Technological Revolutions and Art History, Part Three: Q&A
- I just wanna start off thanking all of our presenters, this was a really wonderful group of presentations and I was amazed and surprised by how much overlap all of your fields come together in terms of the questions that you're investigating, things being addressed in the digital heritage realm including like techniques, ownership, access, and I think the four presentations went so well together. Lu, do you wanna to kick off any questions that we might have and direct them? - Sure. I'll start with the first question that came in. This was during Alonzo's talk. This is from Andrew. So I'd like to push back a little bit on the example of high resolution images from the past now being relatively low res, modern cameras are able to achieve high enough resolutions that they're capturing all of the useful data available, getting, for example, the doc matrix of a small newspaper print.
There's arguably no increase, I'm assuming in information for capturing these in a higher resolution. Can you speak more as to if resolution might ever be thought of as sufficient in the longterm? - I guess for me, is resolution ever sufficient? It's hard to know. Clearly things we did 20 years ago we don't know today. I mean, Carla and Mark might have from the technical point of view, some thoughts on it but yeah, it depends on what you're capturing, if you're capturing newsprint and it's printed at a certain DPI, going beyond that DPI doesn't add much value, but if you're capturing faint brush strokes in a painting or watermarks or something, more resolution could be very valuable. And it's not just resolution, it's also spectra, it's also going beneath the surface capturing three dimensions, capturing x-ray.
One could go on and on. So I don't think we yet know everything we need to and data that the more data we have, the more useful it is but that's of course balanced with cost and time and those are limited. - If I can just jump in and add on to what Lon is saying because I think we're in large agreement, that part of the point of collecting information about what you did is that you, hopefully, if you're digitizing something you had a purpose in doing it and so helping other people understand what your goals were when you did it are gonna help them understand if that data works for them.
So it's all about fit for purpose. Is this gonna meet? I did something that met my needs, somebody else has to determine whether that meets their needs. And I think sometimes we also get caught up in the, it's not about how much information, is that good information, is it precise and what is it showing me? So the spectral example is a good one. And if we're talking about just digitizing something that is printed words on a page, getting the content readable, that's a pretty easy metric for people to understand but when you start talking about 3D objects, do I care about tool marks? Do I care about brush strokes? Do I care about what's underneath? All that kind of stuff can impact it.
So the main thing is when you go to digitize you should understand why you're doing it and who you're doing it for so that you can make the right trade-offs about which technique and what level of precision and resolution you need. And then you should record that so that other people have a better chance of being able to re-use your data. - Great. Thank you both. The next question is from Amy. This was, I think also directed primarily to Alonzo.
Amy would like to know if you could speak a little bit more about the Elgin Marbles. - I don't know if I, I often have a slide of them in my talk. I don't think they were in this talk, but sure happy to talk a little bit. Should we, I mean, that gets into complex politics but at the moment for those that know or don't know, the marbles were removed by Lord Elgin, taken to the British Museum. Athens has a new museum at the foot of the Acropolis which has what little was left on display there in preservation in the museum and then casts of many of the missing marbles or the marbles that are at the British Museum. And as with many instances of culture being taken or removed there's of course deep international debate and dispute over where they should be and why.
I'm not an expert I can't get into that but I can easily talk about what digital can do and perhaps we can more and more use digital to give access to those that have lost access. It certainly doesn't supplant having the real thing or even well, actually, even that is debatable. I think some in the art world, are having the same debate over is seeing a super high resolution digital version of a painting where you can zoom in and see brush strokes more valuable than being in a museum where you have to stand 10 feet away from the real object yet people want to see or be near to the real or what they believe is real. But in many cases that the caves of Lascaux are a similar sort of example of real versus replica.
You can't go in the real, the real Lascaux anymore because of the damage that visitation has done. So there was the first replica then there was a digital Lascaux three, Lascaux four, the one that's down the hill, sort of a half a mile, quarter mile down the hill from the real site is a huge visitor center and operation. And one goes in a full replica which is better lit, temperature controlled one can get a much better experience of the cave there than in the real cave. But of course it's not the real cave and it's an odd case where one actually travels all the way to a remote place, to the countryside to the real place but you actually don't visit the real Lascaux, you visit the replica just down the hill from the real.
So you're almost there, but you aren't quite there. So this idea of what is real and what not (phone ringing) I mean the Elgin Marbles of course are complex in that debate too. (indistinct chatter) - If I can, have just something very, very shortly.
So I used to work at the British Museum and began to do 3D scanning there along with somebody called Daniel Pett, who is now head of digital at the Fitzwilliam Museum in the UK. I just shared in the Chat, a link to Daniel's 3D scans of the Parthenon marbles that he published on his personal account, and this isn't in any way to say, here are the 3D scans, this somehow cancels out any repatriation claims. And I think it almost proves that if you have a digital copy of something it doesn't necessarily, it definitely doesn't in my opinion, replace the real thing or the actual conversation that people should be having about the real world objects. But if you want to look at them, download Daniel's data, there's a link in the Chat. - Thank you. The next question, Robert, do you wanna take the next one or? - Sure.
I think this one is for Toussaint. It's from Andrew. I saw on the map of digital basics in Africa that some countries were listed as discontinued.
Can you speak a little as to what happened there if local activists pushed back against the programs or was it Facebook ultimately deciding it was unfeasible or unprofitable? - Sure. I mean, it actually depends on the different countries. In some cases, for instance in Egypt it actually, it was sort of Facebook pulled the plug because the Egyptian government tried to use the project to conduct surveillance and politically target surveillance. In Zimbabwe which was supposed to launch it, there was the Fall of Robert Mugabe and a change of ownership in the telecom company which was partnering with Facebook and that put a brake on it. But in none of the countries really we've seen, I mean, we've seen the kind of mobilization that happened in India and that was really quite striking.
There are groups, a lot of groups on the continent that were complaining about digital coding and even when talking about that particular initiative. But what happened in India was amazing that they managed to create a sort of a national debate around that question, bringing in questions of sovereignty, bringing question of justice in the development of digital infrastructures at the national level and so that's something that actually took a lot of work from activists, a lot of coordination. And so yeah, to your question there, it depends on which country and I'll put the link to the article which is open access and it talks about some of the specific countries where it was discontinued. - Thank you. Our next question is a little bit more general, it comes from Michael. Most of the talks seemed focused on collecting and democratization of data. This is great.
Could you also talk about how this data can be used to go deeper into the history? How can this data be used for academic research? And how close is it to enabling academic research for scholars who do not have the ability or financial stability to travel to see this art? - Well, I think that one of the main points that enables scholarship and academic acceptance of these kinds of digital representations is having enough context and metadata regarding to the process that was applied to the digital representation to create it. And if you have a good step by step account of what went on, an academic researcher can read the account and then make an individual decision, is this representation suitable for the re-use purpose that I have in mind? Right? So it's not a generalized answer to types of stuff, it's whether or not there's sufficient context and process history for someone to be able to evaluate its quality and look at the precision of where the points are in space, because it's not resolution, number of points in space for 3D, for example it's that the points are in the right spot, right? So if you look at all these things and that information is transparently associated with the digital representation, then I think academic use is empowered and use for any other purpose is equally empowered. - And I think if I can just tag onto that. One of the things that we've been teaching and pushing people to do for years is to say that, and all of our work is in computational photography, that that original photo set is more valuable than any output, 3D model or RTI or whatever you might make because anybody can rebuild it from that. So if you don't understand exactly what somebody did and maybe they filled holes, and maybe they edited it by hand, and maybe they made it look good where they didn't really have real data, you could go back to the original image set to see what's actually in the empirical original data.
So it's not just the process history, it's also having access to the original empirical data which in our case is photo image sets. So we think that as people start understanding that and moving more towards that then the ability for people to use these kinds of materials in research is radically increased because you understand what you have. - Right. And one small addition is that all of the information and computational photography technologies is contained in the original photographic data sets and we know how to preserve photos.
The repository preservation of photos is well-established. So if you're trying to prevent digital dark age annihilation of this data, if you have it in photographic form, it's got an age up and being able to sustain itself going forward. - Great. Thank you. The next question is from Hannah Phillips, and it looks like Tom Flynn had to hop off.
Alonzo spoke so interestingly about the pitfalls of previous digitization efforts in technology that became quickly outdated. I'd love to hear about the presenters' thoughts on the storage of heritage preservation files. My company is investigating heavily in cloud storage but no tech is future-proof and there's concerns about the longevity of this relatively new option as well as the security of storing these precious files on the server of large corporations, example, Amazon. - I guess I could start. It's tricky and in my talk, one of the reasons I emphasize the idea of releasing copyrights, using copy left, using creative commons is the hope that one, instead of you having the single copy of an image or a 3D model, or a published work, or a painting, but by making it freely accessible to others there will be digital replicas floating around, not just on Amazon's web server, or Google's, or Apple's, or whoever's but on many.
And we've lived for several hundred years in the world of copyright and ownership and rights and it takes a lot for folks to get used to the idea of opening up their data but I think if we start to do that that can perhaps solve. One could argue that Amazon is safe because it's got replicas and redundancy around the world but as with any technology as you note, it could fail. So I think our safest solution is to learn how to share and make our data accessible and put enough metadata on it so that someone else can use it and can continue to repurpose it.
That's the way to make your data last, otherwise we're better off with the stones and the originals. - Yeah. I'd also like to just briefly add that it matters how the data is stored.
If you store data in open formats, right? The longevity and the ability of digital curators to go in and work on that material is enhanced. If you store it in proprietary formats, you're almost guaranteed that at some point that's going to become obsolete and the Smithsonian guidelines or Library of Congress, thank you, guidelines for preservation state only non-proprietary open formats should be used. So that's a big part of it is to translate your data into open formats and also keep the data archive separately from whatever its preservation layer is, right? Because if it's put into a particular type of presentation format, that those go obsolete incredibly quickly, and so if your data is locked into something like that, you're gonna lose your data or it will become unusable quickly. So if your data is separated from how you present it that also adds a lot to its preservation and the key is just to have multiple copies, right? Because any one of these corporate things can be purchased by somebody else and in an instant a money decision is made and it's gone. - Great. Thank you.
(murmurs) Where were we? Question just disappeared. - Sorry. I tried to respond to one of the- - Oh, sure.
- I haven't somehow moved to the answered... - That's all right. Mind if I just, I'll just address it to you then.
Toussaint, thank you for increasing our awareness of the divides that exist in the democracy of data. I'm curious if the group has considered what is an optimal path forward? Obviously, corporations drive a lot of technical revolutions. So how can we begin to be more democratic with data if we were to remove corporate influence and implementation on and in the field of data, which includes imaging? - Yeah. Thank you. I'm gonna plug here a recent event that we organized at our lab called Digital. It was about reclaiming digital infrastructure for the public interest and we actually had a few proposals that were put forward by a few thinkers. Ethan Zuckerman from the University of Massachusetts Amherst who has this proposal to build, building off on some of the public broadcasting methods that have been used for instance, development of the BBC and to apply them to digital contexts.
Todd Davies, who talks about the idea of a corporation for public software. So I'm gonna draw up in the Chat box a link to a write-up of that event and in there you'll find actually a range of possibilities that are just waiting to be funded (chuckles) and to be put money into and that a lot of that comes down to political will and that political will and money. So I'll, yeah, I'll just put the link and I think that gives, that opens up the field of possibility and I think that's really what we need to have in mind all the time.
- I think just to add to that great answer. Not all of data is being stored in big corporate repositories, right? Most universities have repositories of data, the internet archive is out there and interested in collecting and storing data. So people, again, it goes back to that multiple copies but they don't all have to be in the control of Google and Amazon and pick your favorite (laughs) one. So I think that that's part of the education process of putting tools in people's hands, making them aware of what's possible, training them to use it, and then also providing places for them to put things whether that's their own repositories or repositories that they can add their data to.
- Yes. And one thing that we're working on and trying to get some funding for is to develop turnkey ways of installing repositories for cultural communities. For instance, there's a software repository package called Fedora which is an open community development, it's all open source.
And if you had the digital lab notebook creating submission information packets, if you install the properly configured Fedora, suddenly you'd have your own repository. And of cultural communities around the world, each had their own repositories with some material linked to the rest of the world and accessible, this is gonna decentralize the storage and accessibility of cultural materials and the fear that it's all gonna be dominated by Google or Facebook, God help our souls is reduced. - Yeah. And I think the thing that we see over and over is we're talking to indigenous communities and other less resourced and underserved communities is that part of the control that they want is not just over telling the stories and the narrative and what gets digitized but also in what gets shared at all. And they have legitimate reasons for some of the material to be digitized for their own preservation, protection, sharing within their own communities that they do not want to share with the rest of the world and so they're very, very reticent to put that kind of material into repositories that are run by other people. So they need the ability, but therefore there's lots of stuff sitting around on hard drives that's not gonna have much longevity, right? It needs to have a better solution than that.
So there's some give and take and that has to happen here and sort of ways that we can help people understand how to manage their data and hopefully be able to give them ways, as Mark said, to install their own repository software and think about repositories, not just, I have a bunch of files sitting on a hard drive somewhere that's gonna fail in the next two years. - Yes. And if the material is prepared in the form of linked open data then, and it's given permissions to be accessed and go out and look for similar bits of information, this material potentially can link up and so that you have individual repositories with a small portion of the overall story but the data is enabled in each of these repositories to be found and linked together so that you can follow Ariadne's thread from one piece of information to another along the relationships that link this information together.
So the structure of the information being in a linked data form, hopefully structured with some form of intelligible conceptual model makes all this much more possible. - Just adding to that. Speaking of linked data, one of the challenges I continue to see is institutions, libraries, academics continue to publish and they write papers that often link to heritage and valuable data, but typically papers because of the way journals and other things work, museum catalogs and so forth, have just an image of whatever the object was and they don't actually link back to the original object which hurts the whole idea of keeping the chain of linkage going so that the data is preserved. So often what's happens, I mean, you look at art museums and you have old catalogs but the links to whatever the object, if the object was lost in a fire it's moved from one collector to another, the linkage is gone. So that's something else we have to change.
It's not easily done, just like linked data and all these other things. We've been talking about it for 20 or 30 years and we've made progress but we still have quite a ways to go. - Yes, in the process of transforming digital representation and their metadata into linked data is very complex. So to the degree that you can automate this, right? In an accurate way that reflects the original ordinary language meanings of the stuff then that empowers the creation of linked data because what the digital lab notebook does is something that only a thousand people in the world could sit down and do themselves and nobody's gonna go through the process to learn how to do it and certainly not gonna spend the hours it takes to formally do this after they've documented something. So it has to be automated so it just happens for the user and it's part of the whole scientific workflow that's going on if you're gonna have this actually happen. - Great. I think we're gonna have to wrap it there.
We've already gone over by about half an hour and there's more questions which means it's been a great discussion. And I just wanna thank everyone once again, all of our presenters, all of my colleagues at The Frick. This has been recorded and at some point it will be put up online for everyone to view and share. So with that, just think about March 14th for the fourth and final session and everyone have a good afternoon, good evening, wherever you're calling in from, and thanks again.