communication tools in floss communities

5 downloads 4613 Views 289KB Size Report
creating products like wikis (Wikimedia), blogs (Wordpress), or social networking ... seemed to focus mainly on the development team of FLOSS communities.
IADIS International Conference on Web Based Communities 2007

COMMUNICATION TOOLS IN FLOSS COMMUNITIES A LOOK AT FLOSS COMMUNITIES “AT LARGE” BEYOND THE DEVELOPMENT TEAM Andreas Meiszner The Open University, UK

ABSTRACT During the past years a high number of tools were developed to enable and facilitate communication and collaboration in online communities, web based content management, knowledge building and learning. These tools provide new opportunities for communication, collaboration and content management (storage, retrieval, presentation and classification/taxonomy). Many of these tools were built by Free/Libre Open Source Software (FLOSS) communities; creating products like wikis (Wikimedia), blogs (Wordpress), or social networking solutions (Elgg). Research that was conducted in this domain seemed to focus mainly on the development team of FLOSS communities (Crowston, 2004); e.g. the KDE case study (Hemetsberger, 2004) or the Freenet case study (Krogh, 2003), and not on the community at large. The development team and enhanced team (core team) are only one part of the community and relatively small in comparison to the entire community, with the largest part being the users (Giuri, 2004). Focusing on the core team only, and not on the community at large, might lead to a very different understanding of how FLOSS communities communicate and the type of tools used for communication, collaboration and interaction and for relationship and trust building. To better understand FLOSS communities as learning environments it would be required to have a look at the communities at large and not only parts of it. An earlier study already initially analysed how learning occurs in FLOSS communities (Lin, 2005). This study focused on the learning activities and knowledge creation through mailing lists. To conduct further research in this domain it would initially be important to know which communication tools are currently used within the FLOSS communities at large. The Freenet case study, which focuses on the core team, for example,. suggests that mailing lists are the main way of communication within the community (Krogh, 2003). Another case study on the Apache community concluded that the main way of communication is based on forum posts, but this focuses only on the Apache support community (Hippel, 2002). These two examples indicate that communication within FLOSS communities and the tools used might differ, depending on the part of the community being reviewed (the core team or the community at large). KEYWORDS Free / Libre Open Source Software (FLOSS), Communities of Practise, Virtual Learning Environments, Informal Learning, Communication Tools, Online Communities

1. METHODOLOGY The objective of this basic research is to obtain an initial understanding of the communication tools that are used within FLOSS communities at large (communities). It does not attempt however to examine how these tools are used. This might be subject of further research activities and as a follow up to earlier research such as that carried out by Crowston and Howison (Crowston, 2004). At this paper statistical data on the degree of usage of the tools is provided only for forums; since only for this tool quantitative data were publicly available. Looking at the number of existing communities (there are already more than 130.000 communities registered at sourceforge.net) it would be impossible to review them all. Therefore only those communities were reviewed that produce software that are developing communication tools to be used within online communities. The opensourcecms.com website provides a comprehensive list of FLOSS with around 130 featured solutions. For each of these solutions the website provides access to an online demo, receiving around 20

111

ISBN: 978-972-8924-31-7 © 2007 IADIS

million hits on all demos together. Thus this list seemed to provide a good overview of the main available FLOSS solutions in this area. Following Pór (Pór, 2004) communities are build on platforms providing key modules for: 1. Knowledge development and sharing 2. Relationship and trust building 3. Community facilitation and management 4. System administration and customisation, typically through a web interface Key modules consist of sets of tools that enable the communities to function. Pór further provides a brief overview on existing tools and their relationship to the above cited four points: 1. Tools within point 1 might consist of forums, wikis, blogs, news, or content and document management systems. 2. Tools within point 2 might consist of profiles, calendars, “who is online” functions, messaging and chat, or tools for eliciting common views and intent (votes, polls, ratings,…). 3. Point 3 addresses the basic requirement that a system (platform) is in place that allows users to register, and the customisation of point 1 and 2 type tools (e.g. forum sections) 4. Point 4 refers to the administration side of the used system. Points 3 and 4 are therefore basic conditions for virtual communities; meanwhile the variety of the tools as mentioned at point 1 and 2 might vary within the different communities. Each of the featured websites was initially scanned for its existence of a community following Pór’s classification and the following criteria were applied: • Existence of a community (point 3; Pór, 2004) • Existence of community tools (point 1 and 2; Pór, 2004) To this a further criterion was added: a minimum degree of activity within the community. This is to eliminate the “inactive” communities (Wenger, 1998). As a result totally new projects or projects without activity during the past month were excluded After applying these criteria and pre-scanning all communities, 80 communities could be identified for being subject to a further detailed review and the collection of available quantitative data. The reviewed communities consisted of the following groups as presented at table 1. Table 1. Reviewed communities as classified by opencms.com

- Portals (CMS)

27

- Blogs

13

- Forums

12

- Image Galleries

6

- Wiki

5

- Lite

5

- e-Commerce

4

- Groupware

3

- e-Learning

3

- Miscellaneous

2

Total

80

For all the communities reviewed, with the exception of 4 communities developing wikis and the Elgg community, forums were the main place of visible communication and also the main source of available quantitative data, like e.g. number of members.

112

IADIS International Conference on Web Based Communities 2007

Additionally it must be stated that communication through mailing lists was not covered by this research. One reason for this is that almost all of these communities are listed at sourceforge.net and therefore a mailing list is made available through sourceforge.net. A second reason is that mailing lists are mainly used by the communities’ developer / core team (Hemetsberger, 2004, Krogh, 2003) and this research was aimed at looking at the communities at large. Furthermore, it needs to be stated that only the “main” website of each FLOSS community was subject of this research. This is important to note since in many cases there is more than one community website for one FLOSS project, to e.g. provide support in languages other than English. Finally, the data collected on, for example, “polls”, or “who sees what” tools might be underrepresented. Polls might have been absent at the moment of review, although they are generally used; or there could be a “who sees what” tool available, but since no other community member was registered at the time of review no information was provided on this.

2. FINDINGS 2.1 General Findings Three communities had more than 100.000 members, with the following 11 more than 10.000. Out of this 14 only 1 was established in 2005, another 2 in 2004, and the remaining 11 were established between 2001 and 2003. “Established” in this case means: the earliest detectible date when communication started. Communities might have existed prior to that date and were just transferred to the website examined at a later date. The main source of information to track back communication was through the forum and thus no quantitative data could be collected for the 4 wikis without a forum and Elgg. For none of the communities communication prior to 2001 could be detected. For 70 communities start dates could be identified as presented at table 2. Table 2. Year in which activity started

Start of activity Year Number 2001 5 2002 13 2003 18 2004 15 2005 13 2006 (Oct.) 6 Though all of the communities develop software that provides a broad range of communication and information tools, the ones being actually used within these communities appeared to be rather narrow as shown at tables 3 and 4. This is also true for the 3 reviewed e-Learning communities. Although the e-Learning communities are aimed at producing state of the art virtual learning environments, providing a broad range of communication and collaboration tools, the forum seems still to be the centre of the communities themselves.

2.2 Point 1 Type Tools Ninety-four percent (94%) of the communities had a forum in place that could be accessed through the community’s website. The ones not using a forum were 4 out of the 5 wikis plus the social software Elgg. Forums seemed to be the main tool for communication. A possible explanation for this might be that communication through forums is, similar to the mailing lists used by the core team, the most appropriate way to convert tacit knowledge into explicit knowledge within the collective problem solving processes.

113

ISBN: 978-972-8924-31-7 © 2007 IADIS

However, since 75% of the communities also provided the option of providing information on external communication tools like VOIP, chat and messenger within the member profile, these might also be important tools for synchronous communication. This question also remained open in earlier studies (Lin, 2005) and would need to be clarified by future studies. Another 94% of the communities made a documentation tool available. The documentation provides information on the software, how it is built, how to use it and FAQs. In many cases members were able to edit and enhance the content of this documentation. Table 3. Point 1 Type Tools

Point 1 Type Tools n Forum 75 Doku / KB 75 News (Flash) 68 Wiki 41 Latest News Various 40 Blog 27 Tags 5 Discussion page 4 Video / Podcasts 4 n = 80

% 93,8% 93,8% 85,0% 51,3% 50,0% 33,8% 6,3% 5,0% 5,0%

In all of the 41% of the communities that used a wiki, the wiki was also used for documentation purposes. There seemed to be no type of communities that were using wikis more often than other, except the wiki communities themselves. Wiki communities were also the only ones using discussion pages that could be found at 4 out of the 5 wiki communities. Blogs were less frequently featured in only 27% of the communities and not all of the blog communities used blogs themselves (7 out of 13). On the information sharing side 68% of the communities offered a project related news section, with another 40% offering additional news such as on the latest forum posts, latest blog entries, or RSS news feeds. Tags were used in only 5 (6%) and video / podcasts in only 4 (5%) communities.

2.3 Point 2 Type Tools Point 2 type tools are in most cases fully or to a high degree integrated into the members profile within the forum and only occasionally outside of the forum. Ninety-five percent of the communities provided member profile options. Besides the 4 wikis without a forum, other types of communities provided generally more (or less) the same numbers of profile options. This might be due to the fact that the forum software used already provided a range of options “on board”. Most of the communities that are not belonging to the “forum development category” (as classified by opensourcecms.com) seemed to use third party forum solutions that were developed by the “forum category” communities. In 82% of the cases the profile provided information on the members’ role, its hierarchy and functions within the community and in 80% of the cases also information on the members latest posts, publications, or bloggings.

114

IADIS International Conference on Web Based Communities 2007

Table 4. Point 2 Type Tools

Point 2 Type Tools Profiles Including: Members Roles / functions / Groups Members publications, posts, etc Members MSN, skype, chat, etc. information Members preferences & interests Geo Map Buddy list FOAF My Tasks Own(wiki)page Other point 2 type tools: Calendar / Events Chat Polls "who is online" "who sees what" Shoutbox n = 80

n 76

% 95,0%

66 63 60 37 6 13 2 1 8

82,5% 78,8% 75,0% 46,3% 7,5% 16,3% 2,5% 1,3% 10,0%

20 19 18 46 3 8

25,0% 23,8% 22,5% 57,5% 3,8% 10,0%

In 75% of the cases, members also had the option to provide information on their VOIP and messenger accounts. A brief scan through each community, though, showed that many members did not seem to provide this information. Forty-six percent (46%) of the communities allowed members to provide information on their interests or preferences, though again it seemed that this possibility was not that frequently used. The “who is online” tool could be found at 58% of the communities and in 3 (4%) communities this tool was combined with a “who sees what” tool allowing one to see what other members are currently doing. Calendars and polls (both mainly within the forum) were used by 20%, respectively 19% of the communities; though their actual use appeared to be generally less intensive and the availability of the tools might be seen rather as an “add on” / “nice to have”. Integrated chat tools could be less frequently detected (19%), though many communities allowed members to provide information on e.g. personal IRC accounts within their profiles. A “Buddy” tool was provided in 13 (16%) of the cases and a FOAF (Friend Of A Friend) tool in only 2 (2,5%) of the cases In 8 cases (10%) members had the option to create their own (wiki) page. 4 of these cases were from wiki communities and the 5th the Elgg communities (representing the 5 cases that did not make a forum available). In another 8 cases shoutboxes were provided. The “My tasks” tool could only be detected in 1 community.

2.4 Excurse: Forums For 67 communities (84%) the posts per day could be calculated ranging from 1 post per day to 1.260. There seemed to be no correlation between the period the forum is online and the post per day. The most active community, Joomla, was only online for 410 days. Also the number of members does not seem to have a direct impact on the number of posts and threads per day. Thirteen (13) communities (see table 5) had on an average more than 100 posts per day, without a clear favourite community type. Within this 13 communities the number of members ranged from app. 4.000 to app. 280.000. Eighteen (18) communities had on an average more than 10 posts per member, with three of them having more than 20.

115

ISBN: 978-972-8924-31-7 © 2007 IADIS

Table 5. Communities with more than 100 posts per day

Type - Portals - Forums - Forums - e-Commerce - Blogs - e-Commerce - Image Galleries - Image Galleries - Portals - Blogs - Portals - Portals - Lite

Name Joomla phpBB SMF osCommerce WordPress Zen Cart Gallery

Threads 94.706 345.184 97.826 196.921 87.838 47.590

Posts 516.519 1.900.946 741.921 665.031 441.630 271.281

52.227

254.548

31.130 41.118 6.590 48.112 7.047 774

173.798 189.755 56.855 202.924 54.726 4.336

Forum Statistics Days Post Member online per day 60.750 410 1.260 278.622 1.908 996 65.825 1.167 636 114.589 1.491 446 1.213 364 32.158 1.202 226 1509

169

1107 1.244 380 1.460 508 42

157 153 150 139 108 103

Post per member 9 7 11 6 8

Post per Thread 5 6 8 3 5 6 5

Coppermine XOOPS Textpattern Drupal MODx GuppY

36.805 40.430 8.425 6.531 3.384

5 5 7 8 1

6 5 9 4 8 6

3. DISCUSSION As this review shows, forums seem to be a centre of communication within the FLOSS communities at large. Beside the forum, which is a tool has existed for sometime, the relatively new wikis could be found in more than half of all communities. Wikis seemed to be especially suitable for documentation purposes; hence in all of the cases where a wiki was present it was used for this purpose. Some of the other 75 (94%) communities that provided documentation through other means than wikis also allowed other members to comment, edit and enhance it. It might be assumed that wikis as a relatively new tool are substituting prior existent documentation tools. Blogs are still not that frequently present with only 1 out of 4 communities featuring them on their community site. Providing information on other members’ contributions like prior posts or blog entries seems to be valuable information, hence 80% of the communities provided this feature. Relatively new community building tools like buddy lists, or friend of a friend functions are still not frequently used within the communities, which also applies for tagging, geo maps, and video / podcasts. Though these communities were building and / or integrating most of these tools into the software that they develop, they still do not seem to make use of it themselves. This basic research as presented at this paper was not aimed at investigating how the reviewed tools are used, and in which relation they might be used. It also was not aimed at investigating the learning process (e.g. mainly offline and alone) and learning outcomes within FLOSS communities at large. This remains to be subject of further research activities that need to be carried out.

REFERENCES Crowston, K., The social structure of Free and Open Source software development, 2004. Available at: http://dissertation.martinaspeli.net/papers/roles-and-leadership/crowston-and-howison-2004-the-socialstructure-of-floss Giuri, P., Skills and Openness of OSS Projects: Implications for Performance., 2004.

116

IADIS International Conference on Web Based Communities 2007

Available at: http://dissertation.martinaspeli.net/papers/roles-and-leadership/giuri-et-al-2004-skills-and-openness-of-ossprojects Gosh, R., FLOSSPOLS Skill Surver Report, 2005. Available at: www.flosspols.org Hemetsberger, A., Sharing and Creating Knowledge in Open-Source Communities: The case of KDE, 2004. Available at: http://dissertation.martinaspeli.net/papers/communities-of-practice/hemetsberger-and-reinhardt-2004-sharing-andcreating-knowledge-in-open-source-communities Hippel, E. v., How Open Source Software Works: "Free" User-to-User Assistance., Research Policy, 32, 923-943., 2002. Available at: http://dissertation.martinaspeli.net/papers/social-model/lakhani-and-von-hippel-2002-how-open-sourceworks Krishnamurthy, S., Cave or Community: An Empirical Examination of 100 Mature Open Source Projects, First Monday, 7 (2), 2002. Available at: http://dissertation.martinaspeli.net/papers/basics-and-structure/krishnamurthy-2002-caveor-community Krogh, G. v., Community, joining, and specialization in open source software innovation: a case study, 2003. Available at: http://opensource.mit.edu/papers/rp-vonkroghspaethlakhani.pdf Lin, Y., Diversity of Knowledge and Dynamics of Knowledge Creation in FLOSS Communities, 2005. Available at: http://www.feweb-vu.nl/dbfilestream.asp?id=2474 Pór, G., Collaboration Tools for Communities of Practice, 2004. Available at: http://www.communityintelligence.com/pdf/Collaboration_Tools.pdf Staring, K., Educational transformation through open source approaches, 2005. Available at: http://www.hia.no/iris28/Docs/IRIS2028-1106.pdf Wenger, E., Communities of Practice and Social Learning Systems, 2000. Available at: http://dissertation.martinaspeli.net/papers/communities-of-practice/wenger-2000-communities-of-practice-and-sociallearning-systems Wenger, E., Communities of Practice - Learning as a Social System, Systems Thinker, 1998. Available at: http://www.co-i-l.com/coil/knowledge-garden/cop/lss.shtml

117