We cut the “Unwebbable” article (q.v.) in half for space and, in general, to make it better. (It ended up being a “slog”  anyway, didn’t it, Jared?) The other half set out to discuss graphics or illustration formats that cannot be shoehorned into the Web. And here it is!

Some complex document types cannot now be rendered in HTML (any flavour) and never will be thus renderable.

Here I draw your attention to the age-old accessibility question posed by images. If a picture is worth a thousand words, do you really expect to find that many words stuffed into an alt text? Not really – and besides, longdesc was invented just for the purpose of extended description of an image.

But longdesc was scarcely ever used in the real world. Few authors knew about it, few authors cared, few authoring tools made it easy, and, importantly, almost nobody knew how to write a long description. (The W3C even licensed a manual from the National Braille Association [U.S.] to help people out, but that didn’t work, either.)

With the benefit of years of hindsight, I believe that few images that require extensive description are published on the Web. Almost all the complex images that are published online are just posted without change or annotation, as though the Web were a printout. Annotation is just too difficult.

But imposing structure is even worse. Recall a later myth of the Web – that the SVG format is so clearly superior to other image formats, largely because it give structure to graphics, that it would someday replace those other formats (chiefly Flash). That hasn’t happened either. But SVG had a lesson to teach that never really sunk in: Drawings can have structure. Illustrations are not just blobs; they are made up of components.

Even simple tools like MS Excel break up drawings into little components. (Components are structure.) Advanced software, including drafting and CAD applications, offer reusable symbol and image libraries.

Yet when we try to post these illustrations on the Web, the best we can manage is alt="Drawing of aircraft engine". That may make the page functionally accessible, but it doesn’t make the illustration accessible.

This isn’t going to change.

As standardistas, we bet everything on the gambit of adding more words to a page – describing an image rather than making its components semantically distinct and inspectable. The unintended consequence: We give too little detail in an alternate text but provide no way to non-visually inspect an illustration at macro or micro levels. You can’t get a gestalt of the aircraft engine and you can’t “drill down” to a single turbine blade or bolt.

If the solution I seem to be advocating is “more semantics,” that isn’t going to work, either. Graphics and illustrations (even business charts) populate the earth in too many species to be classified. We’re never going to be able to develop a workable vocabulary, a set of semantic tags, for illustrations. The task borders on pointless.

Some things you just have to look at.

The Web, then, gives us a way to publish and distribute works such as these. But with no means of marking up their structure, these works will never be Web-native. To revisit Case 1, they’ll be like PDFs – something you can publish online, something you might look at in a browser, that nonetheless is not and never will be “Web content.”

Other illustration types that resist structural markup

  • Org charts and flowcharts. Nested ordered lists are a proven failure here.
  • Circuit diagrams.
  • Many graph styles, including radial graphs.
  • Exploded and X-ray diagrams.
  • Blueprints.
  • Timelines.

Fans of Edward Tufte should examine his many case studies. Few, if any, could be marked up with any structure in use today or on the horizon.


The foregoing posting appeared on Joe Clark’s personal Weblog on 2009.07.22 11:27. This presentation was designed for printing and omits components that make sense only onscreen. (If you are seeing this on a screen, then the page stylesheet was not loaded or not loaded properly.) The permanent link is:

(Values you enter are stored and may be published)



None. I quit.

Copyright © 2004–2024