A place where I can talk about my research and other interests -- history, magic, ham radio, technology, and making things.

Friday, October 5, 2007


In a recent class on archival practices, the instructor made the comment that in multi-level archival descriptions of collections, item descriptions are often not included. Items are the lowest level in the hierarchy of archival descriptive practice, and are the individual pieces of the collection. Typically, an archivist's time cannot be devoted entirely to creating item descriptions for each piece of the collection, and so it just doesn't get done.

A historian, researcher or user of archived material primarily interacts directly with those items for their work, rather than just the macro-level collections. They are the end users of the material, and interact with it in ways that archivists may not have time or the inclination to do. Descriptions of the items in a collection could be very useful to those researchers.

This situation possibly suggests a disconnect between the creators of the descriptions and the end users of a collection. One way to handle that disconnect has been to ask the archivists for assistance. Although item descriptions have not been written, archivists have worked extensively to preserve, catalogue and describe the collection. This intimate work with the collection has likely left elements of various parts of the collection in the memory of the individual archivists.

What can be done if the archivist who has that knowledge in their mind does not happen to be at the archives on the day you are in to do your research? Or, what happens when the archivist retires or leaves that collection? Is that knowledge lost as well? The gap still exists, and there are no long-term or permanent initiatives to bridge that divide.

Perhaps wiki technology can mediate that gap.

Multi-level descriptions are useful for finding and identifying material and portions of the collection of interest to the researcher. They are well-defined, clear to understand, and aide in narrowing down where material relevant to the researcher's work may be. These could be left intact, and would form the base of the collection descriptions. They could be made by the archivists, as they currently are, and would not be part of the wiki portion of the descriptions.

The descriptions could then branch from highest level to the lower levels. The upper levels, made by the archivists, would be fixed and unchangeable by users. The lowest level -- the item descriptions -- would be the wiki. The users of that material could voluntarily add descriptions of what they are working with to a wiki as they use those items. As more material gets used, more of it will be described. Others who return to the same material have the option of updating those descriptions by adding their own. They may have even been led to that material by reading the earlier description(s).

This approach would have the advantage of creating and providing more information about the collection. It would bridge that gap between the end users and the creators of the collection descriptions, and provide useful information that otherwise would not be a part of the descriptions. On a conceptual level, it would also involve the end user in the creation process of the descriptions. This might create a greater connection between the user of the material and the collection as a whole.

Issues of trust appear to be one of the major concerns of wikis, but the specialized case of archival descriptions mitigates those issues. The current users of archives, and therefore those who would be most likely to use a system like this, are probably a very small but highly interested portion of society. Disclaimers could be provided at the item descriptions to make it clear that the information there was not created by archivists but rather by the users of the material, and that it can be updated by others. Finally, since archival material is typically used within the confines of the archival institution, limits could be placed on the IP addresses that can post material. In that case, only computers physically in the archives could update the descriptions while they are with the material. Users are often required to sign in with the archival institution upon arrival, so authors of entries to the wiki, even if anonymous or pseudonymous, could likely be determined by the institution.

Moderated wikis also exist, and that could be one other means to control the information if that was deemed necessary by the archival institution.

The item descriptions would result in a richer database for use by researchers. The wiki could be made available online for anyone interested to read and look for information. Researchers could then go back and read their own descriptions if they wanted to it, or see how others described the same material. How descriptions change over time may also prove to be valuable information. The knowledge would be more durable, more complex, and provide an added resource for the users while making them a part of the archival process.


F. Grace Dungavell said...

I really like this idea Devon! It would also be neat if things like images could be added to these descriptions. Since many materials are fragile, a single photocopied image of the item (say a picture) could be uploaded once and then accessed by countless researchers (ignoring potential copyright issues of course). Very cool!

Devon Elliott said...

Photos would be great -- good point! I wonder how big of a role copyright and/or privacy issues would influence a practice like that? Exposing fragile materials to the bright light necessary for scanning might be frowned upon for some items. In those cases, maybe just a low-quality digital photograph could be taken without a flash -- it would be no different than looking at the photo. That way there would still be something more added to the descriptions.

Ben W. Brumfield said...

This is a fantastic idea. I especially like the ability to mine the minds of the "very small but highly interested portion of society" who are using the archives and the finding aids.

In addition to the user-editable aspect of wikis, it seems like this would be a good place to use wiki-style hyperlinks.

Individual editors of the item description -- presumably the archive's end-users -- could explain how an item is related to another item in the collection and link to it. Links could also be made from an item to other elements within the multi-level, hierarchical description.

If these item description links flourish, they can be mined to create an alternative set of finding aids -- a parallel folksonomy that supplements the archivists' hierarchical description. See Clay Shirky's "Ontology is Overrated" (paper, audio) for some of the potential this has.

Quite possibly, the interconnections between items will be used for something altogether different from the archivists' catalog. I've been mining these links within transcribed text to create indexes, for example. Similarly, entire projects on Wikipedia use the categorization feature to manage articles needing attention from subject specialists, administrators, or arbitrators.

Devon Elliott said...

Great thoughts, Ben. Thanks for sharing them, as well as for the link to the article.

I get the impression that some efforts are underway that would help in linking and in "mining" that information within the archival community. The Library of Congress helped to develop Encoded Archival Descriptions (EAD) that comply with SGML and XML standards. Descriptions encoded like that would likely be even easier to plug into a framework like I suggested in the blog article.

Gavin Robinson said...

The UK National Archives are already experimenting with a public wiki to supplement their catalogue. As well as item level descriptions they're encouraging people to post transcripts of documents. I've been trying it out this week and I'm really impressed with it. It just needs a critical mass of users and content.

Devon Elliott said...

Thanks for the link, Gavin. The wiki at the UK National Archives looks very pragmatic. Keeping the wiki as a separate system with links from the catalogue would have its advantages. I found it hard to tell just how much content is on there. Hopefully more people start to use it.

It was interesting to see the input from staff at the National Archives as well. A system like that could act as another layer of institutional memory. Archivists each have specialized knowledge about the collections at their institution, but if that knowledge is not recorded somehow, it leaves with the archivist. If some archivists recorded some of their knowledge into the wiki, that information may be more durable over time. Such information would be useful to users of the collections, and would not be dependent on a particular archivist being "in" on any given day.

french panic said...

Oh how cheery it is to see all of this enthusiasm!

As archives are also notoriously underfunded and understaffed, I am anxious to hear where the $$$$$ is going to come for this type of wiki initiative. The archives I work for has exactly one computer (mine) and nothing electronic available for researchers onsite or on the web. Actually, even our paper finding aids are incredibly incomplete.

Another problem is that since archives are generally underfunded, jobs are being offered with a very low rate of pay, which is tolerable for newbie archivists who want to gain experience. Once they have that experience, they move on to another archives, taking all of their experiential knowledge with them. For example, my low-paying position has had 4 different archivists here in the past 8 years. There is no continuation of knowledge. When I leave here, my familiarlity with the collection will leave with me, and someone new will start from scratch - it is an incredibly depressing situation, but it IS nice that historians are thinking about this situation.

I would loooove to do item level descriptions. Just as I would looooove to catch up on my 6 month backlog of research requests, and my 5 year backlog of processing "new" accessions, and put up a virtual exhibit on the web, and implement an intelligent, useful records management system, and actually be able to preserve items correctly instead of making do with what's available. Oh yeah, and I'd also like to find some time to educate people on what an archives is, exactly, and how valuable it is despite that it produces no revenue.

Any ideas, digital historians?

Devon Elliott said...

Interesting and very pragmatic concerns, french panic.

I think some of your concerns would be addressed through a collaborative approach offered by wikis. In the case of your lament over archivists leaving archival institutions, and taking their experiential knowledge with them, the wiki could provide a storage base for information and knowledge that will last over time. The information would be more durable over time, as it would be rooted in the server hosting the wiki, rather than in the individual archivist.

That you would love to make item descriptions for your collection would potentially be alleviated by wikis as well. Instead of the burden being entirely on you, it is shared amongst those using using the collection. The users interact with the material at that item level, and they could perhaps share their knowledge about it. If you had time to contribute, or were working on something and found a few minutes to cut and paste what you were working on onto the corresponding wiki, even better. Such things are not getting done in most cases anyway for all the reasons you mentioned, so why not try to get the information from another source -- the users?

The money is a lot trickier. Offhand, I'm thinking Gavin Robinson's above example of the YourArchives in the UK might be a good start. That system looks like it is run separately from but linked to the official descriptions. If EAD was used to make the descriptions, I think it'd be possible to make near automatic links to wiki pages using the metadata tags as attributes in the links.

In Canada, there is the Canadian Archival Information Network (CAIN) which allows searches of a number of archival collections. Perhaps an arrangement could be made to use a centralized server like that to host the wiki...?

Ultimately, organizations like the Association of Canadian Archivists would likely need to continue lobbying the federal government for support in such an initiative, as they have done in the past.

Thanks for raising those issues. I'm only taking a course on archives at the moment, and view them very much from a historian's point of view, so it's great to hear from someone actually in the field!