areas.

Approach again, when designers first began emigrating from their islands of desktop publishing onto the online, they requested a seemingly easy query: how may they take with them their favourite colours, fonts and layouts? At first, we had no good reply. HTML provided semantic tags to explain the construction of paperwork, not their presentation. Browsers couldn’t be informed what fonts to make use of, or the place to fetch them. Authors, nevertheless, quickly discovered a sneaky shortcut: the img factor. By making pictures of their textual content, authors may obtain their colourful helveticized designs. Within the course of, all semantics (that’s, details about the which means of components, versus the presentation) was eliminated and even the searchable textual content was gone. CSS was proposed in 1994 to cease this follow; saving semantic HTML was simply as necessary as reaching fascinating layouts.

A number of years later, on the peak of the XML fever, presentational components made a comeback. XSL outlined an XML vocabulary for Formatting Objects; XSL-FO tags that mentioned nothing about being headlines or checklist gadgets, however all about their presentation. Pc scientists have a peculiar manner of expressing concern and doubt. They publish essays with “thought-about dangerous” within the title. This explicit design sample was began by Edsger Dijkstra when he revealed “Go To Statements Thought-about Dangerous” in 1968. The event of formatting objects led me to make use of the identical gadget; “Formatting Objects Thought-about Dangerous” argued that formatting objects have been font tags in disguise and that their use on the net should be averted to protect internet semantics.

Evidently proposals for presentational components return from time to time. The latest incarnation is CSS Areas. One shouldn’t write “thought-about dangerous” articles evenly, however presentational components will not be the one downside with CSS Areas. For many who imagine in significant HTML tags, responsive internet design, and compact CSS code, the introduction of CSS Areas will not be excellent news.

Downside #1: areas use dummy divs

Some articles on CSS Areas have already seemed on the supply code. An article revealed by WebPlatform.org describes methods to obtain a generally used two-column design:

a sample document with columns of text and a view of the CSS region layout

The formatted doc is on the left, and the corresponding areas are proven on the appropriate. The HTML code that generates this structure should be studied in an effort to perceive CSS Areas. Right here’s a snippet:

<part class="web page">
 	<div id="title">	<h1>Area #1</h1>			</div>
 	<div id="intro">	<h1>Area #2</h1>			</div>
 	<div id="col1">	<h1>Area #3</h1>			</div>
 	<div id="col2a">	<h1>Area #4</h1>			</div>
 	<div id="pull">	<h1>(Separate Area)</h1>	</div>
	<div id="col2b">	<h1>Area #5</h1>			</div>

The weather above signify areas, that are containers the place textual content can stream from one to the opposite. Right here is among the corresponding CSS declarations for the #intro factor:

#intro {
	width: 45%;
	place: absolute;
	high: 5em;
	peak: 3em;
	-webkit-flow-from: primary;
	-ms-flow-from: primary;
	flow-from: primary;

The CSS code above says, roughly: flip the #intro factor into a completely positioned factor with a given measurement and place, then discard the content material of the factor and exchange it with content material from the stream known as “primary”. Thus, the h1 factor inside #intro isn’t a headline in any respect—the div factor is a presentational container and the h1 factor is discarded.

The proponents of CSS Areas would possibly argue that, “Sure, the divs are there for presentational functions, however solely components may be scripted on the net and we should subsequently use components.” This underlines an necessary level: it’s not areas per se that which are dangerous to internet semantics, it’s the truth that they’re encoded as presentational HTML components. If we wish areas on the net, we should always discover a strategy to write them in CSS and never in HTML. If CSS Areas are accepted in 2014, we can be caught with completely positioned dummy divs for the foreseeable future.

Not all internet designers are involved about semantics. If areas can present the design instruments they crave, a couple of dummy divs are value the fee to some. Subsequently, let’s analyze CSS Areas from an internet design perspective. Do CSS Areas make good web sites?

Downside #2: areas are usually not responsive

Responsive design is a trademark of fine internet design. We would like our websites to be scalable throughout a variety of units; from small cell phones, to smarter telephones, to large screens. Columns are a neat manner of utilizing the total width whereas protecting line lengths down. The instance from the WebPlatform article flows its content material into two columns. The width of the columns are set in percentages (45 p.c, as may be seen within the instance above). That gives for scaling as the 2 columns will develop with the width of the display. However, there’ll at all times be two columns. Even on a really slender display, the instance won’t change to a one-column structure.

Ideally, you need the variety of columns to be dynamic so {that a} slender display has one column, a medium display has two columns, and an ultra-wide display has three, or perhaps 4 columns. CSS Areas won’t provide you with this. One other CSS specification, nevertheless, describes methods to obtain pages with scalable variety of pages:

four views of a web page demonstrating different numbers of columns for different window sizes

The samples above are produced with CSS Multi-column Structure, a specification for which I’m the editor. And right here is the code:

article {
	columns: 20em

The snippet above tells the formatter that the optimum line size is 20em, and that the variety of columns must be computed accordingly. Thus, on a small display you should have one column, on a wider display two columns, and many others. In impact, multi-column structure robotically creates the optimum variety of linked areas wanted to show the content material. No presentational components are wanted to signify these areas—they’re robotically created.

Can this sort of automated structure of areas may also be a limitation although? Authors must be allowed to put areas far aside, shouldn’t they? This query brings us to the subsequent part.

Downside #3: complicated textual content stream

Specs usually begin out with a motivational instance to indicate how highly effective the proposed performance is and the way simple it may be achieved. The primary instance of the CSS Areas specification is proven under:

diagram of a page with a confusing visual text flow

The textual content stream strikes from area 1, to 2, 3, and 4, following the arrows. Discover how the eyes of the reader must traverse sideways, within the reverse of the studying course, from 3 to 4. These sorts of traversals are usually not widespread in newspaper design, and I’ll argue that they’re complicated to readers and must be averted. CSS Multi-column Structure can’t make textual content stream this manner, and I contemplate that to be a function.

In my (arguably fairly memory-constrained) thoughts, it will be much better to place the determine (A) on the backside of the web page and stack the columns within the studying course left to proper. Like this:

diagram of page with an improved visual text flow

On this design, the utmost size that the attention must traverse backwards is the size of the road. This design can simply be achieved in CSS Multi-column Structure. Thus, in my expertise, CSS multi-column layouts can do a lot of the issues that designers plan to make use of areas for. In case you can’t obtain your most popular columns utilizing CSS Multi-column Structure, you must in all probability rethink it.

Amongst the issues mentioned on this article, that is in all probability the one I’m least anxious about; if sufficient customers are confused, the design will change. But it surely appears wasteful to take a position years of efforts to implement CSS Areas if a lot of the compelling use instances may be achieved via an present mechanism.

Downside #4: verbosity

The motivational instance from the Areas draft additionally has some pattern CSS code connected. The CSS code is proven in an appendix, in all probability as a result of it’s too lengthy to show at first of the doc. The CSS code makes use of 26 strains (after eradicating feedback and clean strains). As well as, there are seven strains of HTML code for the dummy components, bringing the full to 33 strains.

What number of strains wouldn’t it take to encode the extra intuitive design (proven within the determine above) utilizing CSS Multi-column Structure? Three, it seems:

article { columns: 20em }
h1 { column-span: 2 }
img { column-span: 2; float: backside }

For this instance, utilizing CSS Areas is a magnitude extra advanced than utilizing CSS Multi-column Structure. If one have been to assist one structure on small screens, and extra columns on wider screens, the code measurement for utilizing CSS Areas would develop rapidly.

(I need to add that my favored code examples on this piece use options not simply from CSS Multi-column Structure, but additionally from CSS Figures, which add integer values to the column-span property, and high and backside to the float property.)

Downside #5: code reuse

Returning to the article on CSS Areas, contemplate the highest a part of the formatted article. The #intro factor mentioned above is a area set to comprise the introduction, proven in white on a darkish background. Discover how the textual content flows from the darkish field to the subsequent area underneath it.

top of the CSS Region sample document

Do your eyes observe? Are you aware the place the introduction ends? Admittedly, the primary sentence of Finnegan’s Wake isn’t a conventional introduction, however contemplate how this can be used for extra conventional articles. Personally, I really feel cheated when big-font introductions flip into small-font physique textual content in the identical sentence. As a reader, I’d wish to know the place the introduction begins and the place it ends and imagine it ought to finish in an entire sentence. Most publications use my favored model, whereas some model a field of a sure measurement like within the above instance. These publications can be tempted to attempt CSS Areas.

Nevertheless, I don’t assume they are going to be proud of utilizing CSS Areas this manner. It’s problematic to assign model to a field as a result of authors have no idea what number of strains of textual content will match into the field. A standard downside with completely positioned components is that the textual content doesn’t at all times fill the field because the writer intends. Typically there can be unused white house on the backside of the factor. The identical will occur when authors begin utilizing CSS Areas to model the primary (say) 5 strains of textual content; maybe solely 4 strains will match and there can be a visible bounce to the remainder of the textual content.

An identical downside seems within the pullquote that seems in the appropriate column:

portion of CSS Region sample document with pullquote

The pullquote begins with “The oaks…” within the picture above. The area it lives in has a measurement simply large enough to carry the textual content. An authoring software might compute the dimensions of the field and create a custom-made stylesheet. Nevertheless, to ensure that the doc to be introduced “appropriately,” the textual content should stay unchanged (no translations, please!), the identical fonts should be accessible (please don’t flip off downloadable fonts!), and the person should not improve the minimal font measurement (please convey your magnifier!).

Stylesheets written this manner won’t be reusable, every new doc can have its personal.


CSS Areas have been proposed by Adobe in 2011 and the corporate remains to be its primary proponent. It’s laudable that Adobe takes the online severely and that it brings proposals to W3C—definitely a lot better than pushing a proprietary expertise like Flash. Its motivation is to promote authoring instruments that generate CSS code. That’s good, too—the online wants good authoring instruments and Adobe could make them. However CSS Areas, as at the moment proposed, won’t enhance the online. Fairly, it brings presentational tags, verbose code, and per-document stylesheets.

Leave a Comment