AusWeb97-Culture-The link list on network policy
The link list on network policy
Tony
Barry,
Visiting Fellow, Department of Computer Science, Faculty of Engineering
and Information Technology,
Australian National University.
tony@ningaui.anu.edu.au
Abstract
Five years ago the link@www.anu.edu.au email list was formed as a
discussion forum on Australia network policy. The quality of the
debate on this list has been high with participation from senior
levels of universities, the media, industry and government. Much
of the discussion has been aided by participants publishing, via
the web, significant contributions to the debate consolidating
what has gone before and providing links to the main sources on
the topics covered.
A number of the authors have maintained web papers on a dynamic
basis seeking regular input from the list community on
improvement and new information sources.
Discussions on this list have made a significant contribution to
government policy making and provided a rich source of ideas for
the media and source of contacts.
Without use of the web to provide a publishing mechanism for more
thoughtful contributions, this list would have not have been so
successful if based only on email.
The
hypermail archives of the list are now indexed via the harvest
system and provide a unique resource on the development of the
internet in Australia. Analysis of the content provides some
interesting information on similarities between publishing behaviors
on lists and in journals.
Introduction
The link list was established in 1993 as a forum to discuss emerging
issues related to the development of the internet in Australia. The
impetus for this came from a meeting of the Working Group on Local
Systems Interconnection, now the Electronic Libraries Forum (ELF [HREF 1]
). As a
consequence the membership initially came from the library, networking
and standards community.
The list was established by Eric Wainwright, the Deputy
Director-General of the National Library of Australia and myself.
Initially the list was run manually on a unix server. Early in 1995 it
shifted to a majordomo [HREF
2]
server maintained by the Centre for
Networked Information and Publishing [HREF 3]
at the Australian
National University. Unfortunately prior to that time very little of
the discussion had been archived and what there was , was lost during
this transition. The archives of the list are maintained under
majordomo and can be retrieved via email using the normal majordomo
commands. A
hypertext
version of the archive [HREF 4]
is also maintained using hypermail
[HREF
5]
.
This archive is indexed using the
harvest system [HREF 6]
. I maintain a
web site for the list on a Macintosh in my office at ANU, external to
the archive site which is located at CNIP.
Membership
Because of it's history referred to above the list was initially mainly
concerned with academic issues related to AARNet and its performance. the
two government inquiries into the network, the Australia Science Council
Inquiry into
High Performance data networking and the Broad Band Service expert
Group.
With the rapid growth of the network and the divestiture by the AVCC of
the
backbone
network to Telstra the members ship of the group progressively widened to
include
members from -
- The rapidly growing ISP industry
- People in Government grappling with policy and practical problems
which the internet
was creating for the delivery of government service.
- Journalists attracted by a source of expertise and information.
- Interested individuals attracted to a new medium which had the
likelihood of creating
profound social change.
- Political advisers and staffers from both sides of politics
- Many members of CAUL [HREF
7]
and CAUDIT [HREF 8]
and office holders from organisations such as ALIA [HREF 9]
and ACS [HREF 10]
.
Membership distribution
The distribution of the current membership seems to be fairly equally
divided between the higher education sector, government and commercial
interests. as set out in the following table -
| Domain | Number |
| com.au | 102 |
| csiro.au | 5 |
| edu.au | 126 |
| gov.au | 101 |
| net.au | 39 |
| org.au | 16 |
| oz.au | 13 |
| telecom.au | 6 |
| Overseas | 29 |
Total | 437 |
These figures may be somewhat misleading as some members from
government may access the service via commercial suppliers and
the those from overseas domains (mostly .com) may in fact be in
Australia. There are more
detailed figures
by
subdomain [HREF 11]
available. These figures were derived by
sending the "who link" command to the list and therefore suppress
those who did not want to make there membership public. A simple
perl script analysed the data.
Content of the Discussion
From the start many of the faults which email lists often generate -
- Flame wars
- Swamping by "newbies"
- Swamping by those seeking information rather than those wishing to
provide it
were avoided. It is not clear why this was the case. One possibility
is that the list was formed with an initial membership who largely
knew each other and shared preexisting expertise and a wish to
cooperate and share knowledge. this initial group of only a few dozen
set a style of seriousness which continued throughout the continuing
list of the list such that the signal to noise ration has always been
high and the content largely of a substantive nature.
While the membership is not dominated by residents of Canberra a
disproportanite do come from there. as a consequence that has been
some physical meetings of list members and others interested in the
internet in Canberra now semi-formalised as
Internet Reality
Checks [HREF 12]
(an obvious pun on IRC - Internet Relay Chat), which I
started but which have been actively supported for some years now by
Tom Worthington who describes them as
... designed so that members of the net community can meet and
exchange the small amount of very important information which is not
suitable for digital transmission.
Even those who did not attend shared a more
social experience than is usual with just a list server as the
meetings are reported and photos of the meeting published on the
web.
As a social group the list as a group, at the
suggesti
o
n [HREF 13]
of
Jack Gilding [HREF 14]
, raised
the funds to send
Robin Whittle [HREF 15]
on a
trip to Sydney [HREF
16]
to provide briefing information to the
Australian Broadcasting
Authority [HREF 17]
which contributed to their relatively well informed
position in their approach to the control of content on the
internet.
The rarity of transgressions of netethiquette, the low volume of
beside-the-point postings and threads and the large volume of off-list
communications stimulated by postings to link, are all indicative of a
the formation of a
C
y
berCulture [HREF 18]
as describes by Roger Clarke one of the lists most active members..
Distribution of postings
The archives of the mailing list are available from majordomo using
the "get" command to retrieve individual files. Each file cover one
months postings and their names can be obtained via the "index'
command. By retrieving these files and analyzing them with a
perl
script [HREF 19]
it is possible to look at the pattern of
contributions. Ranking the contributors by the number of postings and
looking at the proportion of contributions coming from the heaviest
posters provides information on the activity of the list. my initial
assumptions were that the list might well -
- Follow some kind of 80-20 rule of thumb
- Be dominated by a few heavy posters in a sea of lurkers
- Follow a Bradford type distribution.
To my surprise the latter proved to be true.
The Bradford distribution is an example of a frequency-rank
distribution [BROOKES78] in the field of bibliometrics and provides a
distribution function that predicts in a bibliography of journal
citations what proportion of the citations come from a given
proportion of the most productive journals. There is an extensive
literature which was surveyed in a special issue of Library
Trends [TRENDS81]. There are a variety of similar
distributions in other areas of the social sciences notably in
linguistics and economics.
While the figures covered a four month period (January to April 1997)
an covered 1552 postings. The membership during that time was slowly
changing so there will inevitably be a few per cent fuzziness in the
figures. Nevertheless it is clear that only about one third of the
membership made no postings during that period indication a fairly
high participation rate (279 individuals placed posting out of about
440 members). At the high end, seven members of the list (1.4%)
generated 25% of the postings (so its hardly 80-20) and half the
postings were made by only 4% of the members. The distribution of
usage shown in figure 1.
FIGURE 1
This shows a farly typical Bradford distribution. A reasonably linear
section in the middle with the "Groos" droop [GROOS67] at the start
and a bit of a tail at the low frequency end.
A comparison between the domains of the posters and the membership is
also interesting and is set out in the table below-
| Domain | % Members | % Postings |
| com.au | 23.3 | 30 |
| csiro.au | 1.1 | 1.2 |
| edu.au | 28.4 | 24.1 |
| gov.au | 23.1 | 4.6 |
| net.au | 8.9 | 10.8 |
| org.au | 3.7 | 10.4 |
| oz.au | 3 | 0.3 |
| telecom.au | 1.4 | 0.1 |
| Overseas | 6.6 | 17.6 |
The not unexpected result is that government members, because of the
position, make fewer contributions than other members. This may be
masked however as some members in Government employ use addresses at
commercial ISPs.
The overseas figures are dominated by those in the COM domain
but it is likely that they are actually Australia residents.
Role of the web
From the start the use of electronic publishing has been central to the
discussions.
This manifest itself in a number of ways.
Alerting
Much of the earlier contribution to the list were reproduction of
useful information judged to be of interest to the list derived from
other lists and EJournals. More commonly now is the practice of
posting URLs of items which are useful to current discussion of which
are used to spark new discussions. With the rapid adoption by the
government as a publishing media, particularly for material relating
to the network, such references are often to material published by
Government particular
reports of inquiries
[HREF 20]
, press releases [HREF
21]
and Parliamentary material available from
Parliament direct [HREF 22]
or via the ANU's
Pastime project.
[HREF 23]
In the case of the Federal Government, the list and several members
played a
large part in the rapid adoption of the new media as documented [HREF 24]
by Tom Worthington. He claims that the list accelerated the development
of
federal government web pages by six to twelve months.[WORTHINGTON97]
Drafts of submissions and papers
A number of the more active members have adopted the practice of
circulating drafts of papers they have written seeking comment
from the list before submitting them in final form. Roger Clarke
in particular has used the list as such a mechanism. Some groups
have also used the list as a forum to debate material they were
preparing to submit to inquiries although more commonly they
would post the URL of the submission subsequent to it being
placed. Recently this sparked a discussion on whether such a
practice
might
be in breach of parliamentary privilege [HREF 25]
for material
submitted to Parliamentary Committees.
This practice also caused considerable concern in the Defence
Department which wrote
procedures [HREF 26]
based on
the link experience.
Tom Worthington comments -
The ACS Community Affairs Board used (and still uses) Link to collect
views
on draft ACS policy and for submissions to Government. What appeared to
happen on several occasions was that the draft material was incorporated
directly into Government and opposition policy from Link, even before the
submissions had been formally sent.[WORTHINGTON97]
Some members have taken this one step further notably
Roger Clarke [HREF
27]
who
maintains a number of "live" pages the content of which is in
part derived from list discussions. notable amongst these is the
page he maintains with the Australia Computer Society on
"Regulating
the Net" [HREF 28]
Using the same sample which gave the
membership figures I searched for URLs via a perl script and
counted how often they occurred. Discarding those which occurred
more than three times as these were likely to be home pages from
attached signatures gave an estimate of 808 URLs mentioned in
1552 postings an average of just over one posting for every two
messages which indicates the high reliance on the web as a
mechanism for transferring information rather than just conveying
content in the body of a message.
The list web site
To promote the operation of the list I am developing a web site to assist
the membership
and in particular to promote discussion. The content will include -
- Administrative information eg list commands
- A library of links to topics under discussion
- Background information on the lists such as that contained in
this paper
My approach will be to -
- Use components that are not costly and readily available
- Build a sense of community in the list
- Make the site interactive as far as possible
Nuts and bolts
As a small community of a few hundred there is no need for a
substantial server to maintain the site. In this case it is an
old Macintosh 7100AV following on from the ideas in my paper [HREF 29]
to
the last AusWeb conference.
The server software is the freeware
Quid Pro Quo [HREF 30]
which supports
a number of sophisticated features including server side
includes. Form handling and tasks which require some logic are
handled by MacHTPL [HREF 31]
which is also
freeware and adds simple programming ability to html. Heavy duty
cgi activities can be handled by Applescript and
MacPerl [HREF
32]
.
Both these languages are able to cross execute scripts and
commands. Interfaces to email are provided via
Mondo mail [HREF 33]
which can be
executed from within HTPL documents.
SMTP and POPMAIL are provided via Apple's freeware
AIMS server [HREF 34]
.
and mail list can be supported via freeware
Autoshare
[HREF 35]
which links to AIMS or with
Macjordomo [HREF 36]
.
Remote maintenance is supported using the ftp server capability of Peter
Lewis's shareware
NetPresenz [HREF
37]
.
Administrative components
The server currently supports -
These documents incorporate standard headers and footers via the
"includes" statement in server side includes.
Links are also provide to the archives and the searchable database.
Electronic Library component
A set of pages are being built up on topics which have been discussed
or are under discussion. This will mainly consist of links to items
mention by members of the list as being pertinent and also dynamic
links of searched into the archives database to retrieve postings on
that topic or related to it. Each page has a form in which members can
make suggestion of further items which might be included.
I have a definite philosophy in developing these pages. In a print
library the emphasis is on acquiring documents, often from suggestions
by users of the library, in anticipation of future demand. These
documents are then described via an abstract indexing language,
typically Library of Congress Subject Headings, which has little
relevance to the language used by the users. Queries are then mapped
to this language to retrieve what is pertinent. On the web electronic
documents are already on the "shelves". What a library then collects
should be concepts of interest to their user community to which
documents can be attached. The practice of the print library is
inverted. Rather than assigning topics to documents can be assigned
to topics as suggested by the clientele and the retrieval language can
be the language of the users instead of something abstract developed
prior to need as is the case with subject headings. In the context of
a list this results in pages related to topics under discussion and
the gradual creation of a cross reference (broader term, narrower term
etc) structure as the need develops.
I have used a number of techniques in developing these topical pages
to cut down on the support requirements which might prove useful for
other contemplating something similar. This is set out in Figure 2.
FIGURE 2
Individual URLs are collected in a
WebArranger [HREF 41]
database.
This is an object oriented database which is adapted to maintaining
URL databases. Amongst its many export formats if the Netscape
bookmark format. URLs for each topic are collected in a database
"Note" and then exported to the web site when needed.
The base document for each topic is an htpl based on an shtml
template [HREF 42]
which picks up standard navigation
information, a form for contributions, a link to the bookmark file and
incorporates the bookmark file, with some reformatting via an executed
perl cgi [HREF 43]
, into the document.
The form when invoked submits its data to an
htpl page [HREF 44]
which logs the submitted data and
generates a mail message of the content via Mondo mail.
Possible developments
There are a number of areas in which this site can be enhanced.
Whether this happens will depend on whether the site proves useful to
the list and the direction in which they want it to go. Some
possibilities are -
- Great expansion of the number of topical pages with the
development of a cross reference structure.
- Web access to the nominated items for the topical pages
- A user feedback rating scheme for topical links to provide a
refereeing mechanism
- Add chat and threaded forums possibly via Framechat [HREF 45]
or Interaction
[HREF 46]
- The devolution of some pages
- The use of freeware eg.acgi, Apples v-twin search engine to
index the site
- Development of robot indexing of sites linked from a topic
using Phantom [HREF 47]
Conclusions and musings
The link list has been very successful. While much of the evidence for
this is word of mouth, and perhaps apocryphal, it would seem that the
debates on the list over the years have had some influence on
Government in the increased use of the internet and the web in
particular (if not in policy) , have helped inform the media or some
sections of it and have built a sense of community amongst the
participants.
It is not clear how this has come about. The initial closed membership
selected on their basis of interest is one contributory factor.
Another is the policy orientation of the discussion which attracted,
via word of mouth or email, other participants who shared the values
of the list and an interest in rational discussion of important
network issues. While the character of the discussion has changed over
the years, more recently tending to be more technical, this character
has remained. Hopefully it will continue to do so.
Whether the web site will aid the development of the list or not still
remains an open question but in future this can be found by an
examination of the web site at
http://ningaui.anu.edu.au/link/
[HREF 48]
to see if it continues to be developed.
Acknowledgements
I would like to thank the members of the list for their useful
comments on a draft of this paper, notably Roger Clarke, Tom
Worthington and Robin Whittle.
References
- BROOKES78
- Brookes, Bertram C. and Jose M.Griffiths. "Frequency-Rank
Distributions".
Journal of the American Society for Information Science
. Jan 1978 p.5-13
- GROOS67
- Groos, O.V. "Bradford's law and the Keenan-Atherton Data".
American Documentation
v.18, 1967. p46.
- TRENDS81
- "Bibliometrics". Library Trends
V.30(1), Summer 1981. Special issue devoted to bibliometrics.
- WORTHINGTON97
- Personal communication
Hypertext References
- HREF1
- http://www.nla.gov.au/lis/stndrds/grps/elf/
- HREF2
- http://www.math.psu.edu/barr/majordomo-faq.html
- HREF3
- http://www.anu.edu.au/CNIP/
- HREF4
- http://www.anu.edu.au/mail-archives/link/browse.html
- HREF5
- http://www.eit.com/software/hypermail/hypermail.html
- HREF6
- http://harvest.transarc.com/
- HREF7
- http://www.anu.edu.au/caul/
- HREF8
- http://www.caudit.edu.au
- HREF9
- http://www.alia.org.au
- HREF10
- http://www.acslink.net.au/
- HREF11
- http://ningaui.anu.edu.au/link/membertbl.html
- HREF12
- http://www.acslink.net.au/~tomw/irc/irc.html
- HREF13
- http://www.anu.edu.au/mail-archives/link/link9604/0040.html
- HREF14
- http://www.edna.edu.au/vetwp/
- HREF15
- http://www.firstpr.com.au/
- HREF16
- http://www.firstpr.com.au/contreg/link.htm
- HREF17
- http://www.dca.gov.au/aba/invest.htm
- HREF18
- http://www.anu.edu.au/people/Roger.Clarke/II/EncoCyberCulture.html
- HREF19
- http://ningaui.anu.edu.au/papers/ausweb97/contrib.pl.html
- HREF20
- http://www.nla.gov.au/lis/govnii.html#oz
- HREF21
- http://www.nla.gov.au/oz/gov/media.html
- HREF22
- http://www.aph.gov.au/
- HREF23
- http://www.aph.gov.au/library/trialhom.html
- HREF24
- http://www.acslink.net.au/~tomw/auugpa.htm
- HREF25
- http://www.anu.edu.au/mail-archives/link/link9704/0208.html
- HREF26
- http://www.adfa.oz.au/DOD/AboutDOD.html
- HREF27
- http://www.anu.edu.au/people/Roger.Clarke/
- HREF28
- http://www.anu.edu.au/people/Roger.Clarke/II/Regn.html
- HREF29
- http://www.scu.edu.au/ausweb96/tech/barry2
- HREF30
- http://www.slaphappy.com/
- HREF31
- http://www.ice.com
- HREF32
- ftp://ftp.netcom.com/pub/ha/hal/MacPerl/faq.html
- HREF33
- http://www.acmetech.com/
- HREF34
- http://cybertech.apple.com/AIMS.html
- HREF35
- http://www2.kb.bib.dk/Staff/meh/AutoShare/AutoShare.html
- HREF36
- http://leuca.med.cornell.edu/Macjordomo
- HREF37
- ftp://redback.cs.uwa.edu.au/Others/PeterLewis/
- HREF38
- http://ningaui.anu.edu.au/link/faq.shtml
- HREF39
- http://ningaui.anu.edu.au/link/members.shtml
- HREF40
- http://ningaui.anu.edu.au/link/welcome.txt
- HREF41
- http://www.cesoft.com/webarranger/
- HREF42
- http://www.scu.edu.au/aw/papers/barry/template.txt
- HREF43
- http://www.scu.edu.au/aw/papers/barry/bookmarkscgi.txt
- HREF44
- http://www.scu.edu.au/aw/papers/barry/processform.txt
- HREF45
- http://www.pelicore.com
- HREF46
- http://www.ifi.uio.no/%7Eterjen/interaction/index.html
- HREF47
- http://www.maxum.com/
- HREF48
- http://ningaui.anu.edu.au/link/
Copyright
Tony Barry, 1997. The author assigns to Southern Cross University and
other
educational and non-profit institutions a non-exclusive licence to use
this
document for personal use and in courses of instruction provided that the
article is used in full and this copyright statement is reproduced. The
author also grants a non-exclusive licence to Southern Cross University to
publish this document in full on the World Wide Web and on CD-ROM and in
printed form with the conference papers and for the document to be
published on mirrors on the World Wide Web.
[All Papers and Posters]
AusWeb97
Third Australian World Wide Web Conference, 5-9 July 1997,
Southern Cross University, PO Box 157, Lismore NSW 2480, Australia
Email: Ausweb97@scu.edu.au