Slate is Microsoft's new attempt at a "serious" online magazine edited by Michael Kinsley. Slate is mainly a failure due to its inability to adjust to the online medium, though it sports several promising innovations and appropriate uses of the Web.
This review is based on early experience with the first issue of Slate . In fact, I have not read every last article in this issue since doing so would be too hard to endure. Slate is not designed for online reading: the feature articles are all long, long, long . In fact, often longer than the content warrants. Of course, users could simply print out articles, and the site does provide a printable version for download.
If I wanted to read long printed articles, I would be better off with The New Yorker : much more compact (thinner glossy paper that is better bound than the output of my printer) and better cartoons. Slate does have a comic strip but it is not well suited for the online medium: it is simply a scanned drawing that has been anti-aliased to a not-very-attractive three shades of gray. At 610 pixels wide by 849 pixels tall, only users with large monitors will be able to see all of the artwork. It would have been more interesting to see Slate use its budget and obvious avant-garde leanings to pioneer mainstream use of interactive cartoons along the lines of Mark Hurst's work at the Media Lab.
Even though the articles are too long for comfortable on-screen reading, their presentation is obviously designed for online viewing with liberal use of HTML tables to lay out multiple columns that look better on the screen than on a printout where they lead to excessive waste of paper. The design is claimed to be optimized for Internet Explorer and even though some HTML purists will no doubt deplore this, I find it hard to criticize Microsoft too much for this sin in the current climate on the Web. At least there are no frames.
Use of the Online Medium
The use of hypertext is very conservative with no use of embedded links in the body text. Most articles end with a small set of selected links to background materials or contrasting views but there is no attempt at non-linear writing within the magazine itself. A useful set of links is provided in a column called "in other Magazines" that surveys major stories in other magazines and includes links to any online versions. This column is also one of the few examples of innovative uses of the online medium in Slate : comments on other magazines obviously become much more interesting when the target is a click away from the commentary. Also, Slate takes advantage of its immediate distribution to provide previews of these other magazines before they reach their subscribers.
There are two other uses of the online medium where Slate shows promise compared to traditional magazines: the first is multimedia which is used in very appropriate ways to enrich the stories. No spinning GIF89a's (bravo), but a clip from an Ella Fitzgerald song in her obituary and a video clip of a Republican anti-Clinton commercial in a (very good) column analyzing political advertising. The analysis of the anti-Clinton commercial gains significantly from the user's ability to actually see the video and observe the pacing, tone of voice, etc., and allows the author to provide a much deeper analysis of cinematic effects than one normally gets in the "let's review a commercial" articles in The New York Times where the print medium forces a focus on the text. As we all know, movement, color, and tone of voice have much more emotional appeal than the actual words, so deconstructions of political campaigns need multimedia.
The second good use of the online medium is the so-called "committee of correspondence". This is a too-long (40 screenfulls on a medium-size screen) scrolling debate between five pre-selected participants. The topic of the debate in the premiere issue is "Is Microsoft Evil?" and was obviously chosen to emphasize the editorial freedom of Slate relative to its publisher. The debate format is innovative in allowing the participants to post new arguments daily throughout a week. By restricting follow-ups to once-daily and by having a select group of contributors, the editors avoid flaming and encourage reasoned writing. Even so, the actual quality of the writing is not on the superb level one would expect from an elite magazine. Sure, we are a level beyond netnews threads, but not several levels up. There are not many new arguments relating to Microsoft's possible evilness in the Slate discussion that have not been seen many times in various comp.lang.java flamefests. You know: "Windows'95 = Macintosh'89" on the one hand and "millions of users can't be wrong" on the other. Maybe the topic has been beaten to death on the Internet already and we will see better results in future Slate debates. One benefit of the slow turn-around and selective list of participants is that rebuttals tend to be more credible and researched than the typical netnews comment. For example, when somebody said that Microsoft never innovated but always had second-rate products, Steve Ballmer (the only microsoftie in the debate) could reply by listing their R&D budget and the new recruits in Microsoft Research as well as provide links to several "editors' choice" awards won by certain products. At the same time, when Ballmer tried to minimize the anti-trust issue, there was a rebuttal by an anti-trust lawyer who quoted the actual Department of Justice court papers. As I said, better than your average flame and a level up from netnews, but no new insights.
Now turning to the design of Slate 's website, I sincerely doubt that they have conducted much usability testing since several of their design elements are likely to cause usability problems according to my studies of other websites. The first and most obvious is the use of two home pages which is often confusing to users. The initial home page contains an abbreviated list of the main feature stories and a small subtle link to a fuller table of contents. Following this link leads the user to a larger home page that cannot be seen on a single screen due to its very spacious layout. In fact, it requires 3-4 screens to see the full list and some of the main stories are not visible in the first window.
The table of contents uses several "cute" headlines to link to stories (example: "Varnish Remover" is the link to the analysis of campaign commercials). On the Web, cute links normally don't work since users rarely take the time to download stuff they don't know what is. In print, playful headlines work because the reader can easily glance around the magazine to see what the story is about. On the Web, every click carries a penalty and removes the user from the context of the previous page, so link anchors must be exceedingly intuitive.
In most articles, the author's byline has been made a link to a short (2-3 lines) bio at the end of the article. This type of author bio is left over from legacy magazines and a poor use of the power of the Web: instead, the short bio could have been left in the article as information for users with little interest in the writers and the link should have taken those users who do care about writers to a longer bio with photo of the author and links to other stories by the same author. Interestingly, the only appropriate biography is the one provided by a New York Times columnist who is a contributor to the Microsoft-evilness debate (and the bio lives on his personal website ).
All the web pages in Slate have the same <TITLE>, thus making it difficult to bookmark individual pages or to navigate using history lists. Presumably, the reason for giving all pages the title Slate is to reinforce the notion of a unified magazine as opposed to a collection of stories, but the nature of the Web is to treat pages as the unit of navigation: there are no staples on the net.
The link anchors in the table of contents are all the same color, no matter whether the user has visited the destination article or not. Non-standard link colors are always bad (as further discussed in my column on the top ten mistakes of web design ), but using the same color for visited and unvisited links makes it impossible for readers to see at a glance what articles they have not seen before.
A final usability problem is so grotesque as to be almost absurd: the use of actual page numbers in the navigation bar. Each web page ends with a navigation bar with 23 numbers that take the user to other articles. Oh sure, I really feel like reading article 13 next! Even worse, because different issues have a different number of articles, the same article number does not always refer to the same part of the magazine. For example, letters to the editor was number 21 in the first issue and number 20 in the second issue. What a waste of the navigation bar; it should have been used to teach users the structure of the hyperspace and provide real navigation shortcuts.
Note added mid-July, 1996:
After three issues of
, I stand by my criticism. The editors have partly fixed the page <TITLE> usability problem: recent issues have had a different <TITLE> for most articles. Also, the cartoon has slimmed down to a more viewable width of 586 pixels and introduced modest interactivity. These changes highlight a benefit of online publishing: the possibility for continuous improvement in reaction to user feedback.
The Committee of Correspondence continues to show promise without actually delivering: the week on tax cuts rehashed very familiar arguments and the week on missile defenses was boring. One can only stay "promising" for so long, so I am starting to believe that Slate 's debate format is the wrong one. I think there might be more value in HotWired 's "brain tennis" format where two debaters alternate contributions. Having only two contributors makes it easier for the reader to follow the debate and probably encourages better writing.
(I can't believe it; this is the second time this year I have praised HotWired . Of course, their fonts and colors suck and the threads UI is an experience in confusion. There, I feel better already.)
Note added August 16, 1996: After seven issues, Slate finally improved their user interface on August 16 by unifying the two home pages and making the new (single) home page much more compact, as can be seen in a side-by-side comparison . The real lesson is that a single user interface expert could have (and did!) made these recommendations after an afternoon's worth of interaction analysis whereas it took a bunch of content experts almost two months to arrive at the same conclusion. A powerful argument to involve user interface folks in Web design, if I do say so myself.
Note added November 5, 1997: Michael Kinsley finally admits that the early Slate had too long articles and had adapted too little to the requirements of the Web as a medium. Better late than never, though it still bugs me that they take 15 months to acknowledge something that was obvious to a Web usability specialist after seeing one issue.
(I originally had links to 4 other reviews here, but they have all vanished from the Web. Linkrot , sigh.)