The Web needs Xanadu features

David Carlson mentions Ted Nelson in an E-Media Tidbits post Tuesday. Nelson is a true pioneer — he coined the word “hypertext” before people had personal computers. He’s spent much of his life trying to develop a system he called “Xanadu,” an idea that predated the World Wide Web. One of Xanadu’s features would have been a micropayment system, as Carlson notes. Other features have been largely impractical on the Web, like bi-directional linking and the ability to address individual paragraphs, sentences or even words in someone else’s document. That last feature was critical to Nelson’s vision of a document publishing system where anyone could quote any other document, comment on it, or rearrange it and make something new. The original author would be paid via those micropayments according to the percent of the original work included in the derived work.

Blogs approximate Nelson’s vision of a system for free discussion and quotation, but the design of HTML only allows them to link to an entire document. They can’t link directly to a particular sentence; the reader would have to scroll and skim through the document to find it. And bloggers definitely can’t quote from an original work without literally copying the quote. Xanadu would have let a reader’s browser go grab a particular sentence on the fly from the original publisher’s server and insert it, eliminating copyright problems and defining documents’ relationships within their markup code. The Web would be better with this idea of “transclusion.”

That said, I think Nelson’s current rants ignore the benefits of Tim Berners-Lee’s Web. Among them: Browsing via home and office computers rather than in proprietary Internet cafes, a simple decentralized system that encourages innovation, and use of the Web interface for purposes other than the Xanadu “docuverse,” like online shopping.

Incidentally, Ted Nelson is a fellow graduate of Swarthmore. The alumni magazine did a good article on him and Xanadu a few years back, describing how Nelson has been staggeringly unsuccessful at ever building his Xanadu system. What online publishing needs is someone practical who could implement some of Nelson’s ideas on the existing Web.

Verdana and Georgia harder to get

Over the weekend Slashdot had an item about Microsoft ending public availability of its “TrueType core fonts for the Web.” In other words, Verdana and Georgia can no longer be downloaded from Microsoft’s site. Probably a majority of personal computers do have these fonts installed, and I’m sure the fonts will remain in downloads of Internet Explorer for Windows and Macintosh, but the fact that those fonts are no longer available separately is bad news.

Georgia and Verdana (the font your browser will use to display this site if you have it installed) are two typefaces designed specifically for reading on computer screens, which have much lower resolution than a printed page. Giving fonts like that a wide distribution and making reading easier for everyone is definitely in the interest of site developers and site visitors, but apparently not in the interest of Microsoft. That’s contrary to their earlier stated intentions: “[t]he whole point is to make them available to as many users as possible” (thanks to Adrian Holovaty for that link).

Having fonts designed for easy reading onscreen will also be critical for the success of so-called “tablet” PCs, which Microsoft is pushing heavily. I am actually a huge fan of the tablet PC concept. I think it would be great to be able to browse the Web in bed or read an online newspaper at the breakfast table via a lightweight device with a decent size screen and wireless networking. It’s just a shame that the fonts to make that happen will probably now be proprietary to Microsoft’s tablet devices.

Designing good fonts is hardly an undertaking for amateurs, but someone else should get to work making some more good onscreen typefaces. Verdana and Georgia are almost overused anyway (although probably not when compared to Times New Roman).

Let users find what they want

I’m in the middle of reading Hot Text: Web Writing That Works, by Jonathan and Lisa Price. It’s a good book, even though not all of its common-sense advice is directly relevant to journalism. The Prices suggest how to shorten, simplify and structure online writing. Text needs to be more navigable on the Web than it does in print, they say, because reading online with a lower screen resolution is more difficult than reading on a high-resolution printed page. As a result, readers scan the page rather than reading and only stop to read what they are interested in. The more boldfaced items, lists, headings and short paragraphs the author uses, the more scannable the text becomes.

Facts like this about the very nature of reading online mean news sites need to not just organize front pages better. They also need to write more scannable articles. But by far the most important point here is that readers only read things they are interested in (duh). It’s harder online to use a clever lead or a good picture to get someone to read an article if they aren’t interested anyway, especially since reading that article means clicking a link and waiting for the next page to load. As I’ve said before, I think news sites need to offer more highly specific, high quality, customized information rather than more broad, generic information. If people are looking for a specific bit of information or are interested in the topic, they will stop and read. Otherwise, they’ll just scan by what they find irrelevant, then leave.

Technical restrictions hinder customization

Vin Crosbie makes an interesting point in a Tuesday message on the Online-News discussion list. (Unfortunately Poynter’s Web interface to the list, in a bit of Web stupidity, won’t let me link to his e-mail. It’s dated August 13 and its subject is “RE: [ON] Re: Salon article,” if you want to try to find it.) Vin says, “An analog printing press must produce the same edition for all users at a time (analog broadcast transmitters have this same constraint). The problem with that is everyone doesn’t share the same interests.” He concludes that newspaper and magazine Web sites are reproducing that one-size-fits-all model of publishing on the Internet. Thus, many of them are missing a glaring opportunity to customize their content to specific readers’ interests.

A few weeks ago, I asked why news sites were doing so much broadcasting and so little “narrowcasting”. Vin suggests that inertia may be part of the answer. I think silly technical limitations play a part too.

If the printing press prevents customization offline, static HTML prevents it online. I’ve seen journalistic possibilities curtailed by ridiculous server constraints in all three newsrooms where I’ve done Web work. My college paper doesn’t have access to databases on its Web server, which means it’s never built a content management system. At the daily where I worked last summer, we had to go to the corporate level to add categories to the CMS. I left at the end of my three-month internship and it still hadn’t happened. But at least at both of those places we had access to server-side scripting. At the major metro daily where I am interning now, we cannot write CGI scripts, let alone use databases.

Why is it that people like me can run server scripts and databases on our personal sites for less than $100 a year in hosting fees, but producers at sites with huge budgets and dedicated datacenters are stuck with static HTML and no control over content management systems?

Whether it is due to the blinders of print edition thinking or just incorrect priorities, I am convinced these technical restrictions constrain producers’ imaginations and prevent online journalism from living up to its potential. It’s as if there was some arbitrary restriction requiring print edition newspaper stories to be published sequentially in galleys, sort of like a paper from the 1850s. Yes, the text would be the same, but page design is the defining characteristic of print. Online news sites can go on publishing text in galleys — or they can give up foolish bureaucratic and technical restrictions and let journalists do more with the medium.

Useless URLs at two papers

In some random Web browsing today, I came across completely amateurish coding decisions at the sites of two newspapers with quite respectable print editions.

The Austin American-Statesman recycles URLs. Take, for example, http://www.austin360.com/statesman/editions/monday/news_1.html. “news_1.html,” in a folder called “monday”? Next Monday, the next “news_1.html” will overwrite this week’s version. I understand (even if I do not agree with) business decisions to remove articles from a site after a certain number of days. In that case, readers’ bookmarks will break and links from external sites will break — and visitors will get a “not found” error page. On statesman.com, bookmarked or linked URLs do indeed break after seven days, but instead of an error message the visitor will receive a different article. There is simply no valid technical reason to cause this confusion for readers.

The Orange County Register uses frames. Hello? Frames? This is the year 2002. I think people have known how annoying it is to have URLs hidden in the browser’s location bar since about, hmmm, 1996. Yes, browsers do let you bookmark pages in frames, link to them and share the URLs with friends, but some Web users do not know how to do those things. And the Register is destroying the usability of their URLs for the sole purpose of displaying a banner ad at the bottom of the page? Please. There are modern ways of doing that which do not require frames. At the very least, the Register could program some content management system to automatically spit out a separate FRAMESET for each article, keeping their framed design but providing bookmarkable URLs. The Register’s current design manages to not only annoy all users, but let sophisticated users bookmark stories and return to them without viewing the main ad (it’s in the bottom frame). This sure is a profitable and user-friendly set of HTML code if I ever saw one.

Archiving front pages

Poynter.org does something more sites might want to try: a day-by-day front page archive. A third-party site, the Wayback Machine, can do something like this already, but in-house versions would offer better access speed and finer time resolution (not to mention the ability to serve ads on the page views). ajc.com, where I work now, already archives its home pages for internal use, and I bet many other sites do as well. Writing a simple script to let visitors to navigate to them would be trivial.

As a side note, poynter.org’s centerpiece today happens to be about newspaper usability. In it, Monica Moses argues not just for better design of pages, but better design of text. “Some of the news ought to be formatted as bullet points, lists, Q&As, and segmented text,” she says. The article focuses on print, but her suggestions are even more important for text that’s read online because small, low-resolution computer displays make reading itself that much more difficult.

Correcting errors, or not

Reporters and editors make mistakes — that’s human. But readers of many news sites would have to be superhuman to discover what they were. Many if not most sites do not even include a link to “corrections” in an obvious place like their left navigation rail. Do they even post their corrections? Of the few sites that do offer a prominent way to get to their print edition corrections, I have seen no site so far which does it in the way I think it should be done.

First, the good: The Washington Post’s online corrections system gets very close. Their corrections page is not the text of the corrections themselves, but a list of “recently corrected articles.” There’s another page that archives the print edition corrections by the date the correction was published. Having both of these ways to browse the corrections is excellent. On each corrected article, the correction appears in a can’t-miss spot, to the right of the lead right under the headline.

The only thing missing from this corrections system is actually making the correction in the full text of the article. The Post’s prominent box that currently runs the text of the correction could instead say something like, “this article has been corrected since it was originally published.” Then there would be a link to “see the original version,” which ideally would have an option to show the changes with strikethrough formatting. I can’t stress enough how important I think it is to correct the text of the original article so the articles people bookmark or share with others do not contain factual errors. It is not enough to correct those factual errors on another page elsewhere on the site. In addition to that, though, the original article needs to be preserved for the historical record. The Washington Post’s simpler system of running corrections with the original articles serves the reader almost as well but does not exploit the Web to its full potential. It deserves a B+.

By comparison, the New York Times gets a D. Its corrections page is available in an extremely prominent position in the site’s left rail, but that’s about all the Times has going for its online corrections. Not only does the corrections page not link to the articles that are being corrected, but if you try the laborious process of finding those articles yourself you discover that the corrections page lists their print edition publication date and section, which is not necessarily where they appeared online (example: there is no “Weekend” section on their Web site). Furthermore, the corrections page only has today’s corrections — there doesn’t seem to be a way to get to corrections that ran yesterday. Emblematic of the lack of significant thought put into this page is the fact that the e-mail addresses offered for visitors to submit corrections are not even “clickable” (i.e., coded as “mailto:” links). But the Times’ worst offense by far is that the original articles are untouched. There is no editing of the original text. There is no insertion of the correction onto the article page, as the Post does. There is no indication whatsoever that the Times has published a correction about a story, so as far as online readers would know, it never did.

The Times does beat out the Chicago Tribune, which deserves a D- for its corrections page. Unlike the Times, the Tribune does archive its corrections from previous days. But the front page of corrections is just a list of links to (and first paragraphs from) all of those archived pages of corrections. All the headlines are the same (“Corrections and Clarifications”) and they don’t even list the publication date of each link. Someone needs to mess around with the content management system a little more.

Many other sites seem not to be linking to corrections prominently, or even posting corrections at all. Take my hometown metro daily, the Philadelphia Inquirer. The word “corrections” does not even appear on their site index page, nor on those of some other Knight-Ridder sites I checked. These papers clearly deserve an F. What are we going to do about the fact that so much of the class is failing?

Update August 6, 2002, 2:16 pm: Adrian Holovaty has a great list of how more sites publish corrections.