Peter Batchelor, Sunrise Research Laboratory[HREF1], Royal Melbourne Institute of Technology[HREF2], GPO Box 2476V, Melbourne, Victoria, 3001, Australia. Phone +61 3 9660 3024 Fax: +61 3 9660 2761 peterbat@rmit.edu.au
In late 1995, Liddy Nevile visited 40 heads of departments in RMIT and showed them, in the privacy of their offices, often with doors closed, that there was a significant new technology which might make a difference to their planning and adoption of the then 'mighty interactive multimedia' and flexible teaching pressures within education. Already in Sunrise there were many activities taking place via Internet including the sharing of information, working remotely with students, on-line chatting, etc...
Gary Hardy went on to develop Victoria's outstanding community/government website VicNet and those in Sunrise kept working on what had always been of interest in new technologies, the opportunities for better, more convivial, computational learning environments for individuals and collaborating groups. RMIT, although a partner and leader in this exercise, did not see the need for a coordinated effort to make sense of the web.
A few experiments with such systems as Lotus was offering in late 1995 confirmed that collaborative computational spaces were going to make big differences to the opportunities for shared working activities. About the same time, the Australian government tried to promote the use and production of CDs with some schemes for making them significant export earners for Australian developers. The 'web' was making its presence felt but it was not a useful institutional system.
The need to make sense of the possibilities and to closely link them to the aspirations and needs of RMIT for its clients led to a considerable amount of rethinking of existing computing principles, experience, and practices. This led immediately (perhaps to be regretted later) to the need to separate the infrastructure underpinning the RMIT website into the practical and technical aspects. A committee of review was formed and a year of discussions led to some 'style' guidelines for the immediate future of at least the top-levels of the website (if they could be defined).
This was meant to be a policy about the presentation of material on behalf of RMIT. It seemed there was an idea that the website could be seen as 'publishing', a well established activity, by the university as a whole, the faculties, departments and other senior sections. Below the departmental level, more freedom was advocated because it was known that the RMIT community would not like to have too much control imposed. It was hoped that best-practice demonstrations at the 'higher' level would lead to a general adoption of desirable standards. Strategies which involved deep-planning, such as those which would have investigated the information needs of the university and related them to the structure of the new information system, were rejected as unnecessary. The newly-appointed head of information technology services, a position described as including information management as well as technical delivery, assured those assembled at a public meeting that it was too soon to be developing serious applications on the web. He predicted that proprietary solutions such as those anticipated by Lotus Notes, would at some stage become viable albeit expensive solutions to RMIT's needs on the web. Specifically, programs such as those being developed at Cornell University, which were of significant interest to many in universities if only for their good intentions and leadership in the use of the web as a significant part of the information system of a university, were rejected. Even participation in consortia investigating these issues at no cost was not valued.
By mid-1996, serious users of the web could see that this was developing into an environment which had the potential to significantly impact all information practices within institutions such as universities. A proposal for a major re-think of RMIT's strategy for the web which included an analysis of information management, cataloging and retrieval practices, computer -based activities and needs of the community, including the prospective student market and alumni, the development of a graphical representation of RMIT for a screen environment, and the processes necessary to encompass a wide sector of the RMIT community in improving the web, became a temporary position for a single member of staff.
In mid-1996, Peter Batchelor was employed by the Corporate Affairs Group to work with RMIT staff to effect the changes necessary to make RMIT's website a 'world leader'. His position was confirmed late in the year for 6 months and made equivalent to 0.6 EFT.
What has been achieved, what could be done, and some learnings from the experiences of the last 12 months are considered below.
In early -1996, the RMIT website looked like this(taken from OZeKIDS Good'e's Disc which archived it in early 1996):
It was the work of Richard Muirden who put it together in his spare time and it was constructed similarly to many at the time. Richard's aim was to link information and give access to website making facilities to those who were beginning to want them. His achievement (undertaken more or less in his spare time) was significant.
The physical organisation of the RMIT web-presence was at this stage very difficult to manage, as it still is. Any number of computers on the various RMIT campuses were connected to Internet. Local system managers were responsible for how they were used and when. It was not possible to know who had websites and for what purposes. It was by no means extraordinary to find the Victorian Divers Association homepages in amongst the new mathematics on line courses.
The pressure, however, was to have the 'best' websites, and this was soon accompanied by the desire to have the most functional one, as users, not just authors, were beginning to take websites seriously.
The challenges included:
to make sure that all who should be accessible from RMIT's front page could be found from that page, a not insignificant goal given that nobody had any way of know how many websites there were within RMIT computers at the time;
to make sure that the website was accessible to those who had older hardware or software facilities only, including all RMIT students and staff who did not have private service provider connections and so were dialling in to RMIT and accessing the web off-campus using only text browsers such as Lynx;
to develop an aesthetically satisfying 'look and feel' for RMIT's electronic front door, and perhaps hardest of all,
to aim for consistency at least in the upper areas of the website by getting all involved to make their pages in conformity with the style guide.
The erstwhile head of Computing Services had held an open meeting to explain, to all interested, the constraints and expectations for the website from his perspective. This meeting, which was in a sense very exploratory, led to the development of a boo group who did not attend meetings of the committee which worked on the guidelines but they could see what was being discussed, how changes were likely to be made, and to offer their suggestions.
Typically, within the formal committee and the boo group, there were vast differences of opinion about the look and feel aspects of the guidelines. Less interesting, it seemed, were issues such as what would make the website most appealing to the various anticipated target audiences, guidelines for ensuring that publishing of university material did not contravene what would have been recognised easily as unacceptable limits if the publishing were paper-based, intellectual property rights, and so on. Generally, those working on the guidelines were themselves discovering the qualities and potential of the new technology as they worked. the focus was narrow and the web was so new there were no real experts or authorities to whom they could go for help within their focus. Generally the wider issues were considered to be 'out of scope'.
The atmosphere in which the players worked was noticeably one of collaborative exploration. Fears that students might, in the meantime, do 'terrible' things or that something might blow-up were easily allayed by the experiences of the times - for those involved they seemed much less threatening than had been the activities of hackers in a previous decade of new technological development. Everyone was learning and there was a sense of excitement and satisfaction.
Peter contacted as many people as he could within RMIT and advised them of the assistance he was willing to offer them in their webmaking. Staff from a wide range of areas within the university, educational, administrative and resource management as well as marketing, and international programs, took advantage of the offer and they worked hard to make their websites better. Not all webmakers were committed to the style guide which was constantly being revised as the possibilities changed. Overall, there was a happy state of exploration achieved with the constant buzz of malcontent from users of the web who wished it did this, that or the other - never the same things - and those who wished it looked like this, that and the other - again with no consensus.
As well as the process of pedestrian but steady development of the website, there were distinct processes being developed such as the initiative of Paul Kennedy who recognised the opportunity to make sense of course information across the university. Paul advocated the use of the web to provide RMIT with a single storage warehouse for information so that at least errors would not occur from inaccurate multiplicity but more so that information could be sought by a wide range of users for a range of purposes. Andrea Chester and Gillian Gwynne experimented with new possibilities such as on-line anonymous chat environments to help students think about their personal identity, in a context curriculum subject. Teachers and library staff in the TAFE sector developed on-line courses for administrators who chose to undertake study from their work-desk in preference to attending classes on the RMIT Campus. Courses were developed for students in Asia who were working across national boundaries on international trading business practices. A remote-delivery dance course was established for north eastern Victorian students. The use of CDs for the delivery of material collated and presented Internet-style was discovered.
Individuals and small groups of people around RMIT were doing as they had always done; practising new skills to achieve outstanding results and incidentally raising the standard of both the technological understanding and deployment of the web. But the demands of the webmaster were becoming overwhelming.
The mailto: facility on the RMIT website was making it very easy to communicate with the now identified and approachable webmaster, and many did. Each 'caller' only needed a little help, but offering enough little bits of help can overwhelm anyone. A second 'shadow' webmaster started working. He too was overwhelmed within a short time. As Richard Muirden had found out early on, there are plenty of people who suggest that saying no is a good way to solve the problem, others who say that better organisation is the trick. There are others who wonder why the work is not done. There are few who realise that it can take 10 hours for even a competent person to make a simple graphic work on all combinations of browsers and computers. We believe the job is too big for 1 EFT webmaster!
Among the changes are:
the immediate rationalisation of the images;
the internationalisation of the content;
the support of stylised pages from the top down;
training and support for departments, and
policy development and exchange across the university.
A reasonable set of demands today includes a wish that the RMIT website could offer:
a good marketing vehicle for prospective students and industrial partners
a good graphical (screen?) representation of RMIT which conveyed a desired message
a powerful, integrated Internet with intranet facilities, especially with seamless links to restricted information for those who have appropriate privileges.
Outstanding challenges already identified and listed as priorities include:
making the website suitable for those who are web disabled
linking the website to RMIT's information system so that recording of information does not continue to be duplicated
facilities for direct input of new information will result in instant publishing of the corrected information
the generation of indices which will provide effective and efficient access to web material however it is distributed across the range of RMIT computers
the relation of 'looking good' to 'feeling good' as two facets of user-interface design which together have been long recognised as relevant to user satisfaction (graphics and structure which together with good content create a compelling environment for users)
development of appropriate literacy skills by all in the community expected to use the website for information handling and communication, and
the development of a secure and functional commercial environment upon which RMIT staff and students can depend to augment their many endeavours (where relevant).
Today the website looks different by the week:
A sense of freshness is being offered even if there is not the capacity for instant mass publication. Without the benefit of information about the information handling needs of the community, and how the web can support them, there is no facility for automatic updating of the web. There is no obvious repository for informatio which staff might want to make available for others. In the absence of this, there can be no automatic generation of a web-presence for this material. Someone has to put it on the web, link it into appropriate pages, .... and all this is generally impractical and will only lead to more mess.
RMIT's website today is like a library in a huge abandoned building. There are DLOs (document-like objects but think for a moment of them as books), piled high all over the place. imagine a pile in what used to be the main entrance of the imaginary building. Someone has neated the piles, put interesting ones on top. The piles of books look quite interesting and well ordered. So do the piles in all the other rooms. But what if you want to find a street directory, a book about caterpillers. How would you know if there was one? How would you guess where it might be? the person who tidied the pile of books in one room might remember what was in that room, but what would they know about what is in the other rooms?
structure | metadata | privacy, security etc | web literacy | access for the web disabled
If only we, and many of our colleagues, had been able to learn all about websites before we started building them, what might we have done?
The above question is one many ask and it does seem that there are some principles which can be used to save time and regrets. There is also, however, a cautionary note to be added. We wish to draw attention to the need for many of us to work through processes in order to come to understand how to use them, what they can achieve. The advice then, is build a website for fun, before trying to implement the suggestions which follow.
But if we could start again, we would work first to ensure that the structure of the website was closely related to the purposes for which it is to be used. Such purposes range inevitably from those related to formal publishing, through informal publishing and communicating, to the gathering of information and on to rigorous data collection. Not unreasonably, the website is seen as potentially a vehicle for instantaneous publishing. (It has been demonstrated as perfect for this - but how much of it can be managed with the current system?)
it is useful to think of the web as an interactive, distributed database. This model has the website as a collection to which many can add material, or containing material that many can alter or remove, where the material is stored on a vast array of computers and where there is no central control over what becomes part of or leaves the collection. Instead of a database manager, such as a piece of software which records the presence of all material, there is a giant society of 'document-like objects' which grows and shrinks perpetually.
The notion of distributed database differs from the conventional model in that there is no need for a controlling list of the content of the database if what is contained in it can be accessed appropriately on-demand, one way or another. The 'glue' which makes the disparate objects into a single, identifiable website, is usefully thought of as the use which is made of messaging within the material as much as it is that the material happens to be on computers which are proximal located, or even physically linked by cables.
Standing back from the material on webpages, one can see the pages as places where information is stored, and the metadata attached to the pages as the names of the containers. In the early phases of web use, the units of material have been identifiable by their URLs, with a slight modification available where NAMES have been used within pages to locate particular sections of pages. What has not yet evolved for common use is containers of sets of materials, of pages, but they will be in common use within a short time.
So far, there are sections and pages. What is coming is containers and small units of material which come together to form pages, on the fly. One example of this is where a page is constructed using such tags as INCLUDE and another is what will happen when style sheets are employed. With style sheets, it will not be necessary to supply all the formatting information with each page but possible to provide a page and a reference set of information about the style in which that material is to be displayed. If this can be done for all pages within a container, for example, it will need to be downloaded only once for all the DLOS in that container.
Metadata is the term applied to data which is seen as being about other data. A library catalogue record is metadata about a book, for instance. A 'handle' is a particular set of information which uniquely identifies a DLO, eg a number associated with it according to when it is created, or published. DLOs which are metadata and thus related to other DLOs increasingly are proving useful for the development of efficient searching and indexing facilities. Handles can help as they can be used to identify chunks of information even where that information changes or is replaced. Handles related to persistent URLs (PURLs) make a big difference.
PURLs allow users to identify a DLO by name, say, and then update the actual document which has that name. For instance, 'today's news' will be a different DLO each day if the news is archived whereas it will probably be a webpage that has to be editied each day if not. Having a PURL for 'today's news' makes it very easy to just add a new DLO of news and have it automatically replace the existing one, eg by a webpage which contains a script that tells it to include the file with the latest date from the directory (or folder) called 'daily new items. An additional advantage of using a PURL like this is that the DLO contasining the new items might also be emailed to someone, printed on paper, translated into a foreign language, etc.
So granularity of 'chunks' of web material is one issue, persistent naming of the chunks another, alongside the possibility to archive DLOs, and there are many others, given the likelihood that any information on the web is likely, in one sense, to change.
What emerges in this context then, is the need to develop a data accumulation system which will not necessitate the re-organisation of DLOs when old ones are replaced by new ones, or when changes are made to existing ones. How, for example, can one know when replacing one reference on a web page by another, how many pages out there somewhere have pointed to this page specifically to help locate what it pointed to?
The only solution is a scalable data accumulation system which logically incorporates new material and seamlessly removes discarded material. This is no longer a task which can be given to someone who is 'in charge' of the material but rather has to be designed into the distributed database in such a way as to make it possible for all users/developers to work with it. It is non-trivial to organise this but fortunately, the technology which causes the problem can be used to solve it. Java scripts, wizards and robots can be used to generate new links.
Another area of concern is to do with certification, authentication of DLOs, privacy and access to them. The protection of intellectual property rights, and the freedom of speech and the elimination of undesirable material from the recipient's browser balanced with the publisher's right to freedom of speech, and so on are more issues to be considered.
If the university website is to provide a one-stop information system, the changing of a student's address should be notified by the student (as they say at Cornell, who better knows the correct version) and it should be reflected in all information which refers to that student's address, immediately. Fine, but what if the information is confidential information, or if it is not as definite as an address but rather ranks only as an assertion which will be valued according to who made it, for instance. Issues such as who should have access to this information, is this information being given with some expectation that privacy will be maintained, how accurate does this information need to be, etc, are not unique to universities but they are of major concern to them. How can courses be offered in flexible mode if there is no way to ensure that the 'student' sitting the on-line test is indeed the student, and not their expert friend. How can courses be offered if it is easy for a charlatan to copy them and offer them at half the price, even with a fake graduation certificate, from his own pirate site?
Privacy, digital signature, user profiles, trust management, intellectual property rights and rewards, and other issues in this area, are currently the subject of intensive development for the web (see eg http://www.w3.org[HREF3]).
The two technical if-onlys, in the short term, relate to how directories (or folders) can be set up for the storage of DLOs so they can be used repeatedly in different ways without having to be relocated and what information could be associated with them to make it easier to determine the privacy, certification, etc issues. In these cases, the answer is likely to come from the development of good metadata systems. (In this paper there is no in-depth discussion of metadata issues as it is the intention of the authors merely to locate the need for consideration of these.)
The technical if-onlys lead quickly to considerations of the implications of working on these differences. Websites in universities will hopefully always be open and thus involve hundreds if not thousands of authors. The massive task of making available and encouraging the education of these authors is almost overwhelming.
Our experience is leading us to urge those involved to start thinking of web making as basic literacy and communication activities.
Where reading and writing are esoteric activities, they are privileged and soon enough marginalised within crowded educational curricula, as has been extensively demonstrated in the recent past. When it is essential for all to use the technology AND there is a supporting culture for doing this, developing skill becomes a community activity and the burden of teaching and learning it can be distributed across those who need to learn. In our experience, this does not happen when there is a competitive environment surrounding the use of the technology in question. We draw upon the principles we developed for the introduction of what we called 'personal computing' a decade ago (see http://www.srl.rmit.edu.au/papers[HREF4]).
We advocate the use of web-style interactions with respect to information between students and teaching staff, for instance. Why use PowerPoint when a web browser can be the interface to rich material? Thinking of presentation packages as providing the labelled lid on the information offers a powerful metaphor here. If the lid is 'locked', as with reductionist type presentation packages it usually is, students have to re write all the material themselves. If the presentation is the course view of fine-grained material, students can integrate it into their growing personal knowledge spaces and immediately cross-reference it with what they already have and know.
Within university communities, teaching staff and students are not the only people who need to use the new forms of literacy and communication. Opportunities for administrative staff (and at RMIT security staff, outdoor environmental development staff) and others to work with the new technologies for private and public purposes is important. Just as with microcomputing in the past, helping everyone to understand the use of the technology for personal, private purposes is important. It is fairly common knowledge that women in particular like to know that when they are using computers they are doing something which will make life easier in the long run. The need for the personal valuing of the activity can be fostered and in our opinion should be. Encouraging everyone to work with a new technology demands, initially, a proactive approach and this, in turn, usually depends upon those who know enough, thinking of ways in which they can promote the value of the activity until its actual value takes over. It is the experience of burdensome overheads which so often wards users, sensibly, off new technologies.
In our case, we have used the CD technology alongside the new web technologies as one strategy to lessen the overheads and increase the advantages associated with becoming web-literate (see http://www.srl.rmit.edu.au[HREF4]). In addition, we have engaged people in the developments which we have been working on for them, offering them ownership and experience in the process.
A third aspect of website provision that has recently come to our attention relates to the web disabled. For many reasons, beyond such obvious ones as hardware and software limitations and cabling, a large proportion of the community are being disadvantaged by current website developments Many sight impaired people have found employment in recent years in occupations which allow them to use computers. As more need develops for those using computers to have access to the web, these workers are losing their jobs again because they cannot get access to information on websites which are not designed for access with screen readers. Image maps and GUI interfaces render many pages 'blank' for the sight-impaired. And who, we need to ask, is to be affected this way? It is estimated that all who live more than about 60 years will be web-disabled unless web designs change.
So we are starting to think about our users from a wider range of perspectives than previously. Originally we worried about the interests of school students compared with already enrolled students, and about the difference between the users with Netscape 2 and those with Netscape 3. Then we realised that many only have text access. In paying attention to the needs of the text-only users, we have, in fact as a side-effect, made our website more accessible to those who need screen readers. We have also made it easier to translate into other languages. We have not, by any means, reached sufficient levels of achievement in these areas, but we have recognised the needs and the ways to work on them.
We conclude, from our experience, that successfully handling a university website is just like handling any other new technology: it all depends upon working with people. It is also true, we believe, that it depends upon deep understanding of the technology working in collaborative dialogue with deep understanding of the needs and aspirations of users.
Exemplifying achievements and processes is only a start: best practice examples do not deliver in themselves. Actively participating in the development of a suitable 'popular' culture is essential: this needs to be done with care and skill based upon experience as there is nothing more damaging to the cause than enthusiasm and drive for a technology which later proves terminally flawed. The 'backwash' phenomenon is not new to those who have worked through the last few decades of technological development.
Always having in mind scalability is critical. What is advocated for the institution should make sense at the faculty level, at the departmental level and even in the individual's use of the same technology: inconsistency encourages the blowout of overheads for the learners and users.
In the same way, timelessness is a priority. Working with systems which have the capacity to become legacies and drain resources and opportunities can be costly. Many proprietary solutions are closely tied to hardware and software bases which do not scale, or work across platform changes. One of the great attractions of web technology is the plug and play aspect of it. It is possible to predict the need for new software and hardware components and avoid the temptation to choose solutions which offer short term gains over long term maintainability. The temptation is often not seen as such, because the implications of choosing a particular practice are not fully anticipated.
We inherited a system which had not become locked into a legacy model, fortunately, but we now work hard to make sure that when we act, we 'do the right thing' and that very often means do what will have long-term, scalable benefits over immediate simplicity. We do feel we have dug deep enough to achieve our ambitious goals for the website, however.
* The use of the word 'handling' draws attention to the need, within organisations, for reliable and persistent means of identifying 'document-like objects', digital materials in this case.
When racial intolerance as an Australian cultural value was portrayed broadly by Pauline Hanson, it was deemed helpful to immediately reassure overseas students resident in Australia, and their parents and friends back home, that this was not a common value, and especially not tolerated within the university communities. Information to this effect was instantly made available on the RMIT website and also access information for those who were threatened or wanted help was posted conspicuously on the website.
Among the changes are:
This included finding ways of making the best quality, smallest versions of commonly used images available so that when a browser was seeking yet another RMIT logo, it could simply re-use one already in cache. A central collection of commonly used images has been made available to those who need these images, and now they are pointed to be websites from all over RMIT, even those not necessarily located on the main RMIT server.
RMIT students and staff do not have PPP connections to the RMIT system from outside the campus and so there was a lot of pressure from those wanting to access the web from outside to make sure it is at least meaningful, if not pretty, from there.
More recently these issues of speed and efficiency are being noted as relevant also to the web disabled.
Internationalisation of the content
The RMIT grew as the product of those within RMIT who were using the web. It did not reflect the interests of thsoe who might come from the outside, let alone those who were not yet university peple at all, such as prospective students, their parents, and friends. Several incidents such as the question of friendliness of Australian universities towards students from malaysia help catalyse the action to re-think the audience of the website, shifting it away from 'vanity' publishing on a large scale towards functional publishing.
Support of stylised pages from the top down
The process of working towards an acceptable style guide and then encouraging staff and students within RMIT to voluntarily adopt the recommended style has been supported by the provision of tutilege for those who are trying to do this. The webmaster and assistant aim to help users develop the necessary understanding so they can be independent rather than make the pages for them. This support has been free to departments.
Training and support for departments
In addition, training courses have been developed all over the university in response to the pressure on teaching staff to earn extra funds for extra facilities. These courses are not yet to be co-ordinated, moderated, and their competitive nature and lack of certification leaves them variable in content, quality and cost.
Policy development and exchange across the university
The development of web-using policies and increased professional development of staff is also patchy. Departments are generally left to choose to participate or not and they have done so with varying levels of success. One problem with this approach is that many of the successes are not known about, and those who makes a lot of noise about their achievements can easily become the 'masters' who set the standards. Innovations often take some time to be seen for what they are.