[My live talk at the Shape of Things to Come conference at the University of Virginia, March 27, 2010. It is a riff on a paper that will come out in the proceedings of the conference.]

As I noted in my paper for this conference, what I find interesting about this panel is that we got a chance to compare two projects by Ken Price: the Walt Whitman Archive and Civil War Washington. How their plans and designs differ tell us something about all digital humanities projects. I want to spend my brief time spinning out further what I said in the paper about control, flexibility, creativity, and reuse. It’s a tale of the tension between content creators and content users.

But before I get to Ken’s work, I’d like to start with another technological humanist, Jef Raskin, one of the first employees of Apple Computer and the designer, with Steve Jobs, of the first Macintosh. Just read the principles Raskin lays out in 1979 in “Design Considerations for an Anthropophilic Computer”:

This is an outline for a computer designed for the Person In The Street (or, to abbreviate: the PITS); one that will be truly pleasant to use, that will require the user to do nothing that will threaten his or her perverse delight in being able to say: “I don’t know the first thing about computers.”

You might think that any number of computers have been designed with these criteria in mind, but not so. Any system which requires a user to ever see the interior, for any reason, does not meet these specifications. There must not be additional ROMS, RAMS, boards or accessories except those that can be understood by the PITS as a separate appliance. As a rule of thumb, if an item does not stand on a table by itself, and if it does not have its own case, or if it does not look like a complete consumer item in [and] of itself, then it is taboo.

If the computer must be opened for any reason other than repair (for which our prospective user must be assumed incompetent) even at the dealer’s, then it does not meet our requirements.

Seeing the guts is taboo. Things in sockets is taboo. Billions of keys on the keyboard is taboo. Computerese is taboo. Large manuals, or many of them is taboo.

There must not be a plethora of configurations. It is better to manufacture versions in Early American, Contemporary, and Louis XIV than to have any external wires beyond a power cord.

And you get ten points if you can eliminate the power cord.

Many digital humanities projects implicitly believe strongly in Raskin’s design principle. They take care of what to the content creators and designers seems like hard and annoying work for the end users, freeing those users “to do what they do best.” These editorial projects bring together at once primary sources, middleware, user interfaces, and even tools.

Like the Macintosh, this can be a very good thing. I mostly agree with what Ken has just said, that in the case of Whitman, we probably cannot rely on a loose network of sites to provide canonical texts. Moreover, students new to Walt Whitman can clearly use the contextualization and criticism Ken and his colleagues provide on the Walt Whitman site. Similarly, scholars dipping for the first time into ethnomusicology will appreciate the total research environment provided by EVIA. As Matt Kirschenbaum noted in the last session, good user interfaces can enable new interpretations. I doubt that many scholars would be able to do Hypercities-grade geographical scholarship without a centralized Hypercities site.

But at the same time, like Raskin, sometimes these projects strive too hard to eliminate the power cord.

Raskin thought that the perfect computer would enable creativity at the very surface of the appliance. Access to the guts would not be permitted because to allow so would hinder the capacity of the user to be creative. The computer designers would take care of all of the creativity from the base of the hardware to the interface. But as Bethany Nowviskie discussed this morning, design decisions and user interface embody an argument. And so they also imply control. It’s worth thinking about the level of control the creators assume in each digital humanities project.

I would like to advance this principle: Scholars have uses for edited collections that the editors cannot anticipate. One of the joys of server logs is that we can actually see that principle in action (whereas print editorial projects have no idea how their volumes are being used, except in footnotes many years later). In the September 11 Digital Archive we assumed as historians that all uses of the archive would be related to social history. But we discovered later that many linguists were using the archive to study teen slang at the turn of the century, because it was a large open database that held many stories by teens. Anyone creating resources to serve scholars and scholarship needs to account for these unanticipated uses.

When we think through the principle of unanticipated uses, we begin to realize that there is a push and pull between the scholar and the editor. It is perhaps not a zero sum game, but surely there is a tension between the amount of intellectual work each party gets to do. Editors that put a major intellectual stamp on their collection through data massaging and design and user tools restrict the ability of the scholar to do flexible work on it. Alan Burdette of EVIA was thinking of this when he spoke about his fear of control vs. dynamism this morning.

Are digital humanities projects prepared to separate their interfaces from their primary content? What if Hypercities was just a set of KML files like Phil Ethington’s KML files of LA geography? What about the Grub Street Project? Or Ken’s Civil War Washington? This is a hard question for digital projects—freeing their content for reuse.

I believe Ken’s two projects, one a more traditional editorial project and one a labor of love, struggle with how much intellectual work to cede to the end user. Both projects have rather restrictive terms of use pages and admonishments about U.S. copyright law. Maybe I’m reading something into the terms of use page for Civil War Washington site, but it seems more half-hearted. You can tell that here is a project that isn’t a holding place for fixed perfected primary resources like Whitman’s, but an evolving scholarly discussion that could easily involve others.

Why not then allow for the download of all the data on the site? I don’t think it would detract from Civil War Washington; indeed, it would probably increase the profile of the site. The site would not only have its own interpretations, but allow for other interpretations—off of the site. Why not let others have access to the guts that Raskin wished to cloak? This is the way networked scholarship works. And this is, I believe, what Roger Bagnall was getting at yesterday when he said “we need to think about the death of the [centralized website] project” as the greater success of digital humanities.

Jim Chandler and I have been formulating a rule of thumb for these editorial projects: the more a discipline is secure in its existence, its modes of interpretation, and its methods of creating scholarship, the more likely it is to produce stripped-down, exchangeable data sets. Thus scholars in papyrology just want to get at the raw sources; they would be annoyed by a Mac-like interface or silo.  They have achieved what David Weinberger, in summarizing the optimal form of the web, called “small pieces, loosely joined.”

On the other hand, the newer and less confident disciplines, such as the digital geographic history of Civil War Washington, Hypercities, and Grub Street feel that they need to have a Raskin-like environment—it’s part of the process of justifying their existence. They feel pressure to be judge, jury and executioner. If the Cohen-Chandler law holds true, we will see in the future fewer fancy interfaces and more direct, portable access to humanities materials.

Of course, as I note in my paper, the level of curation apparent in a digital project is related to the question of credit. The Whitman archive feels like a traditional editorial project and thus worthy of credit. If Ken instead produced KML files and raw newspaper scans, he would likely get less credit than a robust, comprehensive site like Civil War Washington.

The irony about the long-suffering debate about credit is that every day humanities scholars deal with complexity, parsing complicated texts, finding meaning in the opaque. And yet somehow when it comes to self-assessment, we are remarkably simple-minded. If we can understand Whitman’s Leaves of Grass, surely we can tease out questions of credit and the intellectual work that goes into, say, complex KML files.

To help spur this transition along, Christine Madsen has made this weekend the important point that the separation of interface and data makes sustainability models easier to imagine (and suggests a new role for libraries). If art is long and life is short, data is longish and user interfaces are fleeting. Just look at how many digital humanities projects that rely on Flash are about to become useless on millions of iPads.

Finally, on sustainability, I made a comparison in my paper between the well-funded Whitman archive and the Civil War Washington site, which was produced through sweat equity. I believe that Ken has a trump card with the latter. Being a labor of love is worth thinking about, because it’s often the way that great scholarship happens. Scholars in the humanities are afflicted with an obsession that makes them wake up in the morning and research and write about topics that drive them and constantly occupy their thoughts. Scholars naturally want to spend their time doing things like Civil War Washington. Being a labor of love is often the best sustainability model.

Rob Townsend, the Assistant Director of Research and Publications at the American Historical Association and the author of many insightful (and often indispensible) reports about the state of higher education, writes with some telling new data from the latest National Study of Postsecondary Faculty (conducted by the U.S. Department of Education roughly every five years since 1987). Rob focused on several questions about the use of technology in colleges and universities. The results are somewhat surprising and thought-provoking.

Here are two relatively new questions, exactly as they are written on the survey form (including the boldface in the first question; more on that later), which you can download from the Department of Education website. “[FILL INSTNAME]” is obviously replaced in the actual questionnaire by the faculty member’s institution.

Q39. During the 2003 Fall Term at [FILL INSTNAME], did you have one or more web sites for any of your teaching, advising, or other instructional duties? (Web sites used for instructional duties might include the syllabus, readings, assignments, and practice exams for classes; might enable communication with students via listservs or online forums; and might provide real-time computer-based instruction.)

Q41: During the 2003 Fall Term at [FILL INSTNAME], how many hours per week did you spend
communicating by e-mail (electronic mail) with your students?

Using the Department of Education’s web service to create bar graphs from their large data set, Rob generated these two charts:

Rob points out that historians are on the low end of e-mail usage in the academy, though it seems not too far off from other disciplines in the humanities and social sciences. A more statistically significant number to get (and probably impossible using this data set) would be the time spent on e-mail per student, since the number of students varies widely among the disciplines. [Update: Within hours of this post Rob had crunched the numbers and came up with an average of 2 minutes per student for history instructors (average of 83 students divided by 2.8 hours spent writing e-mail per week).]

For me, the surprising chart is the first one, on the adoption of the web in teaching, advising, or other instructional duties. Only about a 5-10% rise in the use of the web from 1998 to 2003 for most disciplines, and a decline for English and Literature? This, during a period of enormous, exponential growth in the web, a period that also saw many institutions of higher education mandate that faculty put their syllabi on the Internet (often paying for expensive course management software to do so)?

I have two theories about this chart, with the possibility that both theories are having an effect on the numbers. First, I wonder if that boldfaced “you” in Q39 made a number of professors answer “no” if technically they had someone else (e.g., a teaching assistant or department staffer) put their syllabus or other course materials online. I did some further research after hearing from Rob and noticed that buried in the 1998 survey questionnaire was a slightly different wording, with no boldface: “During the 1998 Fall Term, did you have websites for any of the classes you taught?” Maybe those wordsmiths in English and Literature were parsing the language of the 2003 question a little too closely (or maybe they were just reading it correctly, unlike faculty members from the other disciplines).

My second theory is a little more troubling for cyber-enthusiasts who believe that the Internet will take over the academy in the next decade, fully changing the face of research and instruction. Take a look at this chart from the Pew Internet and American Life Project:

Note how after an initial surge in Internet adoption in the late 1990s the rate of growth has slowed considerably. A minority, small but significant, will probably never adopt the Internet as an important, daily medium of interaction and information. If we believe the Department of Education numbers, within this minority is apparently a sizable segment of professors. According to additional data extracted by Rob Townsend, it looks like this segment is about 16% of history professors and about 21% of English and Literature professors. (These are faculty members who in the fall of 2003 did not use e-mail or the web at all in their instruction.) Remarkably, among all disciplines about a quarter (24.2%) of the faculty fall into this no-tech group. Seems to me it’s going to be a long, long time before that number is reduced to zero.

Roy Rosenzweig and I had a terrific time on The Kojo Nnamdi Show today. If you missed the radio broadcast you can listen to it online on the WAMU website. There were a number of interesting calls from the audience, and we promised several callers that we would answer a couple of questions off the air; here they are.

Barbara from Potomac, MD asks, “I’m wondering whether new products that claim to help compress and organize data (I think one is called “C-Gate” [Kathy, an alert reader of his blog, has pointed out that Barbara probably means the giant disk drive company Seagate]) help out [to solve the problem of storing digital data for the long run]? The ads claim that you can store all sorts of data—from PowerPoint presentations and music to digital files—in a two-ounce standalone disk or other device.”

As we say in the book, we’re skeptical of using rare and/or proprietary formats to store digital materials for the long run. Despite the claims of many companies about new and novel storage devices, it’s unclear whether these specialized devices will be accessible in ten or a hundred years. We recommend sticking with common, popular formats and devices (at this point, probably standard hard drives and CD- or DVD-ROMs) if you want to have the best odds of preserving your materials for the long run. The National Institute of Standards and Technology (NIST) provides a good summary of how to store optical media such as CDs and DVDs for long periods of time.

Several callers asked where they could go if they have materials on old media, such as reel-to-reel or 8-track tapes, that they want to convert to a digital format.

You can easily find online some of the companies we mentioned that will (for a fee) transfer your own media files onto new devices. Google for the media you have (e.g., “8-track tape”) along with the words “conversion services” or “transfer services.” I probably overestimated the cost for these services; most conversions will cost less than $100 per tape. However, the older the media the more expensive it will be. I’ll continue to look into places in the Washington area that might provide these services for free, such as libraries and archives.

From the shameless plug dept.: Roy Rosenzweig and I will be discussing our book Digital History: A Guide to Gathering, Preserving, and Presenting the Past on the Web this Tuesday, January 10, on The Kojo Nnamdi Show. The show is produced at Washington’s NPR station, WAMU. We’re on live from noon to 1 PM EST, and you’ll be able to ask us questions by phone (1-800-433-8850), via email (kojo@wamu.org), or through the web. The show will be replayed from 8-9 PM EST on Tuesday night, and syndicated via iTunes and other outlets as part of NPR’s terrific podcast series (look for The Kojo Nnamdi Show/Tech Tuesday). You’ll also be able to get the audio stream directly from the show’s website. I’ll probably answer some additional questions from the audience in this space.

In prior posts in this series (1, 2, 3, and 4), I described with some glee my rash abandonment of common blogging software in favor of writing my own. For my purposes there seemed to be some key disadvantages to these popular packages, including an overemphasis on the calendar (I just saw the definition of a blog at the South by Southwest Interactive Festival—”a page with dated entries”—which, to paraphrase Woody Allen, is like calling War and Peace “a book about Russia”), a sameness to their designs, and comments that are rarely helpful and often filled with spam. But one of the greatest advantages of recent blog software packages is that they generally write standards-compliant code. More specifically, blog software like WordPress automatically produces XHTML. Some of you might be asking, what is XHTML, and who cares? And why would I want to spend a great deal of effort ensuring that this blog complied strictly with this language?

The large digital library contingent that reads this blog could probably enumerate many reasons why XHTML compliance is important, but I had two reasons in mind when I started this blog. (Actually, I had a third, more secretive reason that I’ll mention first: Roy Rosenzweig and I argue in our book Digital History that XHTML will likely be critical for digital humanists to adhere to in the future—don’t want to be accused of being a hypocrite.) For those for whom web acronyms are Greek, XHTML is a sibling of XML, a more rigorously structured and flexible language than the HTML that underlies most of the web. XHTML is better prepared than HTML to be platform-independent; because it separates formatting from content, XHTML (like XML) can be reconfigured easily for very different environments (using, e.g., different style sheets). HTML, with formatting and content inextricably combined, for the most part assumes that you are using a computer screen and a web browser. Theoretically XHTML can be dynamically and instantaneously recast to work on many different devices (including a personal computer). This flexibility is becoming an increasingly important feature as people view websites on a variety of platforms (not just a normal computer screen, e.g., but cell phones or audio browsers for the blind). Indeed, according to the server logs for this blog, 1.6% of visitors are using a smart phone, PDA, or other means to read this blog, a number that will surely grow. In short, XHTML seems better prepared than regular HTML to withstand the technological changes of the coming years, and theoretically should be more easily preserved than older methods of displaying information on the web. For these and other reasons a 2001 report the Smithsonian commissioned recommended the institution move to XHTML from HTML.

Of course, with standards compliance comes extra work. (And extra cost. Just ask webmasters at government agencies trying to make their websites comply with Section 508, the mandatory accessibility rules for federal information resources.) Aside from a brief flirtation with the what-you-see-is-what-you-get, write-the-HTML-for-you program Dreamweaver in the late 1990s, I’ve been composing web pages using a text editor (the superb BBEdit) for over ten years, so my hands are used to typing certain codes in HTML, in the same way you get used to a QWERTY keyboard. XHTML is not that dissimilar from HTML, but it still has enough differences to make life difficult for those used to HTML. You have to remember to close every tag; some attributes related to formating are in strange new locations. One small example of the minor infractions I frequently trip up on writing XHTML: the oft-used break tag to add a line to a web page must “close itself” by adding a slash before the end bracket (not <br>, but <br />). But I figured doing this blog would give me a good incentive to start writing everything in strict XHTML.

Yeah, right. I clearly haven’t been paying enough attention to detail. The page you’re reading likely still has dozens of little coding errors that make it fail strict compliance with the World Wide Web Consortium’s XHTML standard. (If you would like a humbling experience that brings to mind receiving a pop quiz back from your third-grade teacher with lots of red ink on it, try the W3C’s XHTML Validator.) I haven’t had enough time to go back and correct all of those little missing slashes and quotation marks. WordPress users out there can now begin their snickering; their blog software does such mundane things for them, and many proudly (and annoyingly) display little “XHTML 1.0 compliant” badges on their sites. Go ahead, rub it in.

After I realized that it would take serious effort to bring my code up to code, so to speak, I sat back and did the only thing I could do: rationalize. I didn’t really need strict XHTML compliance because through some design slight-of-hand I had already been able to make this blog load well on a wide range of devices. I learned from other blog software that if you put the navigation on the right rather than the more common left you see on most websites, the body of each post shows up first on a PDA or smart phone. It also means that blind visitors don’t have to suffer through a long list of your other posts before getting to the article they want to read.

As far as XHTML is concerned, I’ll be brushing up on that this summer. Unless I move this blog to WordPress by then.

Part 6: One Year Later

In an article published tomorrow, but online now, the journal Nature reveals the results of a (relatively small) study it conducted to compare the accuracy of Wikipedia with Encyclopaedia Britannica—at least in the natural sciences. The results may strike some as surprising.

As Jim Giles summarizes in the special report: “Among 42 entries tested, the difference in accuracy was not particularly great: the average science entry in Wikipedia contained around four inaccuracies; Britannica, about three…Only eight serious errors, such as misinterpretations of important concepts, were detected in the pairs of articles reviewed, four from each encyclopaedia. But reviewers also found many factual errors, omissions or misleading statements: 162 and 123 in Wikipedia and Britannica, respectively.”

These results, obtained by sending experts such as the Princeton historian of science Michael Gordin matching entries from the democratic/anarchical online source and the highbrow, edited reference work and having them go over the articles with a fine-toothed comb, should feed into the current debate over the quality of online information. My colleague Roy Rosenzweig has written a much more in-depth (and illuminating) comparison of Wikipedia with print sources in history, due out next year in the Journal of American History, which should spark an important debate in the humanities. I suspect that the Wikipedia articles in history are somewhat different than those in the sciences—it seems from Nature’s survey that there may be more professional scientists contributing to Wikipedia than professional historians—but couple of the basic conclusions are the same: the prose on Wikipedia is not so terrific but most of its facts are indeed correct, to a far greater extent than Wikipedia’s critics would like to admit.

For those who have been asking about the article I wrote with Roy Rosenzweig on the reliability of historical information on the web (summarized in a previous post), it has just appeared on the First Monday website, perhaps a little belatedly given the name of the journal.

Given the current obsession with the reliability (or more often in media coverage, the unreliability) of information on the web—the New York Times weighed in on the matter yesterday, and USA Today carried a scathing op-ed last week—I feel lucky that an article Roy Rosenzweig and I wrote entitled “Web of Lies? Historical Information on the Internet” happens to appear today in First Monday. If you’re interested in the subject, it’s probably best to read the full article, but I’ll provide a quick summary of our argument here.

Using my H-Bot software tool, Roy and I scanned the Internet to assess the quality of online information about history. In short, we found that while critics are correct that there are many error-riddled web pages, on the whole the web presents a relatively sound portrayal of historical facts through a process of consensus. With the right tools, these facts can be extracted from the web, leaving the more problematic web pages aside.

Moreover, this process of historical data mining on the web should prompt further discussion about the significance of all of this historical information online. To do some of our own prompting, we had a special multiple-choice test-taking version of H-Bot take the National Assessment of Educational Progress U.S. History exam using nothing but the web and some fancy algorithms borrowed from computer science. [Spoiler alert: it passed.] This raises new questions that move far beyond simple debates over the reliability of information on the web and into the very nature of teaching, learning, and research in our digital age.