Slashdot Mirror

We sent 10 of your questions to usability guy Joe Clark, and he took it upon himself to go a bit beyond simply answering them. In his reply he said, "Answers attached in a valid XHTML file. I would suggest at least retaining the id attributes. I copy-edited all the questions, but the words are all the same; they are now merely spelled and capitalized correctly. I think all the links work." Whatever. We left Joe's formatting intact. It's a little different from our usual style, but variety is the spice of Slashdot. Ask the Expert: Accessibility 1) How far should it go?

by newsdee

Macromedia Flash has integrated many accessibility features in an effort to promote development of content for special needs. However, can we realistically try to turn any multimedia feature into its accessible equivalent? Is it even feasible other than providing a text-only equivalent?

There seems to be a stereotyped understanding of Flash content at work here. Flashturbation is not the only usage of that authoring tool.

I believe the question really intends to ask Are artistic uses of Flash, like Josh Daviss Praystation, really amenable to accessibility? The answer is a qualified yes, and I say that because Praystation-like Flash experimentation is essentially a form of cinema that merely uses the Web as a delivery mechanism. Cinematic experiments of this sort are indisputably a different species from other forms of Flash development.

In that example, the solution is to treat the Flash objects as a movie and apply standard movie accessibility features, namely captioning and audio description. Im not one of those people who believes that abstract, experimental, or non-narrative cinema cannot be captioned and described lots of music videos fall into that category, and theyve been captioned for nearly 15 years. (Description of experimental audiovisual artworks has not really been attempted to my knowledge, but description of abstract art in museums and of non-narrative plays and dance performances in theatres have all been going on for years. Its perfectly possible.)

The challenges, then, are two: Infrastructure and interface. There isnt really a very good way of including captions or descriptions in a Flash file as yet (an infrastructure problem). Macromedia knows all about this (Ive discussed it with them at length, and also written about it), and it will eventually be fixed. (Even finding an example of Flash with captioning is difficult today. Youd think Id have a complete list at the tip of my fingers, but I dont. The Macromedia Contribute feature tour is one case.) I dont know of any Flash animation that was ever described.

The interface problem is: How does the viewer turn captions and descriptions on and off? This isnt like a TV set, where you can manipulate onscreen menus (and how do you manage that if youre blind?) to turn captions and/or descriptions on and off. Browsers are not smart enough to automatically turn access features on and off, though I think a future upgrade of one file format that shall remain nameless will be the first to include such a capacity. At any rate, this may be one of the rare cases where an overt visual change must be made to accommodate accessibility actual selectable buttons to turn CC and DX on and off. (The buttons themselves have to be accessible, i.e., part of the tabbing order and with alternate texts and so forth.)

Now, lets consider other examples of Flash.

Banner ads the really big skyscraper ads that bug your arse on so many sites The usual Flash accessibility features can be used, and you can be smart and include the Flash object inside, say, an iframe element, which provides vast options for accessibility. (You can add a long description to the iframe, though thats questionably useful, and include alternate content in case the main content cannot be loaded, which could be an ordinary animated GIF or still image with alt and title.) Comics Flash-based comics can be relatively straightforward to make accessible (Apocamon doesnt seem too tricky its essentially a panel-based comic strip with a wee bit of animation) or could require full-on cinematic techniques, as with Broken Saints. User interfaces Flash can be and is used as a tidier means of providing a user interface, as at FoxSports.com or in the Neuros audio-player demo. The temptation, as in that last example, is also to use motion graphics and audio, which may require the same CC and DX as before, but many user interface can be made adequately accessible with todays Flash accessibility tools (text equivalents, making objects visible or invisible in the document structure, etc.). Manipulable objects Games (including the Royal National Institute for the Blinds ill-advised consciousness-raising game, no longer online) and even some interfaces (like History of Health Care) may include objects youre intended to grab and manipulate with the mouse. The current Flash accessibility tools are not really up to the challenge of adding keyboard equivalents for such manipulable objects. You could hack it together yourself, but there are no built-in commands or primitives you could use in a standards-compliant way. Intros Skippable intros are just as awful today as the day they were invented. Unfortunately, we cant make value judgements about which information should and should not be made accessible. Even skippable intros have to be made accessible, either by treating them as cinema or simply giving them a few text equivalents. The skip-intro link has to be selectable by keyboard, of course. Tools These interfaces let you do something. One I like a lot, if only because I am a typography queen, is Jeremy Tankards font viewer, though it is admittedly overkill because other font-viewing miniprograms do not require Flash. It may be possible to make the inputs to such tools accessible (you can place the cursor in the right place, operate controls, and so forth), but the results might be intrinsically inaccessible. (Note that artists portfolio sites, font and clip-art vendors, stock-photo houses, and other sites that sell visual imagery using ordinary HTML can be made passably accessible even to a blind person. In the Tankard case, perhaps only the name of the font and the text entered would be rendered to a screen reader or other device.) E-commerce Perhaps the most credible Flash instance, E-commerce sites like Ted Baker (see its Footwear store) may include all the features of the other instances Ive listed here. Since E-commerce is a convenient way to shop for many disabled people, I would strongly emphasize the need for accessibility. But it might be stretching the limits of current Flash access tools, since you have to make an interface, product shots and other images, and text all accessible. Thats not difficult in HTML, but I dont have any examples to point to of accessible Flash-based E-commerce sites that we could use as a comparison; I dont know how hard it would be to make such sites accessible. Aside: The most sophisticated Flash site Ive ever seen is DirtyBastards.com. (No direct hyperlink; consider this the strongest possible warning of adult content. Be very sure you want to look at it.) The usability could use an update, but in general its astounding. Should we ever be in the same city, Ill take anyone who can update that site for accessibility to dinner at the restaurant of their choice.

I would add a proviso here. Accessibility does not relate solely to blind people. As mentioned above, any quasi-cinematic work with audio requires captioning; deaf people need accessibility, too. There is much more attention being paid now to the Web-accessibility needs of people with learning disabilities (the most famous of which is dyslexia), which well get to later.

Learning-disabled people are by far the hardest to accommodate online, and for many HTML pages, they are probably impossible to accommodate in any really helpful way. Flash animations could be a good solution for that group because you can build in many levels of information, use audio and graphics, and provide really good controls for pacing (because having too much information coming at you all at once is a barrier for many people). Inevitably, accessible Flash in that context would limit itself to custom-engineered animations specifically made for that audience; I doubt that general uses of Flash will be upgraded for that kind of accessibility.

Text-only sites are not the alternative to accessible sites. Text-only is not accessible. Well discuss graphic sophistication later.

Biggest problem

by robbo

What, in your opinion, is the most common complaint concerning accessibility and Web sites? In other words, if in the interests of accessibility you could encourage site owners to change only one thing about how they operate, what would it be?

Images. Seriously, if youve got an ordinary HTML Web page and you make absolutely all your images accessible including, crucially, adding alt="" to every spacer GIF and every other meaningless graphic youre four-fifths of the way to being an accessible Web site for the group with the greatest single need, the blind and visually-impaired.

I emphasize coding to standards. Unless you have an airtight reason (like youre stuck using an old content-management system you cannot afford to replace), I really dont want to have anything to do with you unless youreproducing valid HTML. Now, tiny invalidities are just that, tiny: <hr> and <hr/> really are the same thing. And Im sure that ultra-purist geeks will now launch a hypocrisy hunt and comb through my entire Web presence to locate pages with non-valid markup. (Knock yourselves out. I make small mistakes, and have not updated scores of very old pages. Im also a vegan with some shoes and accessories made of leather. Complete purity is sometimes unattainable.) In one of the many ironies of Web development, it is indie developers like me who have a higher success rate in achieving valid, accessible sites even though larger commercial operations are the ones where valid HTML and accessibility are more urgently needed.

In any event, if youre producing tag soup, as far as Im concerned youre demonstrably not all that interested in responsible Web development.

The upside? If you do write valid pages, you have to include at least an alt text for every graphic. For no extra effort (you have to do it anyway), you get basic accessibility.

Number two on the list is navigation. Left-hand and top navbars stacked with link after link are a nightmare to wade through if you have a mobility impairment that reduces your ability to use a mouse or keyboard. (Screen-reader users are not so heavily affected; they can skip entire table cells, for example. I suppose all-CSS layouts are harder to skip through. But thats not the page authors problem; its incumbent on the adaptive technology and browser to clean up their act.)

If youre able to use a mouse, you can just avoid the entire navbar. But a mobility-impaired person may be stuck tabbing from one link to another and thats the best-case scenario. Quite possibly, a mobility-impaired visitor may be using software that cycles through a set of input choices for example, the mouse; then the alphabet keys of keyboard; then the number keys; then the function keys. You may have to wait until the keyboard option cycles back again in order to type repeated keystrokes. (You may have a mental image of a sip-and-puff switch or Christopher Reeve using speech-input software. The principles are the same and so is the inconvenience.)

If you, the page designer, stack 20 or even a hundred links in a left-hand navbar and assume that people can simply tab through them, well, (a) tabbing 20 or a hundred times is something youd never expect a nondisabled person to put up with, and (b) some people will have to wait 20 or a hundred cycles of their software in order to do the equivalent of pressing the Tab key.

The solution? Put skip-navigation links on top of every navbar with, say, ten or more links. (Or fewer. Use your judgement. Section 508 regulations technically require a skip link in every navbar, even for a page footer.)

Note that skip-navigation links have to be visible; a lot of people use hyperlinked single-pixel GIFs with alt texts, but those are invisible to mobility-impaired people, most of whom have normal vision. The links dont have to be ugly or intrusive, but they have to be plainly visible and selectable. (If you want to be thorough, you can give them accesskey and tabindex values.)

Do those two things and your site becomes vastly more accessible to two large disability groups right then and there.

Accessible Slashdot?

by ictatha

How does Slashdot stack up? What about blog-type sites in general? What can be done on these types of sites to make them more accessible?

Mark Pilgrim has fully strip-mined this topic. (He also tech-edited my book and is generally formidable.)

The issue here is random vs. serial access. A nondisabled site visitor can jump around the page. If you can see, its very easy to skim the page, and it is also very easy to zip to what interests you if you can operate a mouse or keyboard well. Nondisabled people have random access to the contents of a page. Many disabled people the blind and the mobility-impaired in specific experience a Web site serially, with one item after another articulated (as in speech or Braille) or selected. The page author can make skipping around easier, and so can relevant software like screen readers, but its still going to be harder to navigate than for a nondisabled person.

Slashdot is dominated by words. The page introducing this interview carried about 6,900 words even with minimal comment expansion. The issue, then, becomes navigation, which I discussed in the previous answer. Adding hyperlinks to skip various navbars would be a good first step.

Slashdot could certainly use better semantic markup. Valid code is a must; I want Slashdot to eat my own dog food. Subject lines of postings could and should be marked up as headings (h1 through h6); font elements could be eliminated; Im not wild about table markup to achieve indention, though making structural hierarchies apparent is not easy at all (perhaps unordered lists with a style declaration of list-style-type: none might suffice). It would then be possible to navigate from heading to heading.

If youre running a more limited Weblog with just a couple of screenfuls of text at a time, then my advice is simple: Write valid code, provide a text equivalent for every image, work on navigation a bit, and youve made a big dent in the problem.

Photoblogs or those containing multimedia are, of course, more complicated, but as long as every photo has an alt text and your multimedia is captioned and described, youre doing well. It is certainly easy to add alt texts to your photos, but captioning and description are hard to do well and are technically difficult to implement. Im mentioning the multimedia case merely for completeness; I dont read any blogs that regularly post video and audio. (I suppose The Ben Brown Show was an example.)

Accessibility

by acehole

Do you think that where companies are being sued or forced into updating their Web pages at great expense to include accessibility for the blind in their Web pages when the blind could easily find another similar service offline is reasonable?

You have inadvertently stumbled across an extensive issue in disability law the question of providing equivalent or comparable access, or access that is equal in dignity to that afforded a nondisabled person.

You can draw parallels with the physical world. Think of barrier-free entrances to buildings. If the main entrance is at the centre of the buildings face but uses a staircase you cant remove, then providing a barrier-free entrance at the left side of that building would probably be considered comparable or equivalent access. But if you force a mobility-impaired person to walk through an alleyway and take a rear service elevator that is otherwise used for garbage, your accessibility probably is not comparable or equivalent. (Thats in the case of a relatively new building. A historic building or another exceptional case might permit different treatment of that sort.)

If we consider information media, theres a distinction to be drawn between old and new media, or non-electronic and electronic forms. Books are the canonical example: They cannot be made intrinsically accessible to a blind person because a book embodies a single immutable form. You have to provide accessibility elsewhere, as through a large-print edition (its a separate form), a Braille edition (also separate), or a talking book (separate yet again).

Electronic (or audiovisual) media can carry accessibility along with themselves:

• You can add closed captions and closed descriptions to a television program, DVD, online video segment, or first-run movie. (Im skipping some technical details in the movie example.)
• You can add closed captions to a videotape.
• You can add accessibility features to a Web site.

(In the first two cases, you could instead add open captions or descriptions that everyone sees or hears, but thats a very unusual practice, and by doing so you essentially create a separate work, just like publishing a large-print, Braille, or talking-book edition of a printed book.)

In all the examples above, you the viewer can activate the accessibility if you need it or ignore it if you dont. Because Web sites are electronic and can carry hidden access features, the answer to aceholes question is no, it is not reasonable to expect disabled people to go somewhere else to get the same information or enjoy the same experience.

Accordingly, yes, Southwest Airlines reservation Web site should be accessible, and no, it is not OK to expect blind people to call a telephone number when nondisabled people do not have to do so. (Read various other reasons why.)

Thats unequal treatment right there. It is not comparable or equivalent treatment, and, I argue, it impugns the dignity of a visually-impaired person who has already made a commitment to independence by using the Web with adaptive technology.

I also reject, in the strongest possible terms, the offensive and offhand claim that accessibility can be achieved at great expense. I believe the colloquial term for a claim like this is bullshit. Updating or retrofitting a site for accessibility does cost more than designing it properly in the first place, but thats true everywhere: Have you costed out adding barrier-free access to an old building vs. including it in the original designs? Retrofitting may cost more, but I deny that the expense is great. Even very extensive sites with huge swaths of multimedia can be made accessible, and it is doubtful that, given the budgets of such sites, the expense would be great.

In any event, developers always find a justification for what they like to do, whether it be Flashturbation or coding custom JavaScript features or whatever else. Its a bit late in the day, in Web-development terms, to claim that accessibility is not one of the arrows in the quiver of the competent practitioner.

Now, another of the subtexts in this question really, it is a spiders web of half-truths, barely-suppressed resentments, and ignorance suggests that the only way to achieve Web accessibility is by being sued or forced. I have consistently argued that lawsuits are the worst way to achieve accessibility, particularly in the U.S., with its poisonous atmosphere. Lawsuits merely get peoples backs up and sour the defendant on the entire concept. Defendants are forced to belittle and invalidate the concerns of people with disabilities merely in order to provide an adequate defense in the case. This is no way to run a railroad.

But lawsuits (and human-rights complaints and other actions) are still necessary from time to time. Disability law is old and tends not to expressly include the Web. (Sometimes it doesnt even include established accessibility techniques for old media, like audio description on TV.)

Its unrealistic to wait around forever for clueless lawmakers, who can barely use a cellphone let alone surf the Web, to update the legislation. To get some kind of jurisprudence on the books, lawsuits and complaints have to be filed from time to time. Sometimes it works, sometimes it doesnt, but the law is a tool that must be available to everyone, including people with disabilities, whose rights have legal standing.

A competent Web developer builds accessible Web sites and does not wait to be asked to do so, let alone sued or forced.

Market for Web developers

by ragnar

Im considering a starting up a Web development firm with a focus on accessibility. I have good relations with the principals of an accessibility testing firm and believe the businesses can complement each other well. Im a part owner of a Web development firm at the moment that isnt interested in pursuing this market, but I believe there is a significant market.

Can you elaborate on the market for Web development firms that focus on accessibility? Aside from the normal perils of launching a new business (which Im fairly acquainted with), can you expound on the market need for firms that endeavor to deliver accessible content?

Deliver[ing] accessible content and starting up a Web development firm with a focus on accessibility are two different things, so lets focus on the latter.

I would say that the market for accessibility-specific Web consultancies is rather small and will have a short lifespan. I can say this with some confidence as I am an authority on accessibility, with a published book to prove it, and I hardly get any business. Even taking other factors into account, I think its the nature of the work. I have various reliable indications that other consultants arent flush with activity, either.

Why?

• Accessibility is neglected. People cant hire you to do what they never knew needed to be done anyway. Nor will they hire you to do what they resent having to do in the first place and will resist doing until their dying breath.
• Contracts are small. Even very large sites tend to be run by CMSs or templates. Once you clean those up, boom, tens of thousands of pages become accessible. There is often not a lot of billable work involved, as I know myself all too well.
• Attainable expertise. If, as I contend, accessibility is merely one of the skills a competent developer must have, eventually all the competent developers will gain that expertise. They wont need outside experts. Even if in-house access knowledge is demonstrably worse than outside consultants, there are all sorts of precedents for companies making do with barely-passable accessibility because its cheaper. There is a preference for meeting the letter of any requirements (whether self- or externally-imposed) rather than doing accessibility well.

Now, what may work massively better is, in fact, accessibility testing (and certification). It is extremely difficult and time-consuming to test site accessibility with actual disabled persons using actual adaptive technology. A firm that updates Web sites to accessibility standards, advises on how to write new sites that conform, and tests them to prove it may be a winning combination.

The issue of then certifying a site as being accessible (or meeting certain requirements) becomes a bit trickier, but Id really like to see someone give it a go. Note that any venture like this will require thoroughgoing knowledge of the Web Content Accessibility Guidelines, the Section 508 regs, adaptive technology, and multimedia accessibility, and that knowledge definitely includes an understanding of exceptions to the rules. I deal with too many people who literally read and literally apply whatever guideline theyve decided is gospel. Accessibility requires human judgement based on knowledge and experience. Dont set up shop without it.

What of dynamic images (charts and graphs)?

by kuwan

I see that Chapter 6 addresses the image problem which you state is a core concern in accessibility. My question is, what is your solution to data-intensive sites that display their information using graphs? For sites that have constantly changing data (stock charts, for example), what solutions/tools are there to make their graphics accessible?

The answer is that such information, in certain cases, cannot be made meaningfully accessible to a blind or visually-impaired person, or probably to a learning-disabled person. Other disability groups should be unaffected.

This, of course, leads me to my perennial complaint about the Web Accessibility Initiative and accessibility advocates generally: Theyve got no style. They have no understanding of graphic design and typography, and they project this ignorance onto the rest of the world.

To use one of my maxims, accessibility opponents think accessibility means a text-only Web site and hate the idea, while accessibility advocates also think it means a text-only site and love the idea. Theyre both wrong.

One consequence of this ignorance of visual design? The implicit claim that every illustration can be epitomized in words. You could only make this claim if you were so visually unsophisticated that you couldnt differentiate one kind of illustration from another. Of course, this is hogwash: The reason why we use illustrations is because words (or numbers) are sometimes too hard to understand by themselves.

A graph of stock performance, radar weather maps, ultrasound images any picture that is worth much more than a thousand words presents a quandary. The goal here is accessibility a disabled visitor must have equivalent access to the information conveyed by the graphic. If the underlying data is numeric, in theory you could provide the underlying data (as through the longdesc attribute of the img element just set up an HTML file, or, theoretically, a spreadsheet or a PDF, that could be loaded to describe the illustration at length).

But remember, all that numeric data was so hard to understand for nondisabled people that it was turned into a chart; now youre expecting screen-reader users to wade through those numbers one at a time? Like packing, unpacking, and repacking a suitcase, converting data to graphics and back again tends to leave something behind in the transformation.

You may have provided a text equivalent in such a case, but you have not provided accessibility.

I am not giving a carte-blanche exemption here. Many charts and graphs have one or two key points that could, in fact, be added to something as simple as an alt text: alt="Graph shows 12.2% increase in HIV seroconversion in gay males 18 to 24, 1996 to 2001". Even severely complex illustrations require at least a structural placeholder, like alt="Hubble photograph of Jupiter, its rings, and its satellites".

Its true that genuinely equal access to the information embodied in complex illustrations can be unattainable. These are exceptional cases, but they do come up.

Useful links:

• National Braille Associations recommended practices for converting illustrations into accessible forms
• PopChart by Corda attempts to automatically write long descriptions of (numerical) graphs

Deaf-blind

by gmhowell (26755)

Text-to-speech works fine for blind people (mostly). Deaf people can see most Web content. What the heck are deaf-blind people supposed to do?

One of the joys of Delphi, GEnie, Compuserve, etc. is that the discussion boards worked fine with simple telnet access, and Braille TTYs. The various Web boards that have supplanted them dont seem like they would work as well (sorry, havent tried any yet; those Braille TTYs aint cheap).

Yes, this is a personal question (see .sig).

I need help with tech solutions for the deaf-blind. Please contact me via E-mail if you have any experience in this.

Well, deaf-blind people are difficult to accommodate. Theyre also rare: Though adequate population numbers are hard to find, perhaps 11,000 deaf-blind people live in the U.S. But in some contexts, the fact that theyre deaf has no bearing on accessibility. Blindness is the issue.

Screen readers (manufacturer list) not only can turn Web sites and computer software into voice, they can also typically output text to Braille displays. (I wouldnt call them Braille TTYs, since those are used to communicate by telephone.) Braille displays are fascinating, rarefied, and costly devices. Tieman, Freedom Scientific, and ALVA are notable manufacturers. Not all that many people use them, in part because not all that many people read Braille (maybe 10% of blind people), though essentially all deaf-blind people read Braille.

Anyway, for a Web site that does not include multimedia, the fact that youre also deaf has no influence on accessibility if youre already blind. For a deaf-blind person using a screen reader with a Braille display, ordinary Web accessibility becomes the issue, though Id say that navigation help becomes much more important there. Experienced speech-output users run speech at superhuman speeds (300 words a minute is not uncommon), meaning you can burn through a page, albeit in serial fashion, pretty quickly. Given that Braille displays provide one or a couple of lines of Braille at a time, its a more time-consuming procedure.

Now, for sites that do contain multimedia, there is no viable option. An obvious course of action (requested by one activist group) would be to combine caption transcripts and audio-description scripts so that one could essentially read a text rendering of a videoclip, but there is no technology that can actually do that yet. (Yet. I have plans.) Combined script-transcripts of this sort have been attempted manually a couple of times (and I mentioned the idea back in 1999), but I dont know of any research on how well it all worked.

Alternative (non-computer) devices

by superflippy

Increasingly, people are using non-computer devices (cell phones, PDAs) to browse Web sites. What alternative devices are disabled people using, and how are they using them in ways Web developers might not have considered (e.g. voice browser in cell phone)?

Im not really up on that topic. The PAC Mate is one such device; its essentially a screen reader without a screen or free-standing computer.

Accessible site, or accessible browser?

by vofka

I am a partially-sighted person, and I have to admit that I do frequently have difficulty with accessibility issues, particularly with large corporate Web sites which all seem to be full-flow multimedia blitzes which require 1600x1200 resolution or higher, and usually override the default browser fonts to make them smaller.

However, there are a number of browsers, such as Mozilla (just one example, Im sure there are others!) which allow the user to zoom the text on a page, to override colour settings etc.

Though it is undoubtedly important for Webmasters to pay great thought to the design of their sites in terms of colour, font size and multimedia content, how much relative importance should be placed on browser design, and the browsers ability to override the design decisions of the creator of a site?

Its important and overlooked. It would be nice if we had a browser that actually supported all of HTML; we dont (no, not even Mozilla). Then it would be nice if CSS1 and CSS2 were fully supported admittedly an onerous task what with the myriad interactions and the various ambiguities in the spec.

At that point, yes, the user customizability in CSS and the many options available in HTML would presumably be up to the user to control. I think its ridiculous that the only really effective way to override a page authors CSS is for you, the harried, humble Web-surfer, to write your own CSS declarations (dont forget !important!) and activate the file in your browser, if thats even possible. This is the sort of thing that should be built into browser preferences, available for easy use. The first time you start up a browser, it should explicitly ask you if you have any accessibility requirements; a lot of people dont even know about what few customization features browsers currently offer.

Ill make another of my analogies. Remember the lack of visual sophistication of accessibility advocates? They want designers to work at their level by providing accessibility, but they never seem to understand that the converse is also true accessibility activists must learn to work at designers level by providing good site design. By the same token, if page authors are expected to use every practical accessibility feature, then browser makers must be expected to support all of them and support them well.

In the immortal words of Comedy Central, Weve upped our standards. Up yours!

See also: User Agent Accessibility Guidelines.

Physical vs. cognitive political clout

by Aquitaine

Dear Mr. Clark,

I am a Web developer for the Program on Employment and Disability at the School of Industrial Labor Relations at Cornell University. Web accessibility is a serious issue for us, and we try to keep abreast of innovative approaches to design so we can find that elusive place where universal accessibility meets intelligent and aesthetically pleasing layout. We recently spoke with Cynthia Waddell (one of 8 authors of Constructing Accessible Web Sites, also out fairly recently) on this subject, but I found her unwilling to commit to anything other than suggestions rather than real technical solutions.

There are two sticky issues that I have encountered. The first is the notion of universal access. Mrs. Waddell indicated that, working with the W3C, she was coming up with a list of Web sites that met Priorities 13 of the W3C WAI and were still aesthetically impressive (she did not have a list ready). As you are no doubt aware, many sites that tout universal access are themselves victims of poor design -- the problem of Yes, its W3C/WAI compliant across the board, but its ugly as sin. Do you believe that a site can have a single interface that is truly universally accessible, or do you believe that sites should have alternate interfaces? (The Web equivalent of Do we have a ramp and stairs or just a ramp?)

Along those lines, it is apparent to me that the accessibility guidelines are designed to be useful in a manner proportional to the lobbying power of disability rights groups. That is to say, blind people and deaf people, although they comprise extraordinarily small percentages of people with disabilities, have an enormous amount of political clout when compared to people with cognitive disorders -- ADD, ADHD, dyslexia, autism, schizo-affective disorder, schizophrenia, et cetera. Because these disability groups lack the considerable power of a strong advocacy group, do you feel that they have been left by the wayside when it comes to Section 508 or WAI? (And do you personally believe that total-WAI compliance is necessary, or just Section 508?)

My apologies for several questions at once, but we take this issue very seriously here and your answers will go a long way to helping us do what we do to better suit the community that ILR serves.

Thanks so much, Samuel W. Knowlton

To answer your first question: A single interface works for most Web sites. You can simply make the site itself intrinsically accessible to most disability groups.

The only alternative the question seems to envisage is specifically custom-designing an alternative interface for disabled users. In other words, a site would exist in two or more predesigned forms. Thats not the only way.

Some work is being done to permit people and the devices they use to specify formats and capabilities they may possess or require. Have a look at Composite Capabilities/Preferences Profiles. It all boils down to semantic markup again. A single HTML page, if marked up properly, could be visited by a plain-Jane browser and displayed in a way thats familiar to nondisabled users; nothing special would happen.

But if you had a CC/PP-compliant browser or other device, and if the page were coded correctly, and if the server understood CC/PP protocols, then the page would automatically reconfigure itself to your needs without the original page authors having to do anything special. In fact, authors could not predict what kind of transformations would occur, nor would they care.

So a few things could happen. If youre totally blind, your page could be rearranged so the search box and content are at the top, with sidebars, navbars, and anything else uninteresting at the end and no images loaded at all. A low-vision person could ask for larger type on content sections and normal-sized type everywhere else, unless a command were issued to blow up, say, a navbar. (There could be continuous interaction between the user and the server.)

XHTML 2.0 might push this concept along a little, what with its section element, but its all still a pipe dream, really.

Now, as to the second question, putting blind and deaf people together in a group claimed to have an enormous amount of political clout is not really applicable to Web accessibility. Deaf people face very few accessibility barriers in using the Web multimedia is pretty much it. Blind people face very large barriers because the Web is a visual medium. Theres a qualitative difference.

Its true that people with cognitive disabilities have been neglected in Web accessibility. Why? Few people in the wider accessibility field have expertise on the topic. The Web Content Accessibility Guidelines 2.0, when its finished, will contain many more provisions for this group.

The qualitative difference remains. It is arguably difficult or impossible to make Web sites most of which are dominated by text genuinely accessible even to certain specific groups with cognitive disabilities. Remedies proposed in the Web Content Accessibility Guidelines 2.0 drafts would not guarantee access for learning-disabled people. Some of those remedies involve adding illustrations (non-text content) to every single page (yes, the Web Accessibility Initiative may issue that requirement) or rewriting the page according to some kind of half-arsed, doctrinaire editing scheme.

There wouldnt be the same jump in accessibility between a noncompliant site and one that meets those guidelines as you would find with, say, visual impairment. Sites would end up being merely less confusing as opposed to not confusing. You might have met the spec, but you could not be sure you had achieved accessibility.

Certain cognitive disabilities do not even require accommodation online.

Moreover, while accessibility for many other disability groups almost never alters the visual appearance of a page (visible skip-navigation links are a counterexample), it could be argued that a page thats truly accessible to people with learning or cognitive disabilities would have to be custom-created by experts. Thats the stark truth involved in achieving high accessibility for this group. You have to alter content as opposed to metadata or presentation. To accommodate other disabilities, you add information, like alt texts; to accommodate certain learning disabilities, you must remove or alter information.

I am in favour of improved accessibility for cognitively-disabled persons, but Ill only support proposals that can be shown to actually make sites accessible to that group. Im also not willing to destroy the Web as we know it ostensibly in order to save it for a disability group whose needs might not even be met in the process.

Nobody has presented credible evidence that current proposals actually will work, and certainly the evidence supporting the current WCAG 2.0 proposals is weak. In other words, if we want to fix this problem, its going to take a lot more work.

And to answer the final question, Section 508 regulations backhandedly incorporate almost all of the Web Content Accessibility Guidelines 1.0, but go beyond the latter in certain respects. Both guideline sets have all sorts of problems, but complying with either of them will assure reasonable accessibility for large numbers of people.

We got a lot of (shall we say) slightly impertinent questions for Web Standards Project co-founder Jeffrey Zeldman, but that's okay. He reads Slashdot and knows the nature of the beast, and he's hard-core enough to give as good as he gets. So set your humor module to high, then sit back and enjoy Mr. Zeldman's (appropriately impertinent) answers to the 12 questions we forwarded to him.

1) Here's my question:
(Score:5, Insightful)
by FascDot Killed My Pr

If you're such a hotshot web designer, why have you committed one of the cardinal sins of web design: Putting an "entry page" that does nothing but suck bandwidth and make it difficult to "back" out of a site?

Jeffrey:

I'll answer this one piece by piece.

"If you're such a hotshot web designer."

Never claimed to be. Roblimo wrote that glowing description. It's not surprising that some of you, who have no idea what I do, were pissed off when those words of high praise took you to a very simple, low-bandwidth, personal site.

I wish Rob had said "Zeldman is a co-founder of The Web Standards Project" (WaSP), and had explained what the WaSP does, maybe even mentioning the role we played in getting Netscape to throw out its old rendering engine and begin building Mozilla around the standards-compliant Gecko core. I'm guessing people would have overlooked my supposed "design sins" or their distaste for the color orange on my personal site if they had a better idea about what I actually do.

For those who don't know, the WaSP organized a petition drive to persuade Netscape to throw out its old rendering engine and build its new browser around Gecko. Then-group-leader George Olsen of WaSP, along with ThunderLizard's Jim Heid, got 2,000 developers to sign the petition. Netscape is a company that listens - at least, for the last two years, it has been listening - and you all know the result: an upcoming browser that is designed to fully comply with HTML 4, CSS-1, the W3C DOM, XML, and EcmaScript.

No disrespect to Roblimo either. I dig the guy. And what he said is true in a sense. I *am* a web designer and writer, and a lot of the work I've done over the past five years *has* gotten imitated, for better or worse. For instance, oddly enough, the original Mozilla.org (http://www.mozilla.org) was copied from the simple HTML-and-CSS layout I did The Web Standards Project (http://www.webstandards.org/): from the technique, to the color palette, to the crude four-pixel black outlines around content areas. Don't bother checking; the new Mozilla layout has evolved away from that original look, though it still bears trace elements of the original design. A lot of you probably do remember the original Mozilla layout. I'm sure when Roblimo saw it, he realized it was copied from http://www.webstandards.org and I think that's the kind of thing he was referring to in his overly kind introduction to my work.

By the way, I wasn't upset by what Mozilla did; I was flattered by it. You may think it is ugly design, though I'm sure that none of you said so to Mozilla because you believe in the project. It's weird to me that the same people who dig Mozilla would be rude in their comments to someone who, at least in a small way, helped influence the direction of that browser, and who also influenced the initial DESIGN of that project, but whatever. I also talk with Microsoft, because the goal of WaSP is to get standards in *all* browsers, and the fact that I talk to engineers at that company may make me evil incarnate in your book. I can deal with that. If we get better browsers, I'll be satisfied.

I do get copied a lot and often those copies are better than the original. In that sense "VIEW SOURCE" functions like "OPEN SOURCE." ;) I am happy when someone takes an idea of mine and makes it better (and their own).

"why have you committed one of the cardinal sins of web design"

I've been designing websites for five years. I don't claim to be a genius and I'm far from the best designer on the planet, but your take on cardinal sins of a profession you do not participate in is about as meaningful as my comments on your programming decisions would be.

You are parroting Jakob Nielsen or some other expert whose work you've read. You haven't read my work on the same subject (no problem) and you don't know my work as a designer (no problem). Just as in programming, design is about decisions. A designer never sins. He/she makes informed decisions. If you get to know the work, you may understand why those decisions were made. If you never bother to engage with the work - if you merely believe that all design must conform with a small set of rules written by one or two people - you don't understand the nature of the thing you are criticizing. Especially if you spend all of five seconds looking at it, and then rush to be the first to post a rant. There are rules of grammar, too, and James Joyce threw them all out. True, I ain't him. But I am me. All designers make decisions, and if the entire web looked like http://www.useit.com I don't think that would be such a great thing. Anything that departs from the look of http://www.useit.com is violating at least a few of Jakob's "rules," and that's the nature of the beast.

"Putting [up] an 'entry page' that does nothing but suck bandwidth and make it difficult to "back" out of a site?"

The bandwidth sucked is exactly 4K. I think you can handle it.

The "entry page" is a temporary placeholder while I rethink the front end of my personal site. Notice the words "temporary," "placeholder," and "personal." The previous front page was navigational in nature, and it is archived at http://www.zeldman.com/mozillatest.html . The name refers to the fact that the page revealed a bug in recent builds of Mozilla. I have left it online so the Mozilla folks can use it to track down and fix that bug, which they are doing now.

Recently, I've been focused on WaSP and A List Apart (http://www.alistapart.com/), a web design magazine and mailing list I co-founded with Brian Platz. The content on my personal site (275+ pages) has not been my most recent focus, so I determined a while back that it was silly to stop the visitor with a primarily navigational page. I should explain that some people visit zeldman.com for entertainment like The Ad Graveyard; a completely different audience visits for web design info, such as the Ask Dr Web tutorial; and so on. To accommodate those very different visitors, I initially had a core page that was navigational. I didn't put real content on page one, because I was accommodating maybe six completely different audiences, and there was nothing in all that content that would appeal to ALL of them. On http://www.alistapart.com/ I start with content on page one, because the audience is more unified.

Even that old navigational page (http://www.zeldman.com/mozillatest.html), with all its rollovers etc., was very low-bandwidth. I recently got to look at it while stuck at an airport in Stockholm. The airport had five Windows boxes sharing one 56K modem. I looked at some of my favorite sites, and they were all crawling onto the screen. I was pleased that my own front page loaded instantly. I design for low bandwidth, which explains why my work rarely looks like that of "such a hot shit designer." Back to the point. I haven't yet figured out how to restructure the front end of my site, so I put up a 4K placeholder with a bone-simple rollover and a 6 second refresh to the single page at my site that I have been focusing on lately.

Now you know why I have a temporary entry page; now you know why it does "nothing" (it is temporary, and what it does is redirect you); and now you know that it does not suck bandwidth.

How it "makes it difficult to 'back' out of the site" is a mystery to me, so I can't comment on that clause in your question. Personally I find database driven pages much harder to navigate and back out of than 4K html pages. And with browsers that suck, frames-based pages can also be tough to navigate. My 4K page is frameless HTML.

2) I have a question:
(Score:4, Insightful)
by Skinka

What's with that small font www.zeldman.com, haven't you read any (web) usability guides?

Jeffrey:

Yes, I've read them, yes I've written on the subject, yes yes yes.

Along with another WaSP member, I helped influence Microsoft to make ALL web text resizable by the user in IE5/Mac for reasons of accessibility and usability, and we are hoping to get the same from Mozilla. (At the moment, this feature is only available in IE5/Mac. It should be in every browser. It's not in the Windows version of IE and it's obviously not in the current version of Navigator.)

Why small fonts? Personal design decision on a personal site. You can enlarge the type in some browsers, not all. That day is coming.

There are methods of CSS that allow you to resize type in *all* browsers.

Why do I often avoid those methods?

Because they are not supported in most "CSS-capable" browsers.

Absolute font-size keywords are broken in Navigator 4 (all platforms) and IE5/Windows.

Percentages and ems are broken in Netscape 4.

Points are meaningless on computer screens, and the reliance on points in Style Sheets is a widespread authoring error.

Until all browsers support standards, designers will be stuck using pixels or FONT SIZE tags. Or simply making no effort at all to control the appearance and size of type on the web page. If this bothers you, join The Web Standards Project.

(Warning: it is orange. If you "can't get past the color" then I guess you'll have to let big browser companies determine the fate of the web.)

Unless you design web sites every day, you have no idea of the compatibility nightmares involved. But in your own work, I'm sure you have plenty of examples of brain-dead decisions by others that force you to use hacks and workarounds. It's the same in web design.

At the moment, the main text at http://www.zeldman.com/coming.html (the one page in all my work that most Slashdotters seem to have looked at) is laid out with ems. This is wonderful, scalable technology. You can easily enlarge or reduce the type in just about any browser.

Except, of course, that it doesn't work at all in most versions of Navigator 4. If you're using Navigator - and you know you are - you will see large ugly type, not the type treatment I intended. Until we have standards, that's just the way it will be.

3) Not to flame, but...
(Score:5, Insightful)
by mr.nobody

I find it hard to ask HTML questions to someone who has committed the cardinal sin of taking away the status bar with JavaScript.

Jeffrey:

Another cardinal sin.

Hmm. Let's see.

The status bar *does* reveal the url of the page it links to - just like an untreated status bar would do. It also provides ADDITIONAL INFORMATION AND COMMENTARY. I guess that's a bad thing. I can't see why, but I guess I'll take your word for it. URL = good, URL + additional information = bad. Because you say so.

The title tag also provides additional information, but Question #12 told me that that was okay, even good. Whew! That's a relief.

4) where's the interview
(Score:4, Interesting)
by geekpress

Jeff, I programmed for a web design company in which design issues totally trumped more practical concerns like download time. (In one case, I was forced to create absurdly complex html tables just so that the designer could get his one-pixel rounded corners on his notecard design.) What do you see as the appropriate balance between aesthetics and practical usability?

P.S. That company is now out of business, thank goodness!

Jeffrey:

I almost always design for low bandwidth.

I was creative director at a web firm and had designed a layout with thin black borders (yes I do this same thing over and over again) for a database driven site that would be creating tables on the fly. The design effect would not render in Navigator, but the page still looked fine in Navigator, even without those little black outlines.

It was possible to FORCE Navigator to display the effect by surrounding every table with an additional, empty table. That is sometimes okay, obviously - I've been doing it since we could begin applying rudimentary styling to tables - but on a large page full of data, it would unnecessarily increase the bandwidth per page, force all browsers to burn cycles as they calculated the appearance of complex table-in-table displays, possibly cause display errors, and completely yoke the content to the presentation, making it that much harder to fix later, when we have better browsers.

So I told the company president it was a usability nightmare and a waste of resources and bandwidth and therefore not worth doing.

He told me to do it anyway.

So I quit my job and started my own company.

Rounded edges, high bandwidth, all that stuff can be fine in the right situations, as long as alternatives are provided and rules of accessibility are respected. Usually I persuade my clients to go in the low-bandwidth direction, and I almost always go low-bandwidth on the noncommercial sites I do (zeldman.com, http://www.webstandards.org/ and http://www.alistapart.com/ ).

But high bandwidth is fine for the right audience. Consider http://www.praystation.com, which is a brilliantly designed site by Joshua Davis. It's amazing work. The audience for that site is primarily Joshua's fellow designers, and most of them have T1 or DSL access. Since the site's goal is to push design as far as it can go on the web, and since the audience is known to have fast connections and a desire to see great design, there is absolutely nothing wrong (and lot right) with the higher-bandwidth road taken by this project.

5) Optimism?
(Score:5, Funny)
by Chalst

How hopeful are you that Microsoft can be coaxed into making IE standards compliant? What exactly do you think Microsoft's motive was in not supporting HTML 4.0 completely?

Jeffrey:

It varies by the hour. Sometimes I think they are going to do this and simply have not committed to it because they're not sure they can pull it off. Sometimes I suspect that as the current market leader (guys with the most users) they think they don't have to bother with this. ("Our way *is* the standard." That kind of thinking.) And sometimes I reckon that they're doing this to fuck up Mozilla. ("You're going to support standards? Well, we have more users. You lose.") I don't *know* what they're thinking, but I suspect that different people there are thinking combinations of all the above.

I do know there are engineers there who are committed to supporting standards. Not only because I've met some of them through my work with WaSP, but also because - in the case of IE5/Mac - they've actually pulled it off. Remember, Microsoft (along with Netscape, Sun, invited experts, etc.) helped come up with these standards in the first place. Why would you design blueprints and then not follow them when you build the house? The engineers who participate in the standards process are committed to complying with standards. Some people in management may not be. Or they may be delaying, for short-sighted competitive reasons, or from fear of committing until they are sure they can do it right.

HTML 4 - the LAST HTML - includes dozens of accessibility improvements, and it is insane for any company not to fully support that. Without full support for HTML 4, millions of web users get hurt. That's morally wrong, and it's also just plain bad for business. I think that in time, all browser companies, including Microsoft, will come to see that. I also think the W3C's recent hiring of a conformance manager (http://xmlhack.com/read.php?item=517) signals that the W3C will soon take a more active role in "helping" companies get with the program, support standards, and stop screwing up developers and web users in a game where everybody - including the browser companies - eventually loses.

6) Balancing Technologies
(Score:5, Insightful)
by Proteus

As you are no doubt aware, the technology that drives web site design is advancing rapidly. However, there are still a lot of users who run older browsers, or prefer to use text-only browsers such as Lynx.

Obviously, one wants to reach as large an audience as possible, but not "lag behind" too far. How do you go about balancing the use of newer technology on a site without alienating users of older software, disabled users, and text-only browsers?

Jeffrey:

Using HTML 4, ALT tags, and the TITLE tag goes a long way toward achieving this goal.

So does using CSS for type, instead of FONT FACE and FONT SIZE tags that yoke content to presentation.

I do both these things, and all the other little things you have to do, for instance with framesets. I think there may be some really old (1996) framesets at zeldman.com where I left out full content inside the noframes tags. I'm cleaning that up as quickly as I can. At ALA http://www.alistapart.com/), wherever I used framesets, I included the full text inside the noframes tags, and I also included TEXT versions of all articles.

The next stage is full separation of content from structure, and that means using HTML 4 and CSS (and eventually, replacing HTML 4 with XHTML; and eventually, migrating to XML).

We can't safely do that yet. Gecko is still in development, Netscape 4 has appalling "support" for CSS, IE5/Windows has better but far from complete support, and the only released browser that gets it right - IE5/Mac - has a 6% market share.

SOON. Not soon enough, but SOON, we will look back on this era of stupidity and laugh. Oh, how we will laugh. This is the TRON era and we are striving to reach the MATRIX era.

7) Reverse scenario question...
(Score:5, Interesting)
by Jonny Royale

Have you ever seen anything come from a browser publisher "extending" a standard (Microsoft, Netscape, other), and thought "Gee, I wish that was in the standard"? Examples?

Jeffrey:

Yes.

LOW SRC was a funky old tag from Netscape (dating back to Netscape 1.1) that allowed you to slip a low-bandwidth image into place, and then have it replaced by the more bandwidth-intensive image when the latter finished downloading. For people with very slow connections, it was a useful hack. It also enabled creative web designers to add a certain amount of "SFX magic" (cough) to even the most primitive pages, viewed by the oldest browsers, under the most adverse conditions. That's gone. Too bad. I miss it.

Because of browser offsets in all released versions of Navigator and most versions of Explorer, I wish the "four horsemen of non-validation" (leftmargin, topmargin marginwidth and marginheight) had made it into HTML 4.0 transitional. We won't need them eventually, but until the browsers are smarter, we still do need them. The W3C is always ahead of what the browsers can deliver, of course; but by discouraging these dumb proprietary tags, the W3C has put us in the position where PAGES THAT WILL NOT WORK without these tags will fail at http://validator.w3.org. That kind of failure discourages developers from building standards-compliant pages. It is a small thing, and it is transitional, but DURING THE TRANSITION, I would have liked to see those four stupid tags get approval with a benevolent sigh.

On the other hand, designers who know what they are doing may include these tags and ignore those validation errors, but don't tell the W3C I said so.

Given the brain-dead way Navigator 4 and IE4/5/Windows handled absolute font size keywords in CSS, I *sometimes* wish font size tags were not discouraged YET. I hate them and hardly ever use them, but (for instance) there's no way to get small type in Linux that is actually READABLE without relying on these dumb old non-standard tags. What I really wish in this case, of course, is that Netscape and Microsoft hadn't fucked up this simple CSS technology. So I take the FONT SIZE tags thing back. Uh, never mind. I just wish Netscape and Microsoft had gotten CSS right the first time.

8) Banners
(Score:5, Interesting)
by TheTomcat

This is only vaguely related to design, but directly related to the web, and functionality.

We all know that banners don't work anymore. The only way a business can profit from banners is to show thousands per day. Most users don't even SEE banners anymore. We avoid them the same way we dig in the couch for the remote when commercials interrupt The Simpsons.

Do you have any suggestions to make future, content-based sites profitable?

Jeffrey:

There are several issues here. One is, a lot of the best work is done as a labor of love, and always will be. Those who need a revenue model before they are willing to even think about working will lose one of the golden opportunities of the web, which is free expression and the building of communities, regardless of financial issues. For instance, Slashdot was born as a community and still is one. Eventually, Slashdot got into a position where it could make money, but Slashdot is true to itself and was not corrupted or changed by any commercial considerations. So it is possible to make a good thing and not blow it when the cash register starts jingling. But a lot of other sites and communities have turned to dreck when money was involved.

We all agree that banners suck - Roblimo even wrote an article for ALA on that subject, back when ALA was just getting launched. With a big enough readership, banners *can* be profitable, as they are at Slashdot. But I agree that most of us just hate 'em.

Sponsorships are another possible means of revenue. "This issue of Webmonkey brought to you by Hewlett-Packard." With an entertaining HP minisite available at the click of a link, for those who care. Kaliber 10000 (http://www.k10k.net) has gotten Apple sponsorship, and all that means is, there's a tiny Apple link in the top right hand corner of the front page. If you click it, you get a popup window with text on why the site's designers like their Macs, and links to some current movies in Apple Quicktime format.

The Cluetrain guys have spoken about this model of corporate sponsorship as well.

I think about it sometimes. For instance, http://www.alistapart.com/ could be "brought to you by" Macromedia or Adobe. But to tell the truth, I don't really pursue this idea because I'm not motivated by money when it comes to creating web content. I simply want to create or choose the right content, and totally control it, and I'm not sanguine that I could do that if I *had* corporate sponsorship. Thinking about it some more is on my to-do list, but it's about 500 layers down in the list. I make enough money designing websites that I don't worry about "revenue models" for my content sites. It is a real issue, though. Just one I haven't bothered with personally, yet.

9) Jeff, your CSS suck
(Score:4, Insightful)
by Nicolas MONNET

I quote from your website:

H1 {font: bold 24px verdana, helvetica, arial, sans-serif; margin-top: 0xp;}
H4 {font: 12px verdana, helvetica, sans-serif; margin-top: 0px; margin-bottom: 10px;}

So why, tell me, WHY did you use PIXELS (px) instead of POINTS (pt), thereby overriding my painfully crafted DPI settings, rendering your all page unviewable on my Linux machine?

Jeffrey:

Refer to the answer to Question 2. Also refer to this Word from the WaSP column:

http://www.webstandards.org/wfw/ieah.html

The best way to style text - and the way the W3C recommends - is to use relative sizes or absolute size keywords.

Both these methods are completely broken in Navigator 4. Totally frickin' useless. Don't shoot the messenger. Netscape agrees, and that's why they threw out their old rendering engine and started from scratch.

And absolute size keywords are stupidly mis-supported in IE4/5 for Windows, where "medium" means large, and "small" means medium.

Faced with this maddening stupidity on the part of browser makers, designers/developers have two choices:

Do not style text at all. Have a nice day.

*OR* rely on pixels, which work in all "CSS-capable" browsers.

I sadly choose the latter until the browsers fully comply with W3C standards.

As to POINTS versus pixels, points are absolutely meaningless on the web, and the fact that they are used by thousands of developers who should know better proves only how little CSS is understood by the development community.

Certain point sizes may work on your platform in your style sheet. That proves that certain point sizes work on your platform in your style sheet. Cross-platform it is not transportable, and points are print-based units of measurement that have no meaningful relationship to the wonderful world of monitor resolution.

For a good discussion of CSS problems, see Todd Fahrner's "Beyond the Font Size Tag: Practical HTML Text and Styling" at http://style.metrius.com/font_size /livetext.html (Unfortunately, even some of *THESE* techniques do not work in more recent versions of Navigator 4.)

In a few months, there will be exactly two browsers that get CSS-1 right: Mozilla/Nav 6 for all platforms, and IE5/Mac which we have now. Since neither has dominant marketshare, developers will still face huge obstacles when trying to do something as SIMPLE and BASIC as size text on the web. Many will stick with pixels, which are the only CSS technique that actually WORKS across browsers and platforms.

In addition to all these nightmarish problems with our browsers, there are special challenges with Linux, because unless Linux users install additional scalable fonts, you can follow all the rules for good CSS, and avoid "problem" font sizes, and still create pages that look jaggy or are unreadable on a lot of people's machines. I worry about this all the time, but I don't have a solution for it. I have actually gone back to using the stupid The way to advance the medium is to get absolute font size keywords and relative font sizes right in CSS, finish implementing HTML 4, and give us the W3C DOM, XML, and EcmaScript. (And then wait two years for users to upgrade.)

10) Pixel based alignment and HTML
(Score:4, Insightful)
by mcelrath

One of the most disturbing trends that I see in web design these days is the trend toward trying to control layout at the pixel level. As HTML (Hypertext Markup) was not intended to be a graphics language, what is your comment on this?

Jeffrey:

Separation of style and content is the way forward.

The problem is the browsers.

When I revised my "Ask Dr Web" tutorial at http://www.zeldman.com/askdrweb/, along with other pages at http://www.zeldman.com/, to use CSS layouts instead of tables, certain versions of Navigator 4 began crashing.

Actually crashing from basic CSS-1.

I wrote about this at A List Apart, ("The Day the Browser Died") and because of this, Netscape invested some time and resources to fixing some of these bugs in Navigator 4. It didn't catch them all, and it didn't catch them in Linux. These bugs will never be fully fixed in Navigator 4, because Netscape is wisely spending its energy to finish the Mozilla browser. Unfortunately, this means that Netscape users will continue to face serious usability hazards throughout the web until Netscape 6 is released ... *OR* it means that developers will continue to use TABLES for layouts for the next two years (as Jakob Nielsen has predicted).

If you look at these pages - http://www.zeldman.com/steal.html or http://www.zeldman.com/icon.html are other examples - you will see that we are talking about extremely BASIC layouts. An expert from the CSS pointers group actually volunteered hours of her time trying alternate combinations of the very basic CSS on those pages to see if she could find ways to stop Netscape from crashing. She could not; neither could I. Netscape did what it could for its 4.0 users, but it can't do anything more until the next generation is released.

On my personal site I made the tough decision to leave these pages as-is. I don't have time to recode them all using tables.

You can agree or disagree with that decision.

Linux folks can either use the Mozilla or Opera betas to navigate those pages in safety and comfort.

It's worth noting that W3C pages also crashed Netscape 4, for the same reason.

What happens when Netscape's browser is this badly damaged? I get hate mail from people who don't understand the issues involved. I also got a letter of thanks from Netscape's Eric Krock, because good companies WANT us to help them find bugs in their software.

As an example of this, many sites (including yours) use font size=1 to acheive a font that is fairly uniform in pixel size across browsers. Anyone with a high-resolution screen will tell you that this is highly annoying, since it results in an almost unreadable font.

See above for the explanation as to why developers are stuck using 1994 technology to support late-1990s browsers. The same questions, the same answers.

The good thing - the ONLY good thing - about the font size tag is that it is user-resizable. The rest of this has been answered above.

Forcing netscape to use a larger font size often destroys the layout of the page. What's worse, some pages use dynamic fonts and other features to force this on the user.

Right, although in some cases it is justified.

As another example, many pages use the table , and layer to specify the exact size in pixels of portions of the page, and then put a little notice at the bottom ("This site best viewed at 800x600") or some such.

Yes, that is usually a bad design decision. Whenever possible, I use what Glenn Davis (WaSP and Project Cool co-founder) calls "liquid" design ... design that reflows to exactly fit the visitor's monitor. That's almost always a better way to go. Examples of my liquid designs include http://www.alistapart.com/, http://www.the-adstore.com, and most of zeldman.com. If you dig long enough in zeldman.com, you'll come upon pages older than the NYC subways, that simply use BAD design ... though at the time, it wasn't all that bad.

Liquid design is not always appropriate but it is generally best.

What are standards groups doing to fix this?

Nothing. The W3C can't make better or more intelligent designers out of people, and neither can the WaSP, whose sole purpose is to agitate for W3C standards in browsers (and eventually in web authoring tools).

We can try to lead by example. http://www.webstandards.org/ is liquid (aside from the front page, which is "semi-liquid" owing to the large low-rez graphic) and it validates.

MEMBERS of standards groups can write articles on the subject and hope that people read them. Of course, if people "can't get past" a 4k splash page, they will not learn about my articles on the subject.

Will I be looking at pages designed for 800x600 (or worse, 640x480) with my 1920x1440 screen forever? Will persons with laptops at 640x480 be unable to read the web soon? Will standards bodies ever require percentage-of-screen width and height specifiers, or even better, implement table width=30ch to specify sizes in relation to the current font size?

Standards bodies can recommend certain authoring practices, and they can develop standards that make such practices possible, but they cannot enforce good authoring.

11) Evaluate Slashdot
(Score:5, Interesting)
by Pseudonymus Bosch

What would you change, what would you add, what would you remove in Slashdot?

Jeffrey:

It's a community and it works. It has achieved visibility, notoriety, and even commercial success without giving an inch. Pretty awesome achievement. What would I change?

Sometimes the longer threads take a long time to load, due to back-end technology, platform and server issues. The technology works better on Linux than it does on my platform of choice (Mac OS) but, hey, that's okay.

I know you want me to comment on the design. Design is subjective. Black backgrounds and teal are not my favorite color scheme (though I used black backgrounds at the 1995 Ad Graveyard (http://www.zeldman.com/ad.html) and the January 1997 Furbo Filters so who am I to talk? The main thing I was trying to do at Furbo was get CSS to work - and to let people know about Craig Hockenberry's and my Furbo Filters, which were the only Photoshop plugins at the time that dealt with the web-safe color palette - at least to our knowledge.)

I might change the color scheme and some other things if Rob Malda went on a crack run and asked me to redesign Slashdot, but that ain't likely to happen. And I think the design of Slashdot is just fine. It focuses you on the breaking stories, allows you to read more (or not), and provides access to almost everything else on the site via small navigation units. In terms of usability it is damn good, and it has plenty of attitude.

Of course, it commits the "cardinal sin" of teal, but I can get past that.

12) Do you agree with Nielsen?
(Score:4, Interesting)
by Pseudonymus Bosch

I have no idea about you and your views, but I have read lots of the Alertbox columns by Jakob Nielsen.

Do you agree with him? Do you disagree? What about?

Jeffrey:

I agree with his comments on oral sex and wearing white after Labor Day.

And I like that he can get $25,000 to talk for an hour. I'll do the same for half that amount.

I also agree with Jakob that most websites should be usable by as many people as possible.

What I have done about that is help found The Web Standards Project, so we can actually achieve that goal instead of using duct tape and lasagna to build sites that work for "most" people. And I try to make my pages accessible in spite of the limitations of current browsers and some of the cross-platform issues discussed above.

If you are interested in my views, you can read them at http://www.alistapart.com/, Adobe.com (here, http://www.adobe.com/we b/columns/zeldman/20000320/main.html, for instance), http://www.webstandards.org/ and of course at http://www.zeldman.com/. If you can get past the 4k splash page.

At least you share the use of TITLE attributes in hyperlinks (a good feature that Slashdot shouldn't chomp away).

Thanks! The reason Slashdot chomps title tags is probably because they are not supported in Netscape yet. They are an important usability feature, and in some browsers they also offer nifty low-grade special effects - along with the opportunity for contextual ampliciation or ironic commentary.

jeffrey

Can't act. Can't sing. Can dance a little.

http://www.zeldman.com
http://www.alistapart.com
http://www.happycog.com
http://www.webstandards.org