OPF Hackathon on Emulation in Freiburg
In the middle of November 2012, the first OPF Hackathon on Emulation took place in Freiburg, Germany. It brought together practitioners from different national libraries, library information services as well as a couple of researchers in the domain. The aim of the three-day Hackathon was to work on practical use-cases and real-live challenges stemming from actual collections and legal requirements. In parallel to the hack and experiment sessions, a couple of presentations looked into a wide range of aspects with regards to emulation and related access strategies. The libraries brought a number of interesting use-cases to the event which were perfect examples for the application of emulation such as:
- Personal archives of people from institutions consisting of a wide-range of different file-types from a certain era
- An old bookkeeping software with company data locked into the built-in
- Several computer games for different architectures
- Old documents of proprietary formats
Additionally, a hardware item was brought by the University of Graphics and Design (HfG) in Karlsruhe, an old end 1990ies iMac which was used to create computer art and it was possible to render a couple of the Transmediale digital art festival CD-ROMs.
The introductory presentation gave an overview of the state-of-affairs in the field of emulation and actual challenges that had been discussed at the iPres 2012 in the four workgroups: “Metadata and Context”, “Digital Art and Computer Games”, “Web archiving and Big Data” as well as “Sustainable Emulation by Community Involvement”. It was followed by an overview on the actual state in metadata standards of PREMIS and the TOTEM technical registry presented by Portsmouth University. The upcoming PREMIS 3.0 will contain elements to describe original environments. During the experiments, the iMac running an Mac OS 8.5 was successfully imaged using a bwFLA workflow and restarted in SheepShaver. The complete system is now preserved in a more durable form and running in an PowerPC emulator which is accessible both locally or via an bwFLA web service. Another use case was an old commercial DOS application dating back to the beginning of the 1990ies from the holdings of the National Archives of the Netherlands. The “Exact” accounting software together with a proprietary database represents economic data of a company which should be made available and exportable again. To compare different original environments a DOSbox, VMware Workstation and QEMU with different MS-DOS setups were prepared. The application is run-able in all environments, but the actual access was impossible because of either missing software license keys for fresh install or username/password combinations for the existing installation. This highlighted the considerations required for future ingest procedures – to preserve relevant functional metadata and user knowledge how to operate the artifact.
The second day of the event was dedicated on the special challenges of digital art and computer games for various platforms. Here D. Espenschied from the bwFLA project of the HfG and A. Lange from the Computer Game Museum in Berlin showed the unique challenges but also great opportunities posed by these types of objects and their preservation. The computer game scene is very active in producing and maintaining emulators. The tools produced by the enthusiasts like the DOSbox was successfully re-used for “serious” preservation tasks like “Exact”. A possible solution, the KEEP Emulation Framework, was presented by J. v. d. Hoeven of the National Library of the Netherlands. The package is able to handle a couple of formats for a few platforms, but still efforts are needed to extend the framework regarding emulators and original environments. Parallel to the presentations and group discussions, some problems were solved: Older DOS software was shown running in different emulation environments. For different personal archives various original Windows environments such as Windows 3.11 or Windows 98 were used to render various types of files. These environments demonstrated that keeping context intact, like the dependency of text documents to an old MS Access database, helps to reproduce the original artifacts. A by-product of the experiments was the insight that older media, particularly 720kByte floppy disks are no longer easy to read in today’s USB floppy drives. A fall back was required, an older PC system with a standard drive was required, a problem that D. Anderson addressed with his plea hardware museums the day before. A. Strauch from the bwFLA project in Ulm addressed two real-world e-science use-cases looking at dynamic and complex objects and presented a questionnaire aimed to find out what would be relevant for emulation in the scientific domain, one example relating to the field of embedded development, the other research data and use-case dealing with financial engineering.
The final day of the Hackathon structured the preservation issues and problem sets of the previous days. The possible access solutions for the individual sets of objects were addressed practically. In addition, individual sub-groups discussed issues of long-term database maintenance, and how future reading room systems for libraries, archives and museums should be designed. Furthermore, a couple of talks on technical solutions were scheduled. J. Latocha from the bwFLA project in Freiburg gave an overview on the potential security and privacy challenges of emulation-as-a-Service (EaaS) and how they can be tackled. The bwFLA project implements an EaaS prototype at the moment. It is a central component for various bwFLA preservation and access workflows. Other issues discussed were (distributed) archives of standard software components required to reproduce the original environments. A joint effort would help institutions to save on setup time and share the knowledge. The presentation by S. Semaan and R. Zahoranski (from bwIDM project that helped to integrate bwFLA components into a IDM Federation) turned to the issue of proper authentication and authorization of users of EaaS and a federated software archive.
The event produced a couple of good results and documented a number of use-cases which can be found in the OPF Hackathon wiki. The challenges of emulation are interdisciplinary and consist of curatorial, legal and technical parts. As D. Anderson put it, the legal part is easiest to solve – by lobbying and policy making. The curatorial practices have to evolve and the institutions will learn how to deploy and use emulation-based access strategies. Solving emulation for art and games should help to deal with most other emulation cases. In order to push the application of emulation forward, it could be promising to create a consortium of relevant libraries and archives.
By Dirk von Suchodoletz, posted in Dirk von Suchodoletz's Blog
There are no comments on this post.