forum: Interview: Geospatial Computing
Leif Isaksen is a Senior ICT Development Officer at Oxford Archaeology and Moderator of the Antiquist forum for ICT/Cultural Heritage. His interests lie in the challenges of integrating different representations of space and time, as well as issues surrounding the openness and accessibility of heritage information. He has written papers on the Network Analysis of transport systems, the semantic web as a tool for data integration and the use of Free and Open Source Software for GIS in Archaeology.
Stuart Dunn is a research associate in e-science methods at AHeSSC, and a co-leader of AHeSSC's e-Science Theme at Edinburgh. He is interested in the theory and practice of geospatial archaeology, the storage and management of geographical data in digital libraries, and the application of GIS in networked environments. He has published on the environmental archaeology and vulcanology of the Aegean region, and is a visiting research fellow in the University of Reading's Archaeology Department.
Leif and Stuart - first of all let me thank you for taking part in this online interview! In less than two weeks time we will be in Edinburgh discussing Methods in Geospatial Computing for Mapping the Past, which is the title of the workshop you have organised with support from the Methods Network. I was recently discussing the nature of Digital Humanities with a colleague who commented that Digital Humanities was (at least for most of its practitioners) about text, i.e. about XML encoding of sources, especially for the historical disciplines in the field. Your workshop takes a different approach and the word 'text' is not even mentioned in the announcement. What do you think geospatial computing has to offer that goes beyond text?
Space vs. Texts
It's hard to know where to begin!
I don't think I'm being contentious by saying that the Humanities (even the past-oriented ones) deal with a lot more than just texts. In fact they potentially deal with anything and everything we do and experience as human beings. I know quite a few archaeologists who would argue that texts aren't the most important product of humanity. I think that the Digital Humanities probably ought to reflect that diversity.
Space, and the way it is perceived, experienced and managed, clearly has a tremendous impact on human behaviour. But much like texts, that process is one of interpretation and negotiation - it's that interaction that we need to try and get a handle on. Modern computer technology, from GIS to Google Earth and even Second Life, has made it far easier for us to model some of the physical characteristics of space, but we are still wrestling with how to use it in order to get an idea of how past cultures understood and utilised their environment. We'd also like to look at some of the more nuts and bolts aspects of using spatial technologies - To what degree can we use different information sets together? How far can conclusions from one spatial analysis be applied to another? How do we integrate time?
I should add that we're certainly not setting ourselves up in opposition to our more literary-minded colleagues. In fact my paper is partly about Ptolemy's Geography - a 'world map' from Antiquity that is entirely textual.
Space vs. Texts
It is quite right to say that we are not setting ourselves up in opposition to those working in textual fields. In fact, I am keenly anticipating the opportunity to facilitate a dialogue between those disciplines and subjects more traditionally associated with geospatial computing, like archaeology. We are way past the point where that latter simply want to put pins in maps - it is about contextualizing, managing, visualizing and researching information about space and time. As the technologies and methods advance, the opportunities to use texts as a source of such information become ever greater and more exciting.
There was a panel on this very topic at the recent Digital Humanities conference in Illinois: http://www.digitalhumanities.org/dh2007/abstracts/xhtml.xq?i...
History vs. Archaeology and 3D models vs truth
I'd agree with Leif that the Humanities deal 'with anything and everything we do and experience as human beings' - my question aimed at the disciplines that approach these questions mainly through text. It seems to me that historians do often think of their work as mainly text based, while archaeologists in general (do have to) approach a wider range of sources. As your workshop aims at a cross-disciplinary dialogue, would you say that History can learn something form the approach Archaeology takes - especially in the field of ICT -, or do you think the difference between both disciplines is overrated?
I have another question regarding how comparatively easy it now is to model 'some of the physical characteristics of space'. I have occasionally heard it said that this is actually a danger as people tend to believe that 3D representations of the past are not a scholarly interpretation, but rather an almost perfect, photorealistic recreation of the past.
Surprising as it may seem, this argument event came up during a recent Methods Network sponsored panel discussion entitled Talking CGI. Someone in the audience said that with CGI effects becoming indistinguishable from 'real' film we are about to loose the only medium that cannot lie to us (she was specifically referring to the written word as being very easily manipulated, but she also mentioned photographs as not trustworthy - and film as trustworthy). Does the audience really have to be reminded that all of our work is interpretation and if so, is this issue more important within the context of methods that use images and maybe even 3D models?
History vs. Archaeology and 3D models vs truth
To adopt the rather annoying habit of some politicians, and take your questions in reverse order...
The concept of interpretive reconstruction in archaeology is not new. It has been noted more than once that Sir Arthur Evans' concrete reconstructions of parts of the Palace of Minos at Knossos, erected in the first half of the twentieth century, bear a suspicious resemblance to Art Deco. It is largely a matter of transparency, and employing rigourous documentation protocols to ensure *interpretation* is not - whether deliberately (or even consciously) or not - objectified. As William Kilbride notes in his positioning paper for the workshop (http://wiki.esi.ac.uk/Standards_and_Metadata), 'cartography is exhaustingly political'.
I will let historians decide if they have something to learn from archaeology in this context. However, one activity which I think _is_ worth flagging is the important issue of 'grey literature', massive corpora of often unpublished texts which report on excavations. There are a number of projects underway to use text mining technology - in some cases very advanced text mining technology - to manage, search and interrogate these corpora. So, while there are obvious epistemic differences between history and archaeology, there are also concrete areas of crossover. I think both of these will be valuable areas for the workshop to look at.
Reconstruction and Reconnection
I'd agree with Stuart, and it is a tremendous advantage of ICT that we can experiment with reconstructions virtually these days, instead of smothering what remains of our past in interpretive icing. I should say, however, that I'm on the record as an advocate of 'ugly' archaeology. I'm greatly impressed by some of the achievements in 3D modelling these days, but I wonder whether the additional complexity required in order to give the right look-and-feel can muddy the waters when it comes to analysis. Fortunately, that's not (yet) an issue we frequently encounter in geospatial inquiry, but it may well be on its way. Software like TimeMap allows us to 'replay' spatial developments - a really powerful tool that brings a lot of very interesting challenges in its wake.
In reference to the other question, it's certainly true that historians' use of ICT tends to dwell on texts a lot whilst archaeologists tend to create databases of 'stuff'. It will be interesting to see whether that causes them to frame their spatial ideas differently. We might expect, for example, that the history folks will concentrate on locational references (reflecting where our ancestors considered themselves to be), whereas the archaeology crowd want to know where things were found (i.e. where they actually were). For that reason I'm really glad we've got Tom Elliott of the Pleiades project talking. In attempting to map the ancient world from both a literary and archaeological perspective, they have to reconnect those worldviews on a day-to-day basis.
geospatial text and outreach to the non-tech-minded
As a historian I do indeed think that History can indeed learn something from Archaeology in this context. It is not that historians do not use GIS, but judging from my last ten years in the field these methods are still seen as a rather specialised approach that the 'everyday historian' does not encounter that often. Nevertheless, there are useful applications - at a recent Methods Network Expert Seminar, Ian Gregory gave a presentation that referred to problems of 'time' and 'space' and did so in the context of talking about infant mortality trends in England and Wales from the 1850’s to the 1900’s - so this is definitely something historians should be interested in (here is the link to a recording of his presentation. (I figure that if Stuart reverts to methods of the political discourse I would the be allowed to answer part of my own question...)
Stuart, I found it interesting that you mentioned text mining: Next week I will attend the Methods Network sponsored workshop Text Mining for Historians, which might be interesting for members of your group too. Let me know if you have any question you would like me to pass on to the text miners... Is there a text mining project focused on 'grey literature' that already has some results available online? I would also be interested to learn whether there is a project that has combined geospatial methods with text mining, for instance to show how certain words or concepts spread through written discourse over time and space - this, I think, could be very useful. (again, here is the historian's preoccupation with text...)
Now, we could continue discussing possible interactions between ICT practitioners in the fields of history, but a posting to the Antiquist mailing list a couple of hours ago reminded me of the importance of another dialogue we should be interested in:
One thing that's concerned me a little of late is that we're getting increasingly good at sharing ideas and working with each other in the Digital Humanities community, but that there isn't as much dialogue with our non-tech-minded colleagues.
If you wanted to convince someone with a limited ICT exposure of the value of geospatial methods for archaeology, where would you point them?
Outreach
The problem of outreach to non-experts is a real problem, and certainly not one particular to this field. It's also a real shame because, in my experience, a lot of non-IT types are very interested in questions of data creation and sharing.
My feeling is that there are two major stumbling blocks at the moment. The first is the language and level of detail we use. Most people not only don't care about, e.g. XML, they're completely put off by the fact that you brought it into conversation because they can't fit it conceptually into the sorts of things they generally think about - it's just so much alphabet soup. What they do care about is how to make their information more easily accessible to other people (so long as it keeps their name on it, of course).
That brings me to the second problem which is that we tend to put the cart before the horse by saying things like "oh - what you really need is an Open Source MappingServer with a KML webfeed". Most people wouldn't have the faintest idea why that would be handy and so you lose them. Taking an approach along the lines of "those are really interesting results - it would be great to see how they look in Google Earth" is much more likely to get a positive reaction, even (especially!) when they ask 'what's Google Earth...?". It's only then, once they've grokked why that would be cool, that you can begin with the "so your data's in these 27 lever-arch folders, huh? Well let's start by..."
Everyone who works in our field has been on a personal journey of figuring out why some things work and are useful and why some things don't or aren't. To have a meaningful and engaging dialogue with others we need to help them make that journey as well, rather than presenting our conclusions as simply facts to be accepted.
I think maybe I got a little off-topic there, so in (brief!) answer to your question: there are lots of places online I could point them to, like Online-Archaeology or the Theban Mapping Project. But archaeologists on the whole don't need much convincing that geospatial methods are useful - the hard part is getting them to adopt techniques that makes their results usable.
Outreach
As Ian Anderson recently remarked in the Digital Historian forum:
As far as preaching to the unconverted I don't think there is a silver bullet. As far as the AHC-UK are concerned we rely as much on word of mouth as anything, with our members and committee spreading the word. If we can get information out in conference packs, posters up in departments etc. that will help, simply raising awareness of our . I suspect, however, that visiting every department/conference might be the only way to make real headway!
I think he might be right insofar as that actively going out to the 'unconverted' might be what is needed, especially as many of them might not even know what tools and methods are out there that could help them. The Methods Network has a series of working papers that address that topic (more to be published within the coming weeks), but I fear that a site like ours might already look too specialised (also, our remit within the AHRC ICT programme is primarily aimed at the specialists). What might be useful would be an easily accessible website that looks inviting to the non specialist and gives easy introductions that rouse interest and then points the people to more detailed resources. If this would also be used to coordinate the different "preachers'" initiatives then it might be helpful. There is a lot of knowledge already available on the ICT Guides website, but I am not sure how many of its users are 'unconverted'.
But you are right, we did get a little off-topic as your workshop is much more about a dialogue between practitioners in the field. What would you say is the biggest challenge in making these research results usable and what are you hoping to get out of the proposed 'domain-wide overview of the methodologies'?
Workshop outcomes
I think that what I'd like to see coming out of the workshop, and in particular the domain wide overview, is tied up with many of the things in the last couple of postings in this thread. As Leif points out in the context of the example of KML web streaming, many of the journeys people make towards the use of ICT in these areas are very personal, to solve particular problems interest to limited numbers of people. I would like some of these journeys to have a broader context, to look at how work done in different top level areas - we have chosen to think of these in terms of scale, heterogeneity and standards and metadata - can cross-fertilize. Now one is not going to do this exhaustively with a two-day workshop, but I think that six closely interlinked Methods Networked papers - i.e. our three keynotes and rapporteurs' responses - highlighting some of the commonalities and creative tensions identified in the discussion would be a good start.
Computing as a journey
I have to say I very much like the idea of the application of ICT as a personal journey and I would guess that this metaphor fits the experience of many practitioners within the (broader) field. There is an unfortunate side to this insofar as too many (still!) go that way alone and develop, or try to develop, solutions for problems that have already been solved - simply because they are not aware of what others are doing. I am aware that you are not necessary aiming at beginners in the field (and neither does the Methods Network), but creating resources that will help them to get to terms with geospatial computing is definitely useful. As would be the creation of a forum (and I do not mean this in a technical sense) to continue such an exchange beyond the event itself.
Getting back to the more specific questions regarding your event, I am especially interested in learning more about standards regarding 'data encoding and the interchange problem'. In the abstract for his paper, Vince Gaffney suggests not to develop 'yet another schema or protocol' to address this, but rather to use existing standards and tools in order to exchange data, especially across disciplines. Without anticipating the debate, which formats would you say have the greatest potential (Vince mentions Atom and others, but says that there is a danger of ''dumbing down' our data')? Do you think that portals and a focus on service based computing can be an answer to that?