Content uploaded by Marc Went
All content in this area was uploaded by Marc Went on Mar 13, 2019
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
Digital Technology, Culture and Society
ISSN: 2470-1475 (Print) 2470-1483 (Online) Journal homepage: https://www.tandfonline.com/loi/rint20
Archaeology of the Amsterdam digital city; why
digital data are dynamic and should be treated
Gerard Alberts, Marc Went & Robert Jansma
To cite this article: Gerard Alberts, Marc Went & Robert Jansma (2017) Archaeology of the
Amsterdam digital city; why digital data are dynamic and should be treated accordingly, Internet
Histories, 1:1-2, 146-159, DOI: 10.1080/24701475.2017.1309852
To link to this article: https://doi.org/10.1080/24701475.2017.1309852
© 2017 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Published online: 07 Apr 2017.
Submit your article to this journal
Article views: 727
View Crossmark data
Archaeology of the Amsterdam digital city; why digital data
are dynamic and should be treated accordingly
, Marc Went
and Robert Jansma
Korteweg –de Vries Institute, University of Amsterdam, Amsterdam, The Netherlands;
Graduate School of
Computer Sciences, Vrije Universiteit/Universiteit van Amsterdam, Amsterdam, The Netherlands
Received 5 December 2016
Accepted 19 March 2017
One of the major initiatives in The Netherlands promoting the use of
the Internet by private individuals was De Digitale Stad (DDS), which
is the Amsterdam digital city. DDS was launched in January 1994
and soon evolved from an elementary bulletin-board-like system to
a full blown virtual city with squares, houses, post-ofﬁces, caf
a metro. Archaeology of the digital city makes it clear that there is
no beaten track for preserving and, after two decades, unwrapping
“born digital”material. During the research to reconstruct the digital
city two routes were tried, one emulating the old system, another
replicating it. The outcome, together with the harvest of two
working systems, is a lesson, a concern and an appeal. From the
experience of reconstructing digital heritage, we draw pragmatic
lessons. Tools for digital archaeology are tried and contemplated.
The lessons, however, do not unequivocally support the use of the
notion “archaeology.”The concern is one of the social
responsibilities. Web archaeology, being part of contemporary
history, confronts the researcher with such issues as privacy and the
ethics of “young”data. A case is made for treating digital data
Web archaeology; De
Digitale Stad; emulation;
replica; research ethics;
privacy; digital heritage;
dynamic treatment of digital
Participants in the digital city had an avatar. DDS, De Digitale Stad, that is the Amsterdam
digital city, was much more than an Internet server, if only because the community shap-
ing it had not settled for a deﬁnite meaning. The systems conveyed a sense of community
building, and although there was not one but many communities, there was this one basic
sense of being a virtual “citizen”expressed by the avatar. The DDS-team was self-con-
scious enough to adorn its second anniversary with a full backup of the system “to be
studied by archaeologists in a distant future”: The FREEZE, 1996.
Two decades do not create a great distance. However, we do embark on the archaeol-
ogy of DDS. Following the actors of 1996, the effort to read and resuscitate vintage digital
material may be called “archaeology.”The metaphorical expression does not come with-
out repercussions. First, the lack of distance poses problems of contemporary history,
hardly associated with the notion of archaeology. Second, notions like digital archaeology
CONTACT Gerard Alberts G.Alberts@uva.nl
© 2017 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License
(http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium,
provided the original work is properly cited, and is not altered, transformed, or built upon in any way.
INTERNET HISTORIES, 2017
VOL. 1, NOS. 1–2, 146–159
or web archaeology suggest a kinship to a media archaeology or to an archaeology of
knowledge in the sense of Foucault (1969) or media theory (Ernst, 2013; Presner, Shepard,
& Kawano, 2014, p. 84ff), which is hardly explored here. The present contribution is about
digital, but very material, old tapes. Getting hold of the vintage tapes is one thing, reading
them and making sense of the content quite another. Different tools were tried and devel-
oped. In the perspective of accessing the data for historical research and possibly museum
presentation, major issues arise. Ethical issues, not unusual for contemporary history, hit
the digital archaeologist in the face. More speciﬁc questions related to the technological
aspect of DDS impose themselves, issues of security and integrity of the data. The ques-
tion of privacy, perhaps not strictly related to technology, poses itself with new urgency.
Archaeologies, emulation vs. replica
The effort to reconstruct the digital city and have it in operation almost naturally proceeds
along two routes: emulation and a replica (simulation). The two notions of emulation and
replica have a long evolution of shifting meanings in art history. Here they are taken after
their, equally unstable, meaning in computer science (Smith & Nair, 2005, Chapter 2). The
resulting contradistinction adds a nuance, informed by computer science, to the discus-
sion on emulation in electronic art (Jones, 2004).
Emulation is the effort to run the original code on a new platform. This does not come
without compromise. The DDS-system as preserved does not easily “un-freeze”. It will not
run, primarily because the physical systems supporting it are not readily available. Migra-
tion to present day systems is tedious, but feasible. Getting the legacy software opera-
tional requires adaptations. The authenticity-question, as to what system one is “actually”
running, always remains.
Replica, the look-alike remake by present day means, seemingly discards the authentic-
ity-question –one is obviously not running the real system –and focuses on presence to
the user, rather than on the system.
The difference is that the mimicking is on a different level and from a different perspec-
tive. The original sense of emulation is that of one artist out of admiration mimicking an
other. Here one computer system is thought to mimic the other. The agency is with the
system; it is placed on level with the artist –in the original sense of emulation. The per-
spective is that of the system; the boundary is between hardware and software, or
between layers of software. In a replica, the perspective is that of the user. Irrespective of
what happens under the hood of the system, its surface performance for the senses is
what counts. The boundary is between the user and the system. No sharp distinction is
assumed between user and system, or between hardware and software. The crucial point
here is that considering something either an emulation or a replica is a matter of perspec-
tive and comes with diverging expectations. Criteria vary from reconstructing the opera-
tion of the system to recreating the user experience. Recreation implies making the
experience present, and is in that historiographical sense presentism. Does the archaeolo-
gist in her interpretation identify (with) the code or with the user? It seems that archaeolo-
gists by default choose the ﬁrst; we are preoccupied with the authenticity of the code. We
are, but in fact we show the alternative route as well.
What approach to choose and which tools to develop is contingent upon the goals set
for such a project, which in turn depends on the context. Heritage has a community on
INTERNET HISTORIES 147
the donating end, a group strong and dedicated enough to not throw away the material
remnants, and on the receiving end, a group feeling strongly about the value of these
materials. In fact there are several groups caring for the DDS heritage and their goals vary.
For the joy of a one-time replay a system may be ﬁred up and run with loose ends. For
public access, by contrast, for example a museum exhibit, integrity and security of the sys-
tem and privacy of personal data are key issues. The mere consideration of the latter pur-
pose was one of the motives inspiring the alternative approach of building a new system
from scratch with the same functionality, a replica.
If preserving and reconstructing data may be called web archaeology, what are the
“scoops”and “brushes”in the digital practice? For the most part tools of digital archaeol-
ogy require manual calibration and application; automated procedures are in their
infancy. Working with legacy digital material brings home one crucial insight: whether the
unearthed objects are data, scripts or full blown software; their archaeology involves get-
ting the code to work. Born digital material is dynamic. Executing a script may yield differ-
ent outputs each time it runs. Static material does not react, let alone react differently.
The archaeologist will not be satisﬁed with images or screenshots.
This is in stark contrast to the existing practices of Web Archiving, viz. to preserve
snippets of the Internet as pages, as snapshots. The maturing of the ﬁeld of Web Archiv-
ing has been well captured in the volume edited by Julien Masan
es (2006). The pioneer-
ing work of the Internet Archive and its Wayback Machine from the late 1990s has been
broadened and institutionalised on a national scale in many countries. National Archives
and Media Archives have automated their harvesting with crawlers and ﬁlters. The
materials they gather, however, are static, or rather, are treated statically. Even if web-
sites contain code, they are saved simply as pictures of pages. Today, the Internet
Archive does more. It replays the harvested pages as much as it can. Our plea is to rein-
state the pages as they were born, not starting from the resulting page, but from the
server. Given the dynamic character of born digital material the archaeological
approach should be dynamic. To put it differently: such material is called “born digital”
to emphasise that its symbols are not just text. The text is considered as working code.
And for working code, emulation and replica offer themselves as feasible approaches,
each with their own tools.
De Digitale Stad: a local history
De Digitale Stad, the Amsterdam digital city, exempliﬁed the electronic social network. It
facilitated the exploration of all the possibilities to connectis, including almost inciden-
tally access to the Internet. It was designed with the city metaphor in mind. On 15 Janu-
ary 1994, the digital city opened its gates. Beyond the practice of earlier FreeNets, De
Digitale Stad appealed to its users to adopt the metaphor and create a true community.
It allowed the users to be “citizens”or “netizens”and enter the unknown world of the
The project was initially funded for 10 weeks by the city of Amsterdam, on the assump-
tion of bridging the gap between local politics and the ordinary citizens. The number of
148 G. ALBERTS ET AL.
subscriptions skyrocketed. After the ﬁrst 10 weeks, with the project clearly growing bigger
than anyone had anticipated, DDS acquired further funding to continue beyond the initial
experiment (Castells, 2001; Lovink, 2002; Rustema, 2001, pp. 42–67).
The ﬁrst version of DDS was a bulletin board system (BBS), a static menu offering the
user a choice of line numbers to continue towards further pages. Imagination was an
essential asset for the user to walk the streets of the digital city.
The second version made a major step to change from the Gopher communication
protocol, used in version 1, towards the newer HTTP protocol still in use today on mod-
ern web pages. Within weeks, yet another version of DDS was released. Through a major
overhaul this version 3 had become a truly interactive system, embodying the metaphor
of a city. The overarching metaphor was further detailed by such facilities as “post
ofﬁce”,“city square”and “caf
e.”These allowed the “citizens”to navigate the city more
intuitively. Users could fetch their mail at an email facility called “post ofﬁce”, they could
set-up their own homepages called “houses”that were reachable by traveling across
“squares”, those were web pages linking between each other, or they could hang out in
e”, which we nowadays see as a chat room. The city metaphor was introduced and
promoted the Internet as a common, a public space, which it hardly was in 1994/1995,
when network services were mostly available in universities, libraries and as private facil-
ities in large companies.
DDS grew amidst optimist expectations, expectations of technology having a demo-
cratisation effect. In the sense of spreading the technology itself among larger section of
society it certainly had this impact. Hope that the technical facilities by themselves would
bridge the gap between politics and public, and thus solve the representation and legiti-
matisation problems of politics soon evaporated. Such hopes were certainly played out in
acquiring the initial subvention of DDS by the city of Amsterdam. The idea of a push-but-
ton direct democracy is a recurring dream, also in DDS circles. In the same vein but stron-
ger and more speciﬁc for DDS were the expectations of an emerging community and a
new kind of sociability. Conceiving DDS as a commons, as a public sphere, deeply moti-
vated many of the early actors. Their motivation was even strengthened when in those
very years in the US the Internet rapidly evolved in a commercial connectivity with a new
kind of economy (Aspray & Ceruzzi, 2008). Lovink (2002) observes that not one but many
communities shaped around the digital city and created their own niches subcultures.
From recent research into the themes of the caf
es and lists in DDS, one may infer that the
electronic facilities did in cases serve as vehicles of emancipation.
Dennis Beckers and
Peter van den Besselaar have in a series of publications shown the dynamics of the various
groups involved in DDS-initiatives (Beckers & Besselaar, 1998; Besselaar & Beckers, 1998).
Thus, while Lovink (2002) characterised the divergence between the communities as cul-
tural differences, Beckers and Besselaar were more political in their analysis by pointing at
divergence through conﬂicting interests.
At a more fundamental level studies in sociology, STS and media students have hinted
at new kinds of sociability emerging in such connected commons as DDS. Castells, famous
for his trilogy on the Information Age (1996), in The Internet Galaxy (2001) presented DDS
and many other case studies as vistas on new forms of society. The more radical approach
in this direction was Howard Rheingold’s(1993) effort to continue where D€
Weber have left us with “Gemeinschaft”and “Gesellschaft.”Reinder Rustema (2001) emu-
lates Rheingold’s search for a “society”beyond Gesellschaft by the example of De Digitale
INTERNET HISTORIES 149
Stad. Sociology has not yet come to conclusions, but without any doubt DDS is part of the
empirical material to be reﬂected upon. Media studies on their part show that the indus-
trial society is superseded by the platform society (Van Dijck, Poell, & Waal, 2016)–of
which then DDS is not a forerunner or prime example.
Local frost and defrost
On 15–16 January 1996, the DDS servers were down for most of the night to allow for a
full backup of De Digitale Stad. A full 1-on-1 disk copy of all the servers running DDS was
created on 3 Digital Linear Tapes (DLT). DDS congratulated itself with a city frozen in time,
preserved “to be studied by archaeologists in a distant future”: the FREEZE. Further heri-
tage material was gathered at “gravediggers parties”and the Amsterdam Museum
installed a small exhibit on DDS.
In restoring old data, it soon came to light that the package would not simply unwrap,
or defrost. The DLT Tapes holding the FREEZE did not easily render their content. After a
good deal of searching for auxiliary hardware, the tapes had been read and converted
into the more common format of compressed gzipped tarball (.tar.gz). Initial attempts to
extract this tarball of 10 Gb failed because, for no apparent reason, it exceeded the avail-
able storage. After several tries, each time with more storage available, the ﬁles were
ﬁnally extracted to a network attached storage (NAS) with 12 Tb of free space. The size of
the completed extraction revealed why earlier attempts failed: the data ﬁlled little over
2.2 Tb of storage space. When searching for the cause why a 10 Gb .tgz ﬁle extracted to
over 2.2 Tb (220x its own size), we detected four corrupted ﬁles, each over 500 Gb in size –
quite possibly the effect of a “decompression bomb.”Omitting these four ﬁles, the extrac-
tion returned to reasonable size, approximately 35 Gb. The project “DDS 3.0 operational”
worked from these cleaned ﬁles.
However, in an effort to understand what had gone awry in handing down the legacy ﬁles,
and to make sure that no major parts of the original ﬁles were missing, we made a detour
going back to the original servers, still extant in the Amsterdam Museum. Other than the
FREEZE the content of these servers was not strictly dated, let alone of the same date as
the FREEZE. Because of its historical signiﬁcance, the original, but not necessarily opera-
tional, hardware is preserved at the store of the Amsterdam Museum. For the purpose of
our research project, we were granted access. For just this one occasion the original hard
drives were retrieved from these servers –and put back.
In order to read the 20-year-old hard drives, vintage equipment was needed, with
the sockets of the cables as major obstacles. The original server had eight hard
drives, connected using three different cable sockets. The interface was SCSI, which
is a parallel interface subject to different standards. The solution was found with the
help of a former system administrator digging up the ﬁtting connectors from an old
drawer. To read out the content of the hard disks a Linux live USB was set-up. The
disks were connected one at a time. A full-disk-image was made –and preserved
150 G. ALBERTS ET AL.
This sidestep of the project greatly improved our understanding of the legacy material
being studied, speciﬁcally the hardware. The newly retrieved content read from the origi-
nal servers, showed sufﬁcient overlap with the FREEZE of 1996 to conﬁrm the adequacy of
the cleaned ﬁles, but being of different date, was not included in our current reconstruc-
In January 2015, after the defrost and clean-up of the FREEZE had been achieved,
ﬁrst forays into the data were started. Since there was little indication of the struc-
ture of the stored ﬁle, except that the system was an old Sun SPARC system, inves-
tigations began with mapping the folder structures and sizes, and listing the
installed software. The ﬁrst observation was that the system was not as systematic
as one might naively assume. In particular, there were no systematic locations for
source code, if preserved at all, going withtheinstalledcode.Harshlessonforthe
archaeologist: with the programs installed and running in 1996, there will not even
be an indication of source code being preserved and, if so, where it might be
stored in the system. Another harsh lesson, even the non-techies have to learn
Software appears in various modes, basically “source code”and “binary.”Pro-
grammes are written by the programmer –with the help of a whole factory of tools
–written in a programming language. This written version is called source code.
This is the version one usually refers to when discussing programs. Source code is
just text, it does not work by itself. A program in source code is translated to ﬁton
a system, in our case on the Sun SPARC station. The result of translation is an “exe-
cutable”or “binary”code. This is the version that does the work. Hence, the digital
archaeologist looks for the “binaries”to see what a system can do, which engines
The tool doing the translation from source code to binary is called translator or “com-
piler.”In some special cases and only for speciﬁc languages, a tool exists which can do the
reverse translation, back from binary to source code, a “de-compiler”.
In exploring the ﬁles, it was of great help that former “citizens”vividly remembered the
avatars. This clue from oral history set the challenge of locating the relevant software.
With the introduction of DDS 3.0, the system had become truly interactive and logging in
became more than having access. Every user would now have an avatar representing him
or her whilst “walking”through the digital city. A program would generate a small icon-
like image representing the user, an avatar. Somewhere on the server there must be such
a piece of software performing that function, the avatar generator. In 1995, for lack of
memory space and operational speed, the system would not allow for pictures or other
complex images to be inserted. Therefore, the avatar generator created simple but effec-
tively distinctive images for every user, varying on a pattern inspired by a character from
the Muppet show, Beaker.
INTERNET HISTORIES 151
Exploration of the frozen data yielded no such ﬁle as avatar or avatar generator. A fur-
ther clue was to search for the programs governing registration to the system, because
that was the procedure including the creation of an avatar. This lead to locating Apache
and its original conﬁguration ﬁle, from there the original registration page was traced. The
registration scripts had been written in CGI PERL, easily readable, and thus the steps of
registration could be traced.
Further recollections from oral history suggested that these icon images were never
called avatars, but DoDoS, in an apparent play on the name of the system DDS. And
“Dodo”is the extinct bird from Mauritius. So, the avatar generator would revive the extinct
bird. The script, now located, had the name Dodo.cgi. This shows the development of ter-
minology of the past 20 years. Where nowadays “avatar”is the appropriate jargon, it
shows that this term was not as pervasive back in the day.
Soft lesson for the archaeologist: follow the challenges set by oral history.
The reconstruction project
With the avatar generator resuscitated in 2015, and several other chunks of software
brought to life in 2016, we could not resist the temptation of trying to get the whole sys-
tem back into operation, project “DDS 3.0 Operational.”The project group soon split up
and worked in two opposite directions. One part of the research focused on what seemed
to be the most obvious thing to do, viz. to try and run the original software again. The
other part focused on the idea of reinstating the user experience, regardless of the
machinery behind the screen. This second route, to replicate the digital city, leads to
rebuilding the system from scratch and working per modern technology and standards.
Preservation of the digital city
The goal of the FREEZE had been to preserve De Digitale Stad as it had existed and run
two years after its introduction, to ensure for future generations the possibility to experi-
ence and to study the early days of the web. However, simply backing up one’s data does
not automatically result in preservation. A 1-on-1 copy may be historically accurate, but
such a copy loses much of its attraction if it cannot be run, if the context of the original
production server is missing. DDS ran on the Solaris SPARC computer architecture.
Because Sun SPARC has become proprietary software and hardware from Adobe, with
prohibitively high licensing costs, virtualisation seemed the road towards reconstructing
152 G. ALBERTS ET AL.
an adequate remplacant context. It proved to be not that easy, since emulation of the
SPARC architecture on a different system has usually low performance, particularly on the
most common architecture in the 2010s, the x64 architecture.
The one research direction was to revive De Digitale Stad in its original state, from the per-
spective of the software running it. Like a hyena going for the innards, the software
archaeologist goes looking for the software that was run, the software that was executed
to create the performance of the system. The search is for the executable ﬁles, or “bina-
ries.”Being familiar with the operating system, mostly Unix, and knowing how to identify
and search for these ﬁles, binaries, archaeologists use such instructions as “grep”and
“ﬁnd”to trace the location of the executables in the FREEZE.
Lesson of exclusion: only close familiarity with the legacy system and its operating sys-
tem will allow one to do the work of a web-archaeologist. To a large degree this is tacit
Knowing which programs did in fact do the job is not enough to run them again. Bina-
ries are executable only in a speciﬁc context, in this case the SPARC context.
At this point the web archaeologist, in general the software archaeologist, bereft from
straightforward automatic tools like a Virtual Machine doing the work, must revert to more
subtle and more individual methods. For the DDS 3.0 Operational project, a pragmatic deci-
sion was made with far reaching consequences. In spite of the dependencies –i.e. the points
where the programmer had created constructions particular to the speciﬁc machine –we
chose to emulate on an x64 architecture. The consequence being that the binaries on the old
system (Sparc) are of no use and one has to create new binaries for the new system (x64). We
had to go back to the source codes, the programmes as written, and would in that mode of
the programme have to deal with the dependencies each individually. The task at hand was
now to ﬁnd the original versions of the software running in 1996 and compile these pro-
grams anew for the operating systems coming with the x64 architecture. Fortunately for this
project a good deal of the source code, even if not systematically preserved, was retrieved
scattered throughout the FREEZE. For future archaeology there is a use for systematic tools of
de-compilation, in cases where there is no source code is available at all.
The positive side of emulating on an x64 architecture that this system is so common
that one may expect it to survive for the foreseeable future. We are good, not for archaeo-
logical stretches of time, but at least for one or two generations. One may hope that the
present work of getting a version of DDS running needs not be repeated from scratch in
20 years from now.
The downside is that corners of DDS with the heaviest dependencies are hard to restore.
In particular, the system for authentication, logging in, was most speciﬁc for the SPARC archi-
tecture and has therefore been left out of the present emulation. By consequence the avatar
generator, Dodo.cgi, was found and brought back into operation, but the programs constitut-
ing its context, logging in, are not. So, we can play and make avatar-wallpaper today, but we
cannot have our DoDo walk through the system for us. When the former systems administra-
tor was consulted on how to emulate this feature –which he had in fact programmed –his
answer was: “consider not to.”In terms of DDS 3.0 as it ran in 1996, one can only visit the city
INTERNET HISTORIES 153
in the “guest”mode. In other parts of the DDS program functionality was restored by prag-
matic patches. Thus, this emulation comes with compromise.
Along the route of emulation, the digital city has been restored as close to the original
project as possible. In as far as it functions, it does revive with a feel of authenticity –so
say DDS’s former inhabitants. As an extra beneﬁt the emulation preserves code of histori-
cal interest and allows comparison to modern standards. It reveals the challenges as the
DDS developers perceived them and the answers they chose when designing the early
pieces of Internet. It makes the look and feel of born digital media accessible, including its
inner workings and some of the thoughts behind it.
The other approach was to revive the digital city from the point of view of the end user,
replicating the original system as closely as possible using current technologies of 2016.
This was feasible because, compared to today’s projects, the size of the digital city was rel-
atively small and quite straightforward. Although providing static images only, the Way-
back Machine did show the appearance of DDS to the end user. The programming for the
replica was done in parallel to, and strongly inspired by, the emulation. While missing out
on historical accuracy in the back-end software, the replica proves the feasibility of creat-
ing a user interface very near to the original, and in practice indistinguishable from the
emulation –so say again DDS’s former inhabitants. The user will “walk”through the city
without noticing it is not the original software. In that sense the experience is effectively
Major advantages of the replica are, beyond its technical maintainability and sustain-
ability, its security. If one were to consider the creation of a publicly accessible version of
DDS 3.0, for example as a museum exhibit, a replica would offer a doable solution;
whether feasible in terms of museum practice, remains to be seen. As long as mainte-
nance is kept up a replica can be secured and it could be ﬁlled with part of the legacy con-
tent upon authorisation by the, former, users.
The archaeologist’s brushes
If preserving and reconstructing data may be called web archaeology, what are its brushes
and spades? Digging up the digital city has in large part been a manual labour. In fact
relied heavily on the tacit knowledge typical of craftsmanship. It required familiarity with
Unix and other operating systems. Key element in the tacit knowledge is practical insight
in the way server systems were usually built up, preferably joined with expertise in today’s
systems. It took the joint forces of former system administrators who had not forgotten
their trade and were willing to share, and archaeologists who are able to absorb. The latter
are talented computer science graduates using some of their academic lessons and
heavily relying on their experience of working as system administrators to pay for their
studies. Their concerted effort has allowed for a “handmade reconstruction”of DDS.
But in a digital environment, should one not expect “systematic and automated tools”?
Some tools do exist in the realm of software archaeology. For example, source code is
154 G. ALBERTS ET AL.
compiled into executable code. For some situations, e.g. programs written in high level
programming languages, tools do exist for the reverse process: decompilation to recon-
struct source code from executable ﬁles. Further automated tools are dearly wanted, like
excavation tools to help to recognise the various types of ﬁles.
Beyond the archaeological metaphor
As far as craftsmanship goes, the metaphor of archaeology sounds attractive and serves
well. But in fact, what we are describing here has little to do with archaeology proper, but
all with handing over, i.e. heritage through living tradition. It was not Pompeji but Amster-
dam 1996 with its inhabitants still around today.
Once access is gained, from 2015 through unfreezing the FREEZE and now by its
dynamic reconstruction, research is not solely focused on software. The ﬂoor is open for
an analysis of the content. Approach and tools are quite different from the above.
Whether historical, sociological, anthropological or phrased by media studies, the further
research questions involve a completely different set of tools. Technologies of searching
through data of ﬁltering and of visualisation are available in the computer sciences and
the data sciences. Some of the more sophisticated tools are being developed in a branch
of data science going under the name of forensics.
In every part of the research on DDS, even in the most technical niches of the recon-
struction process, the “archaeological”research is mingled with dialogue. Oral history
helps. More than that, the intermingling reminds the researcher that in fact the historical
approach is the umbrella underneath which it all makes sense: the dusting and scooping
and ﬁtting fragments together. Our archaeology of the web has all the beneﬁts and the
pitfalls of contemporary history. The takeaway message is that in spite of what the recon-
structed operationality of the software may suggest to some, the historical distance
remains. It is in the very process of reconstruction that the archaeologist is reminded of
the ineradicable historical distance.
Manipulating such “young data”on people acting as inhabitants of DDS two decades ago,
and existing as fellow citizens today, may well produce a moment of shivers. The archaeol-
ogist ﬁnds herself swimming in a pond of personal information. Applied to the study of
“young data”the metaphor of archaeology is brutally misleading. The work is social sci-
ence or contemporary history and carries with it the social responsibilities tied to such sci-
ences. Accessing young data poses major privacy issues.
In 1994 a major shift occurred that allowed people to share intimate experiences with
the click of a button. Privacy was a difﬁcult question and people did not, nor could they,
predict the long-term implications of posting their information. Google, Facebook,
Amazon and many other tech companies use these data to their advantage. Where users
used to be proud “citizens”or “stakeholders in the commons”, today they are seen as nat-
ural resources for those companies having evolved into platforms. The meaning of “data”
has shifted dramatically in the past 20 years, from information bearing, to monetary gain.
The notion of privacy has changed even more.
INTERNET HISTORIES 155
As a researcher in web archaeology, one will quickly ﬁnd oneself dealing with legal and
ethical concerns regarding the privacy of historical subjects. The web was ﬁrst and fore-
most a communication medium and as such is ﬁlled with personal information of its users.
Personal information is not just the information directly linked to one individual person.
Any piece of information, that could possibly be linked to an individual, counts as personal
information. Linking data to a username would link all these data to an individual, once
that username has been linked to a person. By consequence, most of the web is subject
to privacy laws, as most of the content is about, or created by, individuals –regardless of
whether they can be directly identiﬁed through this content.
The ensuing legal considerations may vary greatly depending on national or state laws.
And since the Internet hardly stops at national borders, legal considerations are compli-
cated even further. Inside of the Netherlands, when handling personal data, the concept
of “purpose limitation”(Dutch: doelbinding) is a core concept of the privacy legislation. Its
purport is that personal data should only be used for the purpose for which they were
acquired (Ketelaar, 2000). Medical data are to be used for health care only and not by
insurance companies; income data gathered for taxes should not be shared beyond tax
administration, etc. An exception to this rule is made if the purpose is historical research.
However, leaning on this exception will only turn the legal consideration into an ethical
one, which it already was.
By publishing web archaeological ﬁnds, the researcher will encroach on the privacy of
the group being studied. And due to the freshness of the sources in the FREEZE, chances
are that the group being studied is still alive and might object to the personal data being
spread. Therefore, researchers must weigh the potential harm their research might cause,
against the potential gain for society of including the personal data under consideration.
The more so while, as researchers, we are inherently biased towards publishing. Therefore,
other researchers must be consulted, and in an ideal world ethical committees should be
installed overseeing projects of web-archaeology (Markham & Buchanan, 2012). In
the DDS case, our provisional measure has been that any access to the retrieved data for
the purpose of research is given under an agreement of conﬁdentiality between the
researcher and the Amsterdam Museum, procuring the source material.
If for historical research, privacy issues may be addressed in similar ways as other research,
with an extra caveat, because of the rapid changes in Internet practices, the discourse
takes a different turn in museum context. Suppose the purpose were to create a public
exhibit out of legacy web sources, not only should the above reticence towards the publi-
cation of personal information be taken into account, but technical matters need to be
considered as well.
Systems built in the 1990s were not created with today’s practices of collecting data in
mind. Technologies of protection have evolved accordingly. The legacy systems, even if one
ﬁrst century. This thought has considerably strengthened the inspiration to build a replica
next to the emulation, which in terms of safety and security must be judged hopeless.
156 G. ALBERTS ET AL.
Treating digital heritage dynamically
The established practice of archiving the legacy of the web is to store “snapshots”, that is
take a momentary image of a website and download it. This can be done automatically by
so-called crawlers and yields enormous haystacks of information on the history of the
Internet. Not only the Internet Archive and the Wayback Machine operate like this, national
libraries and archives have adopted and standardized this approach (Mason 2007). The
Internet Archive will replay the pages thus harvested. In doing so, they are reducing web-
sites to mere pages. But the Internet is not a book, its sites are generated dynamically, be
it at a pace of once per day or ten thousand times per second. Upon return the visitor will
ﬁnd a new thing. Michel Serres (2015) reminds us what a parochial way of organising our
knowledge it is to put it page by page, now that we could liberate ourselves from that for-
mat thanks to the very Internet. As early as 2003 Helen Tibbo (2003, p. 16) observed the
dynamic nature of the web:
A related problem is the Web’s dynamic nature. Web archiving initiatives can only preserve
“snapshots”of sites or domains at the expense of their dynamism, rather like insects trapped
in amber. Once snapshots of Web content are located outside the active Web, it is arguably
missing one of its most characteristic properties.
The appeal was picked up by Michael Day in (2006, p. 193). We urge to take conse-
quence of the dynamic nature and call for an adequate, dynamic approach. Far from inci-
dental, the complexity and dynamism of the web reﬂect its digital nature. And the way
we conceive of this heritage should change accordingly. The web should be seen not just
as text and image. It is working text: code or software. Preserving the web, thus conceived,
may well burst the frame of archiving. In that sense web archaeology is a truly new ﬁeld,
an extension of archiving proper.
The dynamic character of a website’s content expresses the underlying code. Whether
simple CGI scripts, markups or complicated software, executable ﬁles lend the web its
dynamic character. This code is the “working text.”Without it the documents as the user
sees them are different, static. The dynamic approach to web heritage implies to take the
text on the surface inclusive of the underlying code, its context. To be able to contain the
context of a web page one cannot assume static snapshots as a solution. To properly pre-
serve web content for future research, it must be stored with its dynamic elements in
1. In Dutch “gedeponeerd in een archief ter bestudering voor archeologen in een verre toekomst”,
in “De digitale stad bestaat (bijna) twee jaar”. Post on the DDS, 1996. https://hart.amsterdam/nl/
page/37138/gevondenfreeze http://web.archive.org/web/20100830120819/ http://www.alme
2. Unpublished reports of student work.
“History of Digital Cultures”is a regular graduate course by Gerard Alberts in the joint MSc Computer
Sciences programme of University of Amsterdam and Free University Amsterdam. “DDS 3.0 opera-
tional”was a special course in 2016 taken by Marc Went, Robert Jansma, Ronald Bethlehem, Tim
INTERNET HISTORIES 157
Veenman, Kishan Nirghin, Millen Mortier and Thomas Koch. Reports on this work are available on
[Re:DDS 1.0]. The authors further extend their gratitude to Tjarda de Haan and the team at Amster-
dam Museum for support throughout the project and hosting the presentation, to Waag Society for
facilitating our meetings, to Jesse de Vos and Dennis Beckers for guest lectures, to Theun van den
Doel for tireless moral support, and most of all to former system administrators Michael van Eeden
and Paul Vogel for immediate help and for sharing their implicit knowledge of the systems. We
gratefully acknowledge the Digital Preservation Coalition for distinguishing The Digital City Revives,
of which our DDS 3.0 operational is a part, with the 2016 National Archives Award for Safeguarding
the Digital Heritage.
No potential conﬂict of interest was reported by the authors.
Notes on contributors
Gerard Alberts is an associate professor for history of computing and history of
mathematics at the Korteweg - de Vries Institute for Mathematics at the Uni-
versity of Amsterdam. He is editor of the Springer Series History of Computing
and a member of the editorial board of IEEE Annals of the History of Comput-
ing and of this journal.
Marc Went is a master student at the joint MSc Programs Computing Sciences
and Information Sciences of the Vrije Universiteit Amsterdam and University
Robert Jansma is a master student at the joint MSc Program Computing Scien-
ces of the Vrije Universiteit Amsterdam and University of Amsterdam. He is
also a research assistant at the Amsterdam Museum involved in sustainably
preserving De Digitale Stad in the project “DDS herleeft”.
Marc Went http://orcid.org/0000-0002-1133-5805
Robert Jansma http://orcid.org/0000-0003-1121-4189
Aspray, W., & Ceruzzi, P.E. (Eds.). (2008). The internet and American business. Cambridge, MA: MIT
158 G. ALBERTS ET AL.
Beckers, D., & Besselaar, P. (1998). Sociale interactie in een virtuele omgeving: De Digitale Stad [Social
interaction in a virtual environment: The Digital City]. In Informatie & informatiebeleid 16-4.
Besselaar, P., & Beckers, D. (1998). Demographics and sociographics of the digital city. In T. Ishida
(Ed.), Community computing and support systems —Lecture notes in computer science (Vol. 1519,
pp. 108–125). Berlin: Springer Verlag.
Castells, M. (1996). The information age: Economy, society and culture (Vols. 3). Cambridge, MA:
Castells, M. (2001). The internet galaxy: Reﬂections on the internet, business, and society. Oxford: Oxford
Day, M. (2006). The long-term preservation of web content. In J. Masan
es (Ed.), Web archiving (pp. 177–
199). Berlin: Springer Verlag.
Ernst, W. (2013). Digital memory and the archive. Minneapolis: University of Minnesota Press.
Foucault, M. (1969). L’arch
eologie du savoir [The archaeology of knowledge]. Paris: Gallimard.
Jones, C. (2004, June). Seeing double: Emulation in theory and practice. TheEerl King case study. Paper
presented at the electronic media group. Annual Meeting of the American Institute for Conserva-
tion of Historic and Artistic Works. Portland, OR.
Ketelaar, F.C.J. (2000). Elke handeling telt. archiefdiensten en de wet bescherming persoonsgeg-
evens [Every act counts. Archival agencies and the law on the protection of personal data]. Neder-
Lovink, G. (2002). Dark ﬁber: Tracking critical internet culture. Cambridge, MA: MIT Press.
Markham, A., & Buchanan, E. (2012). Ethical decision-making and internet research. Recommendations
from the AoIR ethics working committee (version 2.0). AoIR. Retrieved from https://aoir.org/reports/
es, J. (Ed.). (2006). Web archiving. Berlin: Springer Verlag.
Mason, I. (2007). Virtual preservation: How has digital culture inﬂuenced our ideas about
permanence? Changing practice in a national legal deposit library. In M.V. Cloonan & R. Harvey
(Eds.), Preserving cultural heritage, Library trends, 56-1 (summer 2007) (pp. 198–215) [Special issue].
Presner, T., Shepard, D., & Kawano, Y. (2014). HyperCities. Thick mapping in the digital humanities.
Cambridge, MA: Harvard University Press.
Rheingold. H. (1993). Virtual community, homesteading on the electric frontier. Reading, MA: Edison
Rustema, R. (2001). The Rise and fall of DDS: Evaluating the ambitions of Amsterdam’s digital city
(Unpublished master’s thesis). University of Amsterdam, Amsterdam. Retrieved from http://rein
Serres, M. (2015). Thumbelina: The culture and technology of millennials. London: Rowman & Littleﬁeld
Smith, J., & Nair, R. (2005). Virtual machines: Versatile platforms for systems and processes. San
Francisco, CA: Elsevier.
Tibbo, H.R. (2003). On the nature and importance of archiving in the digital age. Advances in
Van Dijck, J., Poell, T., & Waal, M. (2016). De platformsamenleving. strijd om publieke waarden in een
online wereld [The platform society. The struggle on public values in an online world]. Amster-
dam: Amsterdam University Press.
INTERNET HISTORIES 159