|
|
The growth of multimedia computing, followed by a recent push towards publishing on the World Wide Web, is rapidly changing the publishing industry. Editorial staff, working under pressure in printed and online publications, need to use a growing diversity of representations for planning, creating and reviewing content. We present a study of a number of publishing sites, describing how such representations are critical to ensuring quality in the editorial process. Following this, we discuss design implications for better representational tools.
Field study, collaboration, multimedia, publishing, World Wide Web, representation.
In an increasingly competitive market, the traditional media publishing organizations (i.e. newspapers and magazines which are printed on 'web presses') seek to maintain their stronghold as primary information and news providers. A perceived customer demand for high quality graphics and design means that increasing value is placed on how a publication looks as well as what it says. With the advent of the World Wide Web as an alternative medium, printed media organizations (along with TV and radio broadcasting stations) are exploring ways of providing news and other information resources online. However, they must now also compete with new forms of interactive information services that are rapidly appearing on the Internet, such as zines (or e-zines; online magazines) and search engines.
As a result of these competitive pressures, major changes are happening behind the scenes. In particular, there is an expansion of the editorial production process and the resources it entails. By this we mean diversification in the ways people involved in creating and editing published content collaborate with each other and interact with different media and technologies to construct their product - in this case, evolving from printed towards online content. Whereas print publishing has traditionally involved editors, reporters, photographers and graphic designers working together to create and edit paper-based content (i.e. text and artwork), online publishing requires an additional set of skills to not only create and edit, but also to engineer, test, maintain and update dynamic content (such as animations, hotlinks and live chatshows).
All of these varied activities capitalize upon and demand an ever broadening range of supporting technologies. These include text and HTML editors, site management tools, multimedia authoring tools and databases. Growing diversity in the work required to produce content also demands increasing multidisciplinary collaboration between many people who must continuously acquire new skills and develop new processes to keep ahead of the competition.
An key issue for designing future technologies to support this complex domain, therefore, is how editorial content is developed and supported by multiple representations created by different people for different purposes. In particular, what kinds of representations of content are used and what kinds of planning representations are necessary. With so many different kinds of interdependent work going on, the process of representing and re-representing content and plans is critical to the success of the editorial process.
This paper reports on an extensive field study of the changing practices of the publishing and multimedia industries. We focus on how multiple representations are critical to meeting increasing demands for high quality design and graphics in printed publications and for dynamic multimedia in online publishing. The objective of our work is to inform the design of innovative products at Apple. Rather than simply critique the 'single user-interface' that resides in the box on the desk, or evaluate existing classes of technology, our approach is to consider how computing technology and other artifacts are used together in mediating between people and the work to be done. We also consider details of the social, organizational and physical setting. This is in order to understand how technology and artifacts are used in context to mediate between, and to coordinate the efforts of multiple individuals and groups (cf. [6, 13]).
Whilst there is ample research literature on the design and evaluation of collaborative multimedia technologies in the laboratory (see below), there has been scant research so far on the rapidly changing needs of users in the field. A major aim of this paper, therefore, is to present the multimedia publishing industry as a domain of work where collaborative multimedia technologies could play a valuable role.
We begin by reviewing some notable characteristics of this work domain. We then discuss in more detail various representations that are used in the editorial process, outlining their salient characteristics. Finally, we consider how new technologies could be applied to help overcome some of the problems currently experienced with existing representations. In particular, we consider whether the advantages of the many physical representations used can be compensated for or emulated electronically, given the need for representations of dynamic content in online publications.
Before we present some of our observations we outline the background and scope of the study we conducted.
Literature on multimedia publishing is plentiful (e.g. [1, 5, 29]). There is also ample research on the design of multimedia collaboration technologies (e.g. [2, 10, 11]). However, we have found a lack of HCI and CSCW literature on how work actually gets done and how technology is used by collaborators in the multimedia publishing industry. This contrasts with many insightful accounts of other settings such as office work [25], engineering [22], software design [28], airport ground operations [27] and stock trading [15].
There is some reporting on the use of technology to support activities that play a part in multimedia publishing such as collaborative design [21] and co-authoring [3]. These studies have focused on detailed observations of the minutiae of particular kinds of work. Whilst this work can inform designers about particular technology features needed to support effective collaboration, a wider picture of real work settings is also required. Moreover, we argue that multimedia publishing deserves greater attention from the CHI and CSCW communities as a domain with a unique set of characteristics and particular technology requirements.
Another reason for our interest in the multimedia and publishing industries is that a very large proportion of Apple's installed customer base is within these markets. So, it is important for us to have a good understanding of these customers to anticipate their future needs.
During a six month period we conducted an in-depth field study of a number of print and online publishing companies, located in San Francisco. These included a traditional daily newspaper (The San Francisco Chronicle - 'The Chron'), a high-tech monthly magazine (Wired) and a number of web sites (SFGate - The Chron's own web site, HotWired, Suck and Salon).
We set out to develop a detailed understanding of how content is created from the news event to the published article. In order to do this we interviewed, in situ, 25 people in different departments from 6 sites. These included reporters, photographers, designers, section editors, managers, coordinators, engineers, system support and advertising staff. We also conducted observations of work over 4 months, mainly at The Chron, but also at SFGate and HotWired1.
We collected video and audio material for most of the events and activities that we attended, including editorial meetings and newsbreaking incidents (e.g. a false alarm for a 'cop shooting' and a fire at the San Francisco Opera House). We also tracked photographers and reporters who were 'on the beat' for both online and printed news stories. In addition to studying the various sites, we also examined the publications and spent considerable time interacting with and evaluating a range of the media online websites.
In the following sections we overview the work settings, supporting technology and key characteristics of the multimedia publishing sites we studied as context for our more detailed discussion of how representations are used.
1. We were assisted by a group of Computer Science students from Stanford University who joined us for part of the study.
The main study site, The Chron, is a large regional newspaper with 350 editorial staff. Editorial work in this, as in other newspapers, is largely separate from the 'back shop' printing work and the advertizing section. The editorial staff are located mainly on a single floor in two large, open-plan office spaces with some small meeting rooms and private offices for senior staff. Everyone, apart from the photographers, has a personal desk and these are separated by elbow-height partitions (ideal for leaning on when chatting). The physical layout of the office, therefore, lends itself readily to widespread and immediate communication; just by standing up people can see each other and call out to one another to ask questions or shout instructions.
The same kind of layout, of low partitions in an open plan space, was also common to the other sites. In general, it is important for people to be able to see and talk to one another at all times. Having open-plan offices increases the ease with which people can gain rapid and informal access to each other and allows them to monitor each other's work
Most of the staff at The Chron use a 15 year-old proprietary editorial computer system. Although dated and lacking in multimedia capabilities, the system has proved to be very robust, reliable and effective at supporting the collaborative process of writing 'copy'. There is now, however, a gradual move over to a 'Wintel' platform with PCs emulating terminal functions of the old editorial system but enabling a wider range of capabilities. In contrast, the staff in the art, design and photography departments only use Macs.
The other sites we studied, all use a combination of Macs for multimedia authoring and Sun workstations and servers for network and web-site management. The larger, more established sites (Wired and HotWired) have also implemented various software solutions for tracking content. HotWired, for example, uses Microsoft OfficeTM, though it has now outgrown this product and is seeking a better solution. In fact, in most of the organizations we studied, many of the supporting technologies and software support were being outgrown and in the process of being replaced.
There are a number of notable characteristics of the work of multimedia publishing. Most of these are common to other industries such as software development, stock trading and air traffic control. We summarize these to give a broad picture of the multimedia production workplace. They are:
Fast Pace: Content such as text, art work, and layout design must be generated with rapid turnarounds dictated by weekly, daily or hourly deadlines and meetings. As part of the frenetic collaborative activity, vast amounts of paper and computer files in many formats are constantly passed around people and machines both locally and remotely.
Diversity: Many kinds of work are implicated in the editorial production process. Typical activities include information gathering (e.g. interviewing or taking photographs, researching from library archives or online services); creating and processing content (e.g. writing, processing photographs, designing and engineering); editing content (e.g. reviewing, correcting, proofing); and managing projects (e.g. planning, coordinating, holding meetings).
Continuous Switching Between Representations: During the production process many different physical and electronic representations of the content and the process are required to serve a range of purposes. These include sketches; schematic layouts of single pages (called dummies) or of the entire publication or part of a web site; unformatted text; text with annotations; low-resolution photos; high-fidelity static proofs or (in a web site) dynamic dress rehearsals (putting the web site on an internal server); spreadsheets; schedules; content plans (called budgets) and many more. We found, both at the newspaper and at the more hi-tech online sites, that people continually switch between different kinds of representations of the same content. This usually includes frequently printing out representations for the purpose of rapid annotation, for marking the 'passing-on' of responsibility, or for exploiting large areas of wall space for planning and review purposes.
Complexity: A complex collaborative process is required to bring all of the content together and achieve a satisfactory level of quality. In the process of changing hands, every story, photo, graphic, page layout, piece of code and web page goes through numerous iterations and is worked on by many different people before finally going to print or being placed on a public server. Drafts are repeatedly printed out, photocopied, commented on - both verbally and through written annotations, taken 'back to the drawing board' and used as the basis for updates to online versions.
Each cycle of gathering, production and review takes into account more up-to-date news or recently procured information; corrections based on various editorial criteria (e.g. tone, style, readability, accuracy) and fluctuating space availability (depending on last minute additions or deletions or advertizers buying space). In the words of one of the The Chron's designers, the editorial process evolves around a continuous cycle of 'tweak, tweak, tweak'.
This level of complexity demands continuous electronic and face-to-face communication between collaborators. People constantly shout across the room, send messages and reminders to each other, as well as engaging in discussion and meetings with colleagues.
Mobility: Since there is still such a strong dependency on physical representational media and since multimedia representations are difficult to share online, the various content providers and editors constantly get up from their desk to move around the office (cf. [4, 18]), passing paper drafts and files between each other, reviewing and monitoring other's work, updating new information with each other and so on. In addition to this local mobility, reporters and journalists obviously have to travel over a wide area to gather the news. Editors also leave the office to develop relationships with sources and conduct external meetings.
Routinization: Depending on the type of content and the lead times involved, the editorial process can be very repetitive. Work is metered and constrained by regular weekly, daily or hourly deadlines, meetings and procedural steps that must be followed, although there are frequent slippages marked by even more feverish activity for those involved.
These features of the work of multimedia content generation combine to produce a complex and time-critical collaborative endeavor. One of the most important mechanisms for expediting this work is the use of representations as a means to conceptualize, communicate and coordinate activity. In the rest of this paper we concentrate on representations and show how increasing pressure for higher quality design content and, more recently, dynamic, online content presents an opportunity for the application of new technologies to support this important work domain.
A diversity of representations has always played a major part in the above editorial processes, but now, with high quality graphics, sophisticated page-layout, and new kinds of interactive content (e.g. online animations), new additional representations are becoming increasingly necessary. This is especially so for viewing and reviewing of design content and managing more complex projects. In addition, it means an increase and diversification in the kinds of editorial work that has to be done. This has obvious challenges for technology designers, namely:
There is increasing pressure to produce high quality design and graphics, as well as editorial content. This is particularly pronounced in online publications. A result of this pressure is that there a growing need for new ways to view and interact with representations.
Drawing on examples from our study, we show what kinds of representations are currently used in creating editorial content and planning work. We also show why physical representations are often preferred over online representations (an issue which is not developed in Sumner's similar and otherwise excellent analysis of representations used by software designers [28]). We then discuss implications for designing and applying new technologies to overcome limitations of physical representations. In particular, we emphasize how online representations (which have advantages for some purposes and which are particularly necessary for dynamic content) must either mimic or substitute for some of the key attributes of physical representations.
In traditional newspaper publishing, old printing methods enforced a separation of production and review, which took place in the editorial section, from the layout, which took place in the 'back-shop' where galleys of text (column-width strips of paper) were cut and pasted onto boards alongside photo and graphic prints. Our observations at The Chron revealed that many different kinds of physical representations persist in the editorial process. These include:
Miniatures: Unlike text, of which there is usually only one 'live' copy in existence, there may be many available photos that must be visually scanned to find the right one for a story. Photos both from the wire, via Associated Press's LeafDeskTM, and in the form of developed film, are more efficiently searched through in miniature format.
Schematics: Dummies are page schematics showing measurements and pointers to content (e.g. online file-names). They provide a means for an editor to express precisely how content should be laid out by a designer. The advertizing department also creates miniature dummies showing precisely how much space is in the 'news hole' (space left over after advertizing space is sold). Editors mark-up a paper template to create a dummy. The advertizing department create their dummies online. In both cases, however, the dummies are printed out for communication purposes.
Rough Approximations: Low resolution photos are created and printed out by the photo department and passed around to let editors know what is coming and allow them to mark up size restrictions and captions. Graphics, charts and illustrations are all printed out in various stages of completion for review. Hand drawn sketches are also common as a means of communicating graphical ideas quickly.
Jottings: Reporters use a notepad to make rapid notes of salient quotes and important details such as numbers and contacts when talking to sources. A reporter's notepad is portable, unobtrusive and can be used without breaking eye-contact too often with an interviewee.
High-fidelity Representations: A copy of yesterday's paper is used daily for review purposes. High quality color proofs are created online by the art department and printed out for review and annotation by editors.
Planning Representations: So far we have listed representations of editorial content. However, other planning representations are also used. For example, budgets (daily content plans), schedules and assignment lists (for photographers) are used to plan and coordinate work to be done.
One of our main observations was that physical representations (sometimes created on paper and sometimes printed out from a computer) are essential even when it is possible to do much of the creative work in the production process online. This is true for both traditional and online sites. Further, as demand increases for high quality design and graphics in printed publications and for increasingly dynamic behaviors online, so increasing varieties of representations are used.
A key question for us, since we are interested in opportunities for technology to support the editorial process is: Why are so many different kinds of physical representations being used even at organizations where every employee has a powerful networked multimedia PC? Similarly to [14] and [31], we found that many of the properties of physical representations and the ways they are used are still superior to current online representations. These include visibility, communicativeness, status, permanence and task-adaptedness. We look at each of these in more detail below.
Figure 1. The daily 11:30 review meeting at The Chron
The limitations of most PC displays are such that certain representations like photo negatives and page proofs can not be displayed adequately online. For example, in the photo department, photographers review vast numbers of photos from film laid out on a light table, quickly marking the ones they will electronically scan and process. In the art department, high quality proofs are printed out to gain a sense of what the page will finally look like, which is not possible from looking at a computer screen.
Every day in The Chron, a review meeting is held where the managing editor tacks the pages of yesterday's paper on the wall of the meeting room and everyone sits around and critiques the stories, graphics and layout (see figure 1). Likewise at Wired, next month's magazine pages are posted on a wall in one of the meeting rooms where anyone can come and discuss the content and the all-important design and overall flow from page to page. Sticky notes are also attached to the pages as comments for others to pick up.
The same process of using wall space for reviewing content and structure takes place at HotWired and Suck. Web pages are printed out by their authors or designers and links are represented by pieces of string between the pages. Large representational spaces such as these are ideal for the critical work of collaborative reviewing and negotiation which are at the heart of the editorial process.
The innovative use of paper and string to represent the structure and dynamic aspects of evolving magazines or websites shows the importance of having a means of visualizing what is normally internal to the system and invisible to the reader. Using large amounts of space to see the flow between pages of the magazine or the interconnected links in the web pages rather than only being able to jump between them provides production staff with an, otherwise unattainable, overall picture of their product.
We also observed that, in some cases, information that is represented online has to be re-represented physically because the online version is simply not getting through. This was particularly obvious at HotWired in the process of coordinating various groups collaborating on projects. They had reverted to 'turning the technology inside-out': Every morning, a project coordinator would write up, on a physical whiteboard, the main projects, schedules and deadlines relevant for that day that she had extracted from the online, project management software package. When asked why she laboriously wrote up by hand information that could be readily accessed by everyone on the network, she replied that, owing to the multiplication of projects and people working on them, it had become very difficult to keep track of everything that was going on. Moreover, people had become desensitized to the many email reminders that such software applications provide, so they often forgot their significance immediately after having acknowledged them. Consequently, everyone (including herself) needed to be reminded of what was urgent and what needed dealing with that day. Placing this critical information on a physical whiteboard in a prominent public space, which was clearly distinct from the continuous stream of other online information and messages, therefore, provided a more effective public reminder of what was urgent and needed doing that day. This compares with Rogers' [24] observations on the use of a whiteboard for coordination by a team of CAD engineers where a physical representation was necessitated by a problem with online coordination support.
Figure 2. Dummy representation of the front page
Many physical representations such as annotated print-outs, notes and dummies are used to pass information between collaborators. The front page editor in The Chron, for example, communicates with a page layout designer using a dummy. It is the editor's responsibility to decide how the page will look and what will appear in the various parts. She must quickly communicate this to the designer, so she rapidly marks up the schematic layout on the dummy (figure 2). This contains just about everything the designer needs to complete the page layout (although much is left implicit, relying on her experience with this task and the standard features of a front page). The dummy communicates the filenames for the relevant online content, precise measurements for where and how to lay out the content, and any deviations from standard front page features.
The editor, however, always carries the dummy over to her designer personally. This is because there are always uncertainties to be discussed such as late or missing content, things that are expected to change or be added, and so on. Under pressure of the daily 6:00 pm deadline, with updates and changes continually coming in for the front page, she marches back and forth repeatedly to the designer's desk with changes, additions and deletions.
In addition, physical representations act as tokens for the passing of responsibility from person to person (cf. [14]). When the editor passes the dummy to her designer, the designer knows that she must commence work on the layout. Similarly, at Wired, each project has a physical file associated with it. Whoever holds the file is currently responsible for completing some part of the project.
Tangible representations of work and responsibility (e.g. dummies passed from one person to the next, or files on people's desks or being carried around) are visible, and their significance-in-action recognizable, to others. Their use promotes general awareness of ongoing work and state-of-play and provides critical resources for distributed cognition (cf. [17]) that are not available through online representation tools which are geared towards individual-user task requirements. In this sense, physical representations play a key role in coordinating what is becoming an ever more complicated collaborative enterprize.
Physical representations are good at conveying sketchiness and malleability of ideas (e.g. [16]). Besides, when deadlines loom, people do not have time to prepare polished presentations. It is much quicker for the art editor or anyone else to make a sketch convey the gist of an idea perhaps prior to making proposals or during a meeting for discussion purposes. We saw the use of sketches and the blending of sketches with annotations and print-outs in many different circumstances, especially as a means of communicating to designers or artists what they need to be aiming for.
On the other hand, we were surprised to discover the converse of the sketchiness property; that finality can also be subtly conveyed through the choice of paper over online representations. The art editor told us that her department prefers to present section editors with paper proofs rather than letting them view the online pages in QuarkXPressTM. She and her staff have found that editors are far more likely to suggest changes when looking at the online version.
Partly due to the lack of space constraints, and partly due to the properties of physical artifacts, physical representations are often more versatile than on-screen ones (as long as you don't want to change anything). They can be picked up, marked on, moved about and easily manipulated. They can also be left in public spaces continuously for consultation. For example, we saw, at all the sites we visited, many examples of whiteboards or paper schedules being used for long-term planning information display (from a day's to a few months' activity). Physical representations which are permanently accessible are always easy to consult quickly.
Another advantage of continuously accessible physical representations that we observed is that they are very easy to switch between. It is simple to move between multiple sketches, schematics and proofs on paper, whereas, online, the manipulation of layered windows and switching between different applications is relatively cumbersome.
Figure 3. Daily photo schedule. Photographers' names on the vertical axis and time on the horizontal. White areas represent workshifts with hand written notes marking travel time (hashed areas) and duration of assignment activity
Some representations are designed by editorial staff to solve specific problems. For example, in the photo department, a daily schedule is always available for consultation. Assignments used only to be printed from a database by the coordinator, listed in order of when they began, but it was hard from this representation to answer questions like, "Who's available to take a photo of the Opera House fire now?" or "Who's nearest to San Jose now?"
The city desk (breaking news) photo editor, who has a keen interest in making sure photographers are always available to go out at a moment's notice, introduced a new representation for the assignments to deal with this problem. Figure 3 shows how photographers' names are displayed on the vertical axis and time (over a single day) on the horizontal. Assignments for each photographer are mapped against time so that one can see at-a-glance who's free and where people are at any time. One reporter we spoke to observed that there had been a noticeable improvement in the photo department's ability to respond to an urgent request for a photo since this representation had been created.
Our intention so far has not been to dismiss online representations, but rather to highlight valuable characteristics of physical representations that we can consider designing for or compensating for the loss of. Clearly, there are many shortcomings of physical representations; we observed plenty of these at the sites we studied. For example, physical representations reside in a single place and are laborious to copy. Information on paper or whiteboards has to be hand-copied elsewhere (e.g. from paper to the PC, or from the whiteboard to paper). Where duplicate representations exist, only one version may be updated. A particular shortcoming of physical representations for online content is that they are only able to represent static aspects of content.
Online content, by its very nature, is dynamic and interactive so the 'experience' cannot be viewed and explored using paper tacked to a wall. As such there is a growing need for more fluid ways of representing and interacting with dynamic information. The methods used at HotWired, SFGate and Salon include:
This combination of methods reflects the importance of collaborative reviewing. However, we have seen that standard PCs are not well adapted for representing all of the important aspects of online content and structure.
Our study of multimedia publishing work has underscored the value of diverse kinds of representations when creating content - be it print or online. Physical representations, in particular, have many valuable properties that have important ramifications for designing new technologies. In some cases, it may not be an improvement to force what were physical representations entirely online. On the other hand, we have seen that online content does not always lend itself to being represented effectively on paper. So it appears that there is a need to design devices that emulate (or substitute for) some of the properties of the physical representations that we saw in use (such as the pages and string) whilst still providing flexible means of interactivity. We consider some of these properties in the rest of this section.
Large physical representational spaces are difficult to emulate with current technology at a reasonable price. What tends to happen is that, as display size increases, resolution is lost. Xerox's Liveboard [9] and other large computer displays or projection devices are ideal for sharing and could support dynamic, collaborative reviewing of web content very well. When asked what he thought of the idea of a large display, one of the online site's directors responded:
"What I would love would be a flat panel I could hang on a wall ... For the tacked up paper and string setup we have, a video wall could be really useful, not just for the sake of more expensive equipment, but for working with remote group members, for ease of modification, and for keeping a better record of the evolution of the site. Paper and string also don't scale as well to hundreds of pages."
However, when high resolution is also required (e.g. for viewing hundreds of photos or details of an entire web site) then alternative or additional compensatory design strategies may be necessary. A variety of solutions are already in the pipeline. Staples [26] introduced techniques for overcoming limited display space, such as depth, perspective, light and shadow, transparency and opacity, some of which have been implemented in WebBook and Web Forager [7]. Bifocal, or fish-eye views can also be used as in Distortion-Oriented Displays [12] and The Hyperbolic Browser [19]. Alongside these new display conventions, new techniques are required to, navigate, zoom, or manipulate 3D objects.
It is worth noting here that photography in the multimedia industry is rapidly going digital, in spite of the high cost of cameras (around $15,000 for acceptable on-screen or newspaper quality). With the move towards digital photography, the light table, used for visually scanning film, will soon be a thing of the past. New technologies that enable rapid visual scanning of large numbers of images are needed to provide the same versatility and resolution that it offered.
'Information appliances' are specialized, cut-down computational devices adapted for particular kinds of function at a low price, such as games and web-browsing (e.g. Apple's Pippin) or recipes and nutritional information (e.g. Diba's DibaKitchenTM [8]). This specialization is also apparent in the PDA (personal digital assistant) market where some devices like the US Robotics PilotTM only offer five functions (calculator, schedule, phone list, to-do list and memos) which contrasts strongly with the more general purpose Apple Newton or Microsoft Windows CE operating system-based, handheld PCs [20].
In the light of our study we recommend that specialized augmented computational devices could also be valuable. For example, the Liveboard mentioned earlier is a computer that has an augmented display for sharing in group meetings. A separation of specialized from general-purpose computing, whether by cutting down or by augmenting, means designing more task-adapted devices.
Bearing these developments in mind, we have been reflecting upon the variety of roles of the diverse representations we saw and how they could be better supported. This includes thinking about their context of use, how they are created, shared, communicated, visualized, and so on, amongst the different parties involved in creating content.
Instead of trying to replace paper or other physical artifacts by cramming everything into the networked PC, we need to think about adapting and combining a variety of technologies to support a range of representations. For example, small, portable jotters and more-or-less dedicated, notice-board style, public displays would have different representational properties suitable for different purposes than those currently afforded by existing desktop displays.
To illustrate the need for task-adapted artifacts, we saw that reporters take notes on a paper pad, when interviewing subjects, even though they will have to type them up later. Recordings are not used because it is too laborious to listen to them and then write up notes later. We asked why they do not use laptops and were told that this would seem intrusive because of the clicking keys and having to be perpetually looking down at the keyboard. Furthermore, an interviewee cannot see what is on the screen which is between them and the reporter. When we proposed a pen-based device with recording and indexing capabilities similar to Marquee [30] we received a very much more positive response:
"Oh I would do that, that would be cool. I think it would be only a minor distraction, because the person that you would be using it in front of would want to look at it and see how it works and say 'Wow I want to try this. Can I look at this?' ... just small and hand held would be good ... I mean I wouldn't mind carrying around something like that where you could actually tape the person ... I don't know if photographers would like something like that, where you could take an image and you could have ... no that would be totally cool."
Rather than just thinking in terms of replacing paper-based methods with electronic technologies, we are also considering how to more effectively integrate the two. For example, a task-adapted device application that combines new and traditional technologies, would be to attach small wireless capable devices to paper folders which are carried around and passed from person to person for tracking the ongoing status of a collaborative project. A useful side effect of this is that the person responsible for the current step could, literally, be quickly located or contacted through the device.
The idea of integrating paper-based and electronic technologies seems particularly suited to areas where the practice of moving information on- and offline is particularly prevalent. At The Chron an extreme case of this was in the photography department where assignments (excursions to take photos) and schedules were printed out, annotated, updated back online and reprinted multiple times.
Paper representations are necessary here, so that photographers can quickly consult them to find out what their next job is. On the other hand, if the schedule slips, relying heavily on paper means that the photo coordinator must go to a lot of extra trouble to negotiate with and update photographers in the field or editors elsewhere in the building.
The need to be able to turn computerized online representations into other kinds of offline representations was also described above in relation to the problem of coordinating teams of people working together on the same or different projects at HotWired. In this case the project coordinator resorted to copying important projects, schedules and deadlines relevant for that day from the online project management software package to a large public whiteboard.
Given the preponderance of copying of information between paper, online, on the wall or desk, and so on, it would seem useful to make the persistent practice of moving between online and off-line representations more seamless juxtaposing electronic and paper elements as needed. We are pursuing some of the ideas initially developed for the DigitalDesk [23] that seem particularly applicable here. For example, it would be valuable to be able to remotely view paper-based schedules or whiteboards (perhaps via a video still on a web page) or to enter information on a wireless device in the field to be projected onto a paper schedule or a whiteboard.
The findings reported in this paper have focussed primarily on changes occurring in the work practices of multimedia publishing companies. We have observed an increase in the range of activities and representations needed to create, edit, view and review content - both print and online - and to coordinate the work of putting it all together. We have also pointed out the significance of understanding why traditional representations persist in working practices despite the availability of more advanced technologies. In considering design implications we have been considering solutions that move beyond the desktop into the workplace. These include technologies, such as hand held devices and large displays, that can support the diverse range of collaborative activities and representations used in creating content. Finally, we emphasize a need to understand how technology can also be integrated with existing representations which are often well adapted for their purpose.
We wish to thank our Stanford students, Kathleen McKinney, Josh Elman, Kulin Tantodand and Jim Morris. Also the staff at The Chron, particularly Dave Hyams, Andy Lottmann, Hulda Nelson and Pam Reasner and staff at the other sites we visited, especially Sean Welch, and Eugene Mosier. In addition we thank Paul Dourish, Scott Minneman and Dan Russell for comments on drafts of this paper.
|
|