Wisdom in wiki production/Introducing a wiki for use/Wiki selection

Why choose a wiki? We can study the selection of a wiki from several points of view. As the section Introducing a wiki for use states, the purpose of the wiki is one of the most essential criteria. The technical aspects require attention also; on the whole, there are many issues to consider in wiki selection, and anyone facing the need to choose for the first time might need help. This section of the Wisdom in wiki production material forms a part of a wiki study [http://tampub.uta.fi/tulos.php?tiedot=400 ''Wikiä käyttämään! Tukea wikin valintaan ja käyttöönottoon''] conducted under the AVO project in 2010 (only in Finnish). This section will function as the background for the study, shedding more light onto issues related to wiki selection. Therefore, we wished to include this in Wisdom in wiki production.

About the wiki study
Readers can turn to the study when they want to introduce a wiki for use for the first time and have to consider their long-term wiki use; how do they select the right one? The study reports practical comparisons of wikis in actual use, but testing without possibilities for actual use would remain somewhat limited. Therefore, we use surveys and interviews to collect experiences from parties that are already using the wikis that we might select for the study: their weaknesses, strengths, possibilities. The basic feature lists tell us something about these wikis, but it is only experiences from actual use that give us more in-depth, fine-tuned information. Our principal point of view is that of education. The focus of the study is on the experiences of use of the wikis, both those obtained through testing and those obtainable otherwise. We conduct that part of the study through interviews in organisations that use wikis so that we can obtain material about their experiences of use. The study is more qualitative than traditional application feature reviews. Because there already are rather comprehensive pages available for comparing features of applications (e.g. WikiMatrix and Comparison of wiki software), we do not think it necessary to construct a similar comparison. As applications develop and change, the output would require constant maintenance.

However, this study does not exclude the comparison of the features of various wikis &mdash; this study focuses on features that are seen as critical for the fluency of the use of wikis, features that should be carefully researched when the introduction of a wiki is considered. We present the differences we find when we test and compare the wikis we select for installation in our test set-ups. We report on the differences mainly in terms of our observations and the tested features in real use.

In short:


 * This study does not include feature lists for different wiki platforms like e.g. WikiMatrix &mdash; such lists undergo constant change and users can search for precise, current information in WikiMatrix and on the web pages of the wiki application providers themselves.
 * This study offers suggestions for features, grouped appropriately, considered essential to include in wiki comparisons, to help readers understand which features they should focus on when selecting their wiki platforms.
 * This study does not focus on any individual wiki features but on the objectives and preconditions for working with wikis.
 * We study the important features in our test set-up in more detail than given in WikiMatrix, describing the features rather extensively and generally in view of e.g. the ease of use and coverage of revision control/page history. Our output does not consist only of a feature list with "available/not available".
 * Neither do we rank wikis in any order of superiority, because so many factors influence the issue of which is better. The final outcome reports the purposes for which the different wikis are suitable, how easy they are to introduce for use, how easy they are to use, and which resources would be required. In relation to these issues, we provides verbal descriptions of the testers' impressions of the order of superiority of the wikis in terms of their ease of use in particular.
 * The study relies on the assessments of the selected participants concerning their actual use of their wikis. No test set-up necessarily gives a reliable picture of the actual use of a certain wiki even though certain basic features and their performance can be studied better through test set-ups. The limitation of an assessment like this is, naturally, its case-specificity. Case-specific descriptions still enable us to gain deeper qualitative information and allow us to see the various factors affecting the use of wikis as well as the practices employed in the organisations using certain wikis.

Selection criteria for the wikis in our study
This study includes x number of wiki platforms/programs, roughly to be subdivided in three groups:


 * 1) Free-of-charge Wiki services (e.g. Wikispaces, Jottit, Wetpaint) that get users started easily.
 * 2) Free and open source wiki software, the installation of which becomes appropriate if an organisation decides to start using a wiki.
 * 3) Wikipedia and other wikis of Wikimedia Commons (e.g. Wikibooks, Wikiversity) can be examined as special cases in this study, as the wikis under Wikimedia Commons are implemented on the MediaWiki platform, but they can be applied to many different (educational) purposes, for which they have already been tailored.

The following rough selection criteria were first suggested:


 * open source (for the test set-up at least)
 * free of charge
 * compatible with different operating systems
 * experiences of use available, and/or promising or interesting in character but no experiences available yet
 * page history
 * WYSIWYG edit option (to some degree a least)
 * Finnish language
 * Ranked among top 25 (for example) in WikiMatrix.

We asked for comments to these criteria, explaining them to be rough ideas only (e.g. Finnish language and WYSIWYG were not necessarily seen as very important criteria). A more extensive discussion of the selection criteria and the wikis to be selected is available in the blog Avoimesti tutkien (Openly studying). Discussion has also been ongoing on the pages of (fi) Sometu.

The comments expressed that freedom of cost cannot form a selection criterion, but appropriate performance can. Finnish language was not seen as an essential criterion. (This is partly due to the fact that many wikis that were introduced as non-Finnish may have different language versions available inexpensively, or such can be made in-house.) On the other hand, in the spirit of the AVO project, we considered openness, understood from different points of view, to be an important criterion for the wikis that would be included in the study (e.g. open content, open technical implementation). Questions relating to the ownership of the content (technically and otherwise) were seen as important.


 * Who owns the information?
 * What will be done with the information?
 * How and under which conditions can the information be taken elsewhere (licence)?
 * How do things work technically?
 * Can users back-up copy the data themselves?
 * If the service provider becomes bankrupt, can the data be easily set up in another wiki (exit plan)?

We also decided it prudent to test platforms that have active developer groups and are popular, in other words, platforms that seem viable. Similarly, wikis that are used a great deal are likely (but naturally not certain) to be good measured by some criteria &mdash; precisely because they are used so much.

Because we want practical experiences of wikis in this study, one of the selection criteria (the last one of them) has to be whether we can find organisations that use the wikis that we wish to select for the study.

A crucial criterion, which also forms an essential question relating to the implementation of the study, is the definition of wiki. Which platforms are classified as wikis in our study? This is an interesting question as such, and during the planning of the study, its practical ramifications were evident in the discussions about whether or not other community platforms should be included that either contain wikis or are implemented in a manner similar to wikis. This viewpoint and its consequences to the study are discussed in the following section.

Definition of "wiki" and what the definition means for the selection of the wikis for this study
For the study, it is important to determine what a wiki is &mdash; how extended or closed we want our definition of it to be. The line between wikis and community platforms is not very fixed at all points. If we take a look into a presentation of the history of wikis by Ward Cunningham, the developer of the first wiki (1995), we see certain features that are typical to wikis.

Ward Cunningham, the creator of wikis, defines wikis as follows: the wiki is the simplest on-line database that works. Cunningham highlighted the aspects in wikis that relate to their speed and their ease of use. According to Woods & Thoeny (2007), the wiki can be defined as "a collection of web pages that anyone can edit", which in the view of Woods & Thoeny, leads to the following questions: What is there on the pages? Who all belong to the community of "anyone"? What happens when pages are edited? Are there rules relating to joint ownership of content?

Cunningham defined wikis through certain basic concepts and allowed wikis to develop on their own through practice. A great number of practices were created and all wiki applications work a little differently. There is no generally accepted rule for which applications may be called wikis and which ones must not be called so. The greatest part of the value of wikis is created through the various cultures around them that come about through the actions of people, not so much from the mechanisms in the wiki programs.

We can think of wikis through the metaphor of a box of indexing cards. We can:


 * add new cards
 * write more information on the cards
 * interlink cards
 * sort and search cards
 * copy cards
 * monitor changes made on the cards.

The information on a card would be the body text with the formatting (e.g. bold, underline, italics). The contents can also be arranged as lists and tables. One of the most important features is the interlinking of cards. When you click on a link in a card, you are transferred to the page under the link even if it had not been created before then &mdash; it is created when the link is used, and your may add some contents and save it. The possibility of working with unfinished trains of thought in wikis gives you the easy option of adding information and returning to the context later when you wish.

It is also essential that the indexing cards are kept in a central storage and anyone can use and modify them. Joint ownership of contents is essential in wikis; it is possible to share everyone's special expertise and reach the best possible output in terms of correctness and coverage of the contents (principle of self-healing).

Any simple database is able to perform the simple operations given in our indexing card metaphor. The reason why wikis are so popular is that creating and editing contents is so easy &mdash; unlike other databases, typically. "An easy-to-edit, shared online database of pages that really works? That's a wiki."

Wikis are different from many other internet tools such as email, blogs, discussion forums, content management and publication systems. Wikis are tools that are intended for page creation, and the created pages can function in many different ways &mdash; this is the key to understand the differences between wikis and other web pages and tools. In the following, we present some similarities and differences between wikis and other web tools:


 * E-mail: Similar to emails, wiki pages are easy to create and edit quickly; almost anyone can create a wiki page as well as an email message. Email can be used for communication on one-to-one basis or via group messages and mailing lists. Email systems typically lack a central location accessible to several users at one point in time.
 * Blogs: Wiki pages may look like blog pages but their production logic is different. (Compare to how newest blog posts are shown on top, how comments to posts are with them, and how posts on pages are seen in one view.) Blogs usually work on one-to-many basis, wikis on many-to-many basis.
 * Discussion forums (discussion boards, bulletin boards, internet forums...): In their communication structure (many-to-many), discussion forums are closer to wikis than they are to blogs, but discussion forums are more structured than wikis. We can use threads on wiki pages similar to discussion forums, in which case new comments are shown at the bottom of our wiki pages, but this is a stylistic choice rather than a basic function. On discussion forums, the pages and communication are always constructed in the same way, and they cannot be changed.
 * Content management and publication systems: Wikis, compared to content management and publication systems, share many features in the sense that they are general applications for creating web pages. Content management and publication systems, however, are not based on Ward Cunningham's model which is found behind all wikis. Almost any web site, blog, discussion forum or wiki can be built inside a content management system. Many such systems include wiki extensions. Content management systems cannot be easily used by people other than expert programmers whereas wikis can easily be introduced for use.

The reason why wiki applications are developing into so many different directions is partly that Ward Cunningham did not over-specify wikis to force their development in any way. However, the features of the first wiki he structured have been copied so widely that they now form a de-facto wiki definition. Working from Cunningham's definition, briefly, wikis are required to have the following features:


 * All pages must be stored in a central place so that the wiki can be easily distributed.
 * Anyone should be able to edit the pages. Wikis are flexible, which means that the organisation of the information on a page can be changed, when necessary, by users other than a specific expert or the party maintaining the wiki.
 * Editing and access to editing should be easy, and they should not require special tools. Wikis should be simple so that starting with one is made easy. Wikis should be easy to manage, which will make it easy for anyone to join in and create pages.
 * Editing wiki contents should be simpler than using HTML. Linking pages should take place through e.g. WikiWords or another similar, simple technique.
 * A list of the latest edits should be available.

The standardization of wikis has been suggested but the idea has never taken off the ground. Because the basic idea of wikis is not controlled by anyone, it is not likely that a standard would emerge.

However, wikis have developed in the course of time and the following features have been made available:


 * Revision control: When a wiki page is saved, a new current revision is made, but the previous revision is also kept. The earlier revision can easily be restored.
 * Adding attachments to wiki pages.
 * Backlinks make it easy to browse pages linked to a certain page.
 * Information about changes, e.g. email alerts when a certain page is changed.
 * Search functionality.
 * Print versions of pages, i.e. page versions without navigation bars on the printouts.

The wikis for this study have been chosen with these issues in mind. We have tried to limit the wikis to those that meet our criteria relating to key wiki features; we have regarded platforms with various extra features such as blogs, discussion forums etc. as community platforms. Schwartz, Clark, Cossarin & Rudolph (2004) state that wikis with no extra features are most likely to fulfil the needs of most users, which is why we think it justified to study the key wiki features only. Because many wikis today have a great deal of extra features similar to community platforms, we cannot completely limit our study in this way. Our attention in this study will not be on the possible extra features, but instead, we will focus on the basic wiki features in accordance with the definitions we presented (we might need a classification here to show more clearly what we think the basic wiki features to be &mdash; see also later). We have used a conceptual tool we devised, a wiki continuum, which helps us visualize the characteristics of wikis in comparison with community platforms. Figure 1. Wiki continuum. (Please note. The wikis inside the arrow in the illustration need not be located at the precise spots &mdash; they are only intended to help the visualization of the continuum.)

The reason why we consider classification according to wiki-like characteristics as important is the fact that many extra features may confuse the wiki-like quality of an application and, at worst, make its use complicated. The worst-case scenario is that first-time wiki users would use an application with many extra features and deduce that all wikis are difficult to use and unsuitable for their use &mdash; even if they had been involved in the difficult task of changing the appearance of an application that should actually be classified as a community platform. It is important that those who read our study understand which features (such as page history and revision control features) are important for wikis, in particular, and need to be paid attention to when selecting a wiki platform, and which features are not so important for a basic wiki-like use (such as the options of creating calendars, discussion forums and blogs on the same platform). Of course, users may feel these additional features useful in their own contexts &mdash; but we should talk about community platforms then instead of wikis.

Comparing features of wikis
The basis for the feature selection for our study was the extensive feature list in WikiMatrix. WikiMatrix is a site maintained by the staff of CosmoCode, a commercial German software firm, created for the purpose of comparing wikis. WikiMatrix is based on a highly popular wiki comparison system originally created by Andreas Gohr, a CosmoCode staff member. Because these pages are updated by wiki implementors themselves, it was believed that the information would stay up to date better than if it were updated by a third party. (The requirement in WikiMatrix is that whoever adds a new wiki or maintains information there, belongs to the implementation team of that wiki.) For example, in March 2010, the information concerning only two among the most popular 25 wikis had not been updated in WikiMatrix during 2009 and 2010. Because there already is an established comparison system like WikiMatrix, up-to-date and with good coverage, we did not wish to create one more system by copying the idea. It is worth noting, however, that all wiki applications are not listed in WikiMatrix, even though anyone can enter his or her own wiki application there. Therefore, we asked certain wider social media communities (such as the AVO project and the Sometu-network) to give us some hints about wikis that should perhaps be included in our study.

At first, our idea was to use an expert group to rank the features in WikiMatrix in their order of importance, and then use the ranking to choose the features that were felt to be most essential for comparing and assessing wikis. WikiMatrix also gives short descriptions of what the features consist of. Our closer scrutiny revealed, we were sorry to observe, that not all features in WikiMatrix are provided with descriptions. This is partly explained by the fact that descriptions are peer productions and their standard varies. We decided to let go of the idea of assessing according to the ranking, because our tests showed that we could group the features in more extensive classes. The idea was also brought up that it is not so essential which particular features and detailed descriptions are available for a particular wiki &mdash; but it is essential whether the features can be used for creating the kind of wiki the user wants. In other words, the possibilities enabled by the wiki are important whereas the tools as such are not, because various tools and various combinations of features can be used for reaching the same end. Example: it may be one and the same how the pages in the wiki are managed (search functions, indexing of all pages, page map, tags...) &mdash; the main thing is that their management works in some easy way.

Important features: the wider perspective of features and performance. Below, we will list the features that should be studied on the basis of the above:

Group 1: Technical characteristics impacting the wiki application selection

 * The operating system required for the wiki
 * Wiki programming language
 * Server software
 * Storage format
 * Data security
 * Tailoring possibilities (open vs closed source)

Group 2: How you export data from the wiki (back-up copies, transfer to another platform)

 * The wiki licensing model and its impact on the ownership of the content
 * Ease/difficulty of transferring data (can basic level users do the transfer or does it require technical staff)
 * Data transfer format:,  ,  ,  …

Group 3: Resource considerations in wiki selections

 * Expertise required for maintenance
 * Ease/difficulty of maintenance
 * Wiki costs
 * In-service support (technical support, user support &mdash; please note the developer community also)

Group 4: Coverage of revision control/page history

 * Number of revisions
 * Ease of recovery of an earlier revision
 * Comparisons between revisions
 * Comments to changes

Group 5: Ease of use, comfort, usability

 * Appearance and clarity
 * Tailorability and modifiability with basic user level skills
 * Ease/difficulty of the text editor and editing (incl. adding media elements)
 * Simultaneous use by many users
 * Search functions and ease of finding pages (e.g. indexing)

Group 6: What can be included in addition to text (+ file formats, picture sizes)

 * Attachments
 * Video clips
 * Audio clips
 * Pictures

Group 7: Special uses and special user groups

 * Language requirements
 * Barrierlessness (e.g. key board commands)
 * Formula editor (for e.g. mathematics)
 * Mobile use

Group 8: Possibility and number of additional bells and whistles

 * Possibilities for further programming (open vs closed source)
 * Other tailoring and modification (see section 4)
 * Plug-ins available
 * Calendar
 * Blogs
 * Picture galleries

Schwartz, Clark, Cossarin & Rudolph (2004) produced a list of items to consider when selecting a wiki for use in education. Their main-level criteria include:


 * 1) cost
 * 2) complexity
 * 3) supervision
 * 4) clarity
 * 5) common technical framework
 * 6) features

This classification is more technical than ours.

In usability-related assessments, it would be possible to use various usability heuristics with slight modifications (e.g. Jacob Nielsen's Ten Usability Heuristics). Some of the criteria for high-quality online learning materials, developed in the (fi) Referee project, could also be used; for example some of the criteria in group 4 (Instrumental criteria) could be very useful.