Margaret does something that is of uber importance for all of us: she builds and thinks through the technology of describing, storing and retrieving images. Her tool ImageSnippets™, created for anyone who archives, curates or shares images on the Web (who doesn’t these days) was born out of her passion and scientific problem: the semantic annotation of images, or in other words, the proper preservation of digital memories.
An artist and technologist, or as I called her in one of our emails, an artistic soul, versed in RDF, Margaret finds her most satisfying art making occurring when she has achieved some kind of critical balance of ’empty’ mind while intuitively narrating herself through conceptual landscapes. In-between imagining that our connected information auras make up abstract constructs of an invisible ‘matrix’ if you will – of consciousness, Margaret actually does some very practical work towards a semantic web of images.
I am privileged to have her art & tech soul beautiful ripples around the questions I asked.
In this Dialogue we talk about metadata, psychedelics, formal descriptions, information “auras”.
After having this Dialogue with her, I am not worried about the image of aunt Sally anymore*. And I hope you won’t be too. Here’s why:
This is an example of an archive Margaret gave … and it is so simple in its complexity! Just have a look and enjoy a beautifully archived photo, described with RDF (in Plain English with data that explains which thing is related to which other thing and what that relation is).
* ref: In his book “Everything is Miscellaneous – the power of the new social order”, David Weinberger writes: “When you have ten, twenty, or thirty thousand photos on your computer, storing a photo of Aunt Sally labeled “DSC00165.jpg” is functionally the same as throwing it out, because you’ll never find it again.” [cit. Everything is Miscellaneous – the power of the new social order, p. 13]
Enter Margaret Warren and her world – a gorgeous dynamics of art and technology.
Margaret, first things first, why metadata rocks? :)
First – I just want to say Teodora, that your questions are wonderful! They encouraged me to be playfully philosophical and even ‘artistic’ while answering – even though my work with ImageSnippets often has to border on the much more practical aspects of the semantic web. So thank you for asking such thought-provoking questions that let me play with these interesting ideas!
Metadata does ‘rock’ doesn’t it? Truthfully, I didn’t put that much thought into it. I just needed an extra domain for some internal business and a mail server. I thought it would be cool to use the word metadata somehow and noticed the upper level .rocks domain. I haven’t actually used it as much as I plan to yet.
If you are familiar with Jason Scott from the Internet Archive, he once (now famously) said; ‘Metadata is a love note to the future.’ which has been widely quoted all over the internet. I go one step further and say, “if metadata is a love note, then writing metadata with linked data should be a serenade”.
A research scientist I know, David Fries of Florida IHMC has coined a wonderful phrase called, the “information aura.” It basically means that everything we can talk about, look at or describe – especially in the physical world, can have an information aura around it, i.e. it’s metadata.
The difference with the information aura, is that the metadata doesn’t just stop at the ‘who, what, why, when and where.’ Take a tree for example – not only could you say it was a maple tree, about 18 years old with a specific GPS location, but you could also describe its biological processes, people who have touched it, people who have been married under it, birds who have nested in it, insects that feed on it or in it, fungi that live on it, the color of its leaves during all of the seasons of its life, poetry written about it and really anything else. This is all metadata. The who, what, when, where, how of anything and everything – and it describes life…all life is metadata and we can have a meta-awareness of this.
Going one step further, we can imagine that it is our connected information auras that make up these abstract constructs of an invisible ‘matrix’ if you will – of consciousness.
I like to imagine that what I am doing in some small way – by building ‘linked data image snippets’ as part of the ‘semantic web’ – is like weaving some of these information auras together. Aiming toward some imagined singularity of ‘intertextuality’ (or rather inter-link-uality).
How did you end up navigating the Semantic Web ocean? What lead you to ImageSnippets?
I think it might be easiest for me to combine these questions as each answer leads directly to the next.
I came to the semantic web from anything but a traditional path, in a very funny way. I was somehow guided by chance, fate, curiosity, or a tenacity to solve a problem (depending on your philosophical flavor.)
In some ways, I think I am very lucky I did not take a strictly academic journey, because I can react to exploring this field with a freshness not entirely burdened by so much rigorous theory. But on other days, I feel like I am playing catch up to get the kind of theoretical background a more traditional path of study would have given me.
But here is how it happened:
I grew up with many artists in my family; it’s in my blood. Essential for my soul. My psyche. My father had studied astrophysics, but then became an award winning photojournalist. As a kid, I thought I might become a writer one day because both of my parents were such talented visual artists that I never thought I would achieve their level of talent and I needed to find my ‘own’ thing. But I created art and photography from a really young age, I often accompanied Dad on his assignments and learned to print photos in the darkroom at around 7 or 8. But I also often paired up Dad’s photos with poetry and wrote extensive metadata all over the back of my prints.
I had also had an exposure to the study of electronics and found it fascinating. When I was quite young, I joined the Coast Guard and became a cryptographic electronics technician. I was programming 8 bit microprocessors in the early days of computing and I loved technology, especially systems theory. After the Coast Guard, I went on to pursue computer science academically. I studied programming, math and computer language theory, loving all of it. In 1988, I fantasized that one day I might study in the Symbolic Systems program being offered at Stanford. I even made a copy of the program from the catalog. Money and other choices made me take a detour down a different path and though I knew I wasn’t going to be able to go to Stanford, I thought perhaps I could pursue my own path of self-study and make up a similar interdisciplinary program at another college later. [I even found the xerox copy I made of the program in 1988 in my archives, the front page of which can now be found in ImageSnippets].
For many years thereafter, I worked as a hardware tech, programmer, systems analyst, network engineer and ran my own computer consulting company. I was working with multimedia companies in the early 90’s. I was the geek at the office who first got Mosaic running on the ‘internet’ just a few days after it was generally available in 1993. I was also doing tech work and sound reinforcement for composers/guitarists/musicians and though I am not a musician myself, I built a lot of systems for musicians – analog to digital conversion systems for audio recording, midi systems, composition software and just generally learning all about aural and musical symbolic systems. I continued to take courses here and there along the way, read as many books as time allowed on early AI and played around with programming languages like Forth and Smalltalk. My passion was ALWAYS to learn more about cognitive science generally and of course I never stopped making art.
At some point, in the early, early days of the web, I was building web galleries of my art work (I think some of them can still be found in the WayBack Machine!). And here is where my problem started. I was totally vexed by how I was going to handle the metadata with the images and how almost meaningless I found keywords. Of course I could use all of the technology available to me in writing html files and using databases, DAM’s, etc. But in my AI predisposed mind, I couldn’t get to what I needed. I had a lot of requirements in my mind: disambiguating the keywords, relating the words to the images, having the persistent metadata stored and published properly with the image. I didn’t quite know how to verbalize it properly yet, but all throughout those years, I kept adapting to various strategies as they were emerging – but nothing had gotten it ‘right’. I did know that my problem was about linking DATA, not text. But in fact nothing in self-publishing had gotten it completely right until I did it myself with ImageSnippets – but that was a long way off.
In the late 90’s and early 2000’s I was still trying to solve the problem. But everything was about to change! Sometime late in 2001, something crazy happened. The guitarist I mixed sound for got a call to sub for another guitarist who had cancelled a gig at a BBQ party in Pensacola, Florida. We drove to the address and the sign outside the building read: Florida Institute for Human and Machine Cognition. I was FLOORED. I had no idea a research facility like this existed in Northwest Florida. I went home that night with this feeling that somehow just knowing this place existed was going to change my life. And this, is essentially, how I arrived at the semantic web ocean.
So, I was at this BBQ party hanging out at the mixing board and by chance end up talking with Pat Hayes. Pat Hayes is a prolific and well-known figure in the semantic web world (but I didn’t know that then.) We struck up a conversation about art and cognitive science. Pat happens to also be an artist and of course I always wanted to be a cognitive scientist. I thought that was the extent of it, but then a few months later, I run into him again at a local art gallery (again – quite by chance) and I invited him to come to a local drawing group in Pensacola. At one point, I recall discussing how much I had wanted to study symbolic systems at Stanford and he remarked he had been at CSLI (Center for Study of Language and Information) during those same years. That’s when I went home and found that Stanford program and realized, in fact, that Pat had been one of the Consulting Faculty members of the very program I had wanted to study!
Casually at drawing groups and art events over the next few years, Pat and I would discuss my problem with publishing my web galleries and my interest in formalizing image descriptions. I guess it was around 2003-2004 when Pat was then working on authoring the RDF semantics and I gradually began to realize how RDF could solve my problems.
All these years later, I still can’t get over the great irony that I would, by chance(?) have the opportunity to befriend a research scientist, with a world renowned background in knowledge representation and common sense reasoning who would also happen to be an artist and somewhat uniquely qualified to lead me into this maze of RDF and the semantic web.
A few years of discussions later and along with several other researchers at IHMC, Pat and I started our first project that involved myself as a subject matter expert on my own art work using CMap Tools (Concept Mapping tools). Some really interesting work came out of this, including a paper I co-authored in 2006 with Pat, Tom Eskridge and Robert Hoffman: Formalizing the Informal: A Confluence of Concept Mapping and the Semantic Web .
Several years later, in 2010, I decided to build a prototype and start the development work on what has now become ImageSnippets. I just wanted to see if it could work! By then I was just as interested in the knowledge representation aspect of the project and wanted to do it as much for pure research as anything else. But since I was spending my own real money on it, I figured I might as well also create something that would help me get back to solving my original problem of sharing and publishing the images on the web with linked data.
I hired a few programmers to help me start the actual development, and we all still continue to work together. In more recent years, I have also hired and trained small teams of annotators in the use of the system and constructed use cases to test, measure and gather metrics on various aspects of the user interface. Just recently I have started to build and test website galleries created with a SPARQL query based WordPress plugin. It’s a team effort with the developers and I communicating frequently on what I would like to be able to do, whether or not we can do it and how we should approach it. It’s a daily iteration of my own experimentation, research of practical RDF publishing considerations and the observation of the use of the UI by annotators with different levels of experience. As developers say, we eat our own dog food.
Pat and I created the lightweight ontology used for the main image description relations and he continues to consult on ontology issues particularly. Increasingly, these days we view ImageSnippets as much as a research platform for ontology engineering as for any other publishing function. We have also now co-authored several other papers on LIO and the system, and I did actually make it to Stanford after all! :-) when I presented our initial work on our Lightweight Ontology at the 2010 AAAI Spring Symposium.
In June 2018, we wrote Bounding Ambiguity: Experiences with an Image Annotation System which Pat and I both feel is our best paper to date on the background of ImageSnippets and the methodology behind LIO. I presented this paper at a workshop on Ambiguity, Subjectivity and Disagreement co-located with the July 2018 Human Computation Conference.
My wordpress website (http://www.margaretwarren.us) (and my father’s: http://www.earlwarrenjr.com) website now have galleries populated with images annotated exactly the way I had envisioned almost 20 years ago. :-) The plugin uses SPARQL queries to pull the images from ImageSnippets and expresses the RDF as JSON-LD.
In his book “Everything is Miscellaneous – the power of the new social order”, David Weinberger writes: When you have ten, twenty, or thirty thousand photos on your computer, storing a photo of Aunt Sally labeled “DSC00165.jpg” is functionally the same as throwing it out, because you’ll never find it again.
How do you think it is best for everyone to approach their digital memories?
Well, of course we now know that since this book was written, major advances in facial recognition and cloud storage mean that fortunately (or maybe unfortunately) Aunt Sally isn’t as easily discarded as she used to be (assuming someone uses these technologies) – even though any storage medium is fallible.
I think it is interesting that you asked specifically about our ‘digital memories’ – and not just our image memories. It seems the organization of our experienced and recorded (visual/aural/experiential) memories continues to get blurrier every day. Who would have ever thought, “I’m going to go make some photos today, let me make sure I have my ‘phone’ with me.”
I am a storyteller, an artist, a curator,a metadata-ist and I have a purpose built ImageSnippets, that supports my very particular workflow for the distribution of images in the semantic web and for preserving digital memories. I do believe that it is RDF and the semantic web technologies that hold the greatest promise for continued accessibility of digital assets with their accompanying metadata moving forward into the future. But I am really not sure I have a definite ‘posture’ on what is the ‘best way’ for everyone to preserve their digital memories. Everyone is different and I encounter many people, particularly from younger generations who simply aren’t that interested in their digital memories.
I used to say that one day (if we were lucky) we would all end up in a box of photos in a flea market with very little metadata left to mark our existence – if Aunt Sally was lucky, she might have gotten her first name written on the back of the photo. Now, many of us – particularly those of us who are older – have storage media sitting around (like myself) that can no longer be accessed without a huge effort. Syquest and Zip drives with no SCSI or parallel ports.
Add to that the fact that I am a die-hard film photographer and maker of visual media IRL so to speak, as were my paternal grandmother, mother, stepmother, father, maternal grandfather and stepmother’s father (the last three of whom were all professional photojournalists by trade). Our family has had boxes and boxes and boxes and boxes of slides, negatives and prints – full of REALLY good work and iconic imagery. Lots of us are recorded memory hoarders, but only some of us, archivists.
And despite ALL of the millions of images I have – especially of my father – digital and otherwise, it is the ONE ‘digital’ image memory that I don’t have that I appreciate the most.
In the last days before my father passed away, I made, perhaps the last photo I ever made of him using my cell phone. It was a photo of him signing a Blurb photo book I had made for him of one of his photo essays. A few days later, for some crazy reason, a glitch occurred on the phone and despite all of my technical knowledge, I could not recover that one file. I spent hours with it, attempting every file recovery strategy I could with no luck (yet). I am still occasionally obsessed with wanting to try to get that ONE image off of that phone. To heck with impermanence!
Lately, I have been keenly aware of a pressing need people seem to have to register their time and place in the world despite our impermanence and despite the fact that they then turn around and treat that same reference (the image) ephemerally after its use in social media has served its purpose. The sharing of the images for others on social media affixes some historical mark that, once it’s been done, ceases to need to be a ‘memory’ it would seem. I have read other theories on this behavior. That is it is our current narcissistic self-absorbed culture that leads us to do this; but I think it goes beyond that. I think it has more to do with the speed at which information bombards us and a deep need we have to register our existence even while knowing how tenuous our threads on this earth are.
I went to the Louvre earlier this year and I made my obligatory pilgrimage to the Mona Lisa. All 400 or more of us who were there in that room in that same 30 minute window surged forward toward the painting slowly, waiting our turn, phones and cameras in hand, and I overheard this conversation between a (presumed) husband and wife:
wife: ‘aren’t you going to take a picture with the Mona Lisa’?
husband: (not visibly holding a phone or camera): no – why? Do I need to?”
wife: But, because you have to!
husband: Why? everyone here is getting a picture of it – there are hundreds of photos of this painting on line – everyone here is getting their picture with the Mona Lisa, why do I need one?
wife: because you need your own – so you can say you were here
husband: but I know I was here, what difference does it make to have a photo – I won’t look at it again!
and so the conversation continued like that for another few minutes and I smiled at the philosophical debate, understood his point, wondered at exactly how many photos of the Mona Lisa are uploaded to the cloud each minute and STILL felt compelled to make my own photos. A taller than me person behind me offered to take my selfie since I don’t travel with a selfie stick. Did I post it on social media? No. Have I looked at it since that day? Nope. I knew I was there. Will I put it in ImageSnippets? – maybe now :-) but only because now I have written this story and I could use the image to accompany it.
How about our collective memory – thoughts?
One day, I am going to find a way, semantically, to accumulate a collection of all of the books, stories and films in the world in which a central plot device is around a tethered telephone line. Perhaps the central character needed to find a phone, a telephone booth (or a quarter as in “Dr. Strangelove”), or answer a tethered ringing phone; and whether or not they could answer the phone or not changed the storyline.
Already, these kinds of movies would make almost no sense to those born digital with a cell phone in their hand – unless the phone falls in a river or something.
But it is my own personal sort of semantic challenge – to observe whether or not the world can get that kind of meaning out of our recorded media formalized in such a way that it can be discoverable before tethered phones have been completely forgotten out of our collective memory. I have casually asked this question before on FB, so if anyone knows of a way to get this information, please let me know!
In general, though, there is no doubt of the relationship between collective memory related projects and ImageSnippets. It’s my great desire to find more and more opportunities this year to bring interesting use cases into the system and work with subject matter experts describing images that can be used to formalize knowledge that will carry forth into the future as RDF.
I was in a hotel in Manitowoc, Wisconsin once, where the proprietor had his own collection of photos on display in the common areas. He had these beautiful photos of the fishing and boating industries in the Great Lakes just after the turn of the 20th century. In these images, were all kinds of tools, implements and activities of the fishermen on the docks and on the boats.They were not in a museum, but just a family collection by some one who was personally motivated to preserve the images and this knowledge. There are a lot of these kinds of photographic collections out there that I would love to digitize and preserve with their metadata!
So far, I have spent some time digitizing Porsche related images from Porsche restorer, Gary Kempton’s estate, including them in ImageSnippets and working with the subject matter experts (the employees) who know how to precisely define the parts and work in the images. We also have a few other ongoing projects including a textile museum in Central Florida who has a curated collection of Japanese propaganda kimonos and we have downloaded and triple-tagged many images from the public domain from the Yahoo/Flickr collection and images in the Flickr Commons as part of our research. We currently have almost 60,000 ‘good’ triple tags in the system for these images. ‘Good’ triple-tags are hand curated with our relations as well as some schema.org relations and fully disambiguated with linked data entities from: DBpedia, Wikidata, Yago and Getty’s Art and Architecture Thesaurus.
As an artist and technologist very close to both the analytical and the synthesis part of creation, what do you see in the empty space between envisioning and embedding metadata in a given work?
In general, it can be said, that our minds are always naming and labeling things, constantly describing and defining our experience; judging it and putting it into categories that have meaning for us. Our human minds attempt to connect to each other (at least in part) through these symbolic systems of words and I believe that because we interrelate to ourselves and each other through language(s), we use these labels in our thinking to continually ground ourselves in our ordinary experience – in our experience of something we call ‘reality’.
Zen practice directs the meditator to free your mind of all preconceived notions, to recognize and release the incessant naming and judging. To find the ‘empty space’ with no labels. And in fact, in this way, naming and labeling during the art creation process (for me) is a huge distraction, so I find that making art is often an exercise in being mindfully and painfully aware of the ever chattering monkey mind. I have always believed that music is the superior art form – because of the immediacy with which music communicates – often without words at all.
And yet there is also something very satisfying in having a narrative to support the work. Even if the story of that work is interpreted after the fact.
When I go to a museum and experience visual art, I usually get more from the work by reading something about the artist and the work on the placard next to it. It is these narratives that I believe we can formalize using semantic web techniques.
My most satisfying art making usually occurs when I have achieved some kind of critical balance of ’empty’ mind while intuitively narrating myself through conceptual landscapes.
One of my favorite pieces that communicates some of these ideas quite well is called “Surfin’ with Psyche” : http://imgsnp.co/qddg3
I am often inspired by Greek mythology, Joseph Campbell, Jungian archetypes, and the mysterious links between the mind and our human mythology in general. I love the multi-layered symbolism.
When I made this photograph, I was using a layered in-camera photo technique in which I place objects, transparencies and multiple malleable materials on different levels of glass with spotlights multi-positioned to play with the illumination.
At the time I was making this art, I had been reading some Greek mythology, but the day I physically made this piece, I was not intentionally ‘thinking’ of making a piece about Psyche. The word, Psyche, in Greek translates to “soul” and “breath of life”. It allegorically inspired ideas about psychology and consciousness, and in Jungian terms, she is said to represent the concept of the psychology of the feminine.
I was not ‘thinking’ of any of these things when I was making the art, but instead, I was simply ‘playing’ with color, and light and objects like jewelry, an egg, a test tube and a few transparencies. I created quite a few pieces of art work during that session. But mostly, I was trying not to ‘think’ at all. Creators call this a ‘flow’ state. Miles Davis said: ‘learn everything and then when you are playing, forget that you know it” (I am paraphrasing.)
But while I was creating, I became aware of the shapes I was creating while ‘painting’ with the egg that was beginning to cook a little under the hot lights. Out of these photographs, emerged what I saw as a female torso sitting astride a virtual surfboard of sorts with this kind of enormous brain looking shape attached to her. So I began to work with that concept and kept experimenting, following my curiosity. At some point, I felt the specks of light and bubbles in a deep purple color behind the female shape made me feel as if she was ‘surfing’ through deep space; a cosmic consciousness if you will.
Now – can I guarantee that everyone will grasp the meanings I intended as the creator of this piece? No. Will formalizing all of this narrative help? Maybe? Is it possible to forge this kind of meaning into a ‘semantic web’? This is yet to be discovered. So far, there is no way to really pin down these kind of subtle differences in meaning into a machine readable language. So far. Describing images like this in ImageSnippets is still quite challenging; yet for now, I feel quite satisfied that I can say that the image:
usesPictorially an egg yolk, a test tube, and a brooch and that those shapes lookLike a female torso who conveys the idea of ‘Psyche’ and conveys the idea of deep space.
Your project Porsche Art Car is very much related to psychedelics. In a talk (I wish I could do a semantic search over Youtube and find that link) Terence Mckenna was talking about the self and its atopical nature – the self has no one locus, it is everywhere. If we think through metadata from that perspective, what are your first thoughts?
I think it has probably come as a complete surprise to some people, after seeing my art – particularly the art car – that I have never done psychedelics. I may have had an experience or two with some powerful marijuana as a teenager, but no mushrooms, no ayahuasca, peyote or LSD. My former partner, Gary, used to say that some of the art I make – like the kind I painted on the art car – would drip out of my ears at night while I slept. I have always loved that image – seeing a visual stream of my consciousness pouring right out of my dream state all over the pillow. No hallucinogens necessary. I don’t actually even enjoy getting high. But, come to think of it – quite a few people who have ended up acting as kind of ‘muses’ for my art, have been people who got high a lot more than I did. Maybe one of the roles in my journey has been to be an ‘interpreter’ or translator of sorts, of the psychedelic states in others, ha ha ha.
I do say that the type of painting I did on my art car was inspired by the psychedelic era and it was.
It was my mother, who DID use LSD, and whose neon dayglo painted rear windows in a VW station wagon which she did when I was about 8 years old, that inspired me to paint the art car. Maybe the psychedelic experience was just transferred to me epigenetically, ha ha. I say this mostly as a joke, because she wasn’t using any drugs at 16 when she conceived, carried and gave birth to me. She died when she was 52, over 20 years ago,. She had used a lot of LSD and she also had schizophrenia related issues that I believe were no doubt compounded by her repeated use of psychedelics. I am sure this is a big part of why I have never wanted to do them.
But when I thought of McKenna and this idea of the self having no locus, and my mother, and LSD and a myriad of other persons – Robert Pirsig, Carlos Castaneda, etc. related to psychedelics, schizophrenia, spiritual medicine and the use of psychedelics to treat mental illness; my mind went to R.D. Laing somehow. Laing was a Scottish psychoanalyst and social phenomenologist (who rejected psychiatry and the psychiatric label) and was inspired by existential philosophy. Laing wrote a pair of influential books in the early 1960’s titled: “The Divided Self”, and “Self and Others”. Since I had a mother who was always bordering on some kind of fragmented reality that sort of made sense, but at the same time didn’t make sense; an exposure to Laing’s ideas was extremely useful and from it along with similar observations from others, I latched onto this notion that normal ‘c’ (normalcy) really just falls somewhere between high c and low c on a completely relative scale, not necessarily grounded to any kind of conformity.
Laing’s observations of the treatments for people suffering from schizophrenia went counter to standard treatments of the day – which included electroshock therapy – and other more violent means meant to jolt a person back to a ‘shared reality’. Perhaps existential philosophy, psychedelics and the web have illuminated and improved at least ‘some’ aspects of the violence related to cultural conformity. Now if we could just solve violence related to political and religious conformity…or violence period.
But what is a ‘shared reality’? From a more zen perspective – with or without psychedelics or schizophrenia – one can experience a sense of our detached ‘self’ as a function within a larger organically evolving system. From my own perspective, I think you can get the same kind of thing as McKenna by reading Buckminster Fuller, ‘Intuition’, a not so much scientific – but free-verse poetry book by a scientist in which he intuits about synergetically, mass-interattracted, omni inter-accommodating energy event complexes.
From that vantage point – our self’s ‘information aura’ is projecting its experience into, across, merged, intertwined and interconnected with every other aura as part of a whole that is yet again part of another whole, ad infinitum.
How does that relate to machine-readable metadata? I guess I am just bold enough to believe that the ontology work I am doing with Pat Hayes might just be a little like programming a small aspect of the ‘shared reality’ of the ‘uninformed child mind’ of the semantic web, but maybe we should drop some acid first?
What do you think are the four elemental forces of technology?
Well, my first thought is they are the same elemental forces that have always ever existed – and they are elemental regardless if you are talking about technology or anything else on this planet: earth, air, fire, water….then I had some funnier notions about the four fundamental interactions: gravitational, electromagnetic, and strong and weak nuclear interactions.
And then I decided: Google, Amazon, Facebook and Twitter might be the better answer.
Tell us more about the “radical departure from traditional restoration” and your Porsche adventures…
My porsche adventures started in 2002 when I met Gary Kempton, a very well known, world class restorer of vintage Porsches 356’s and early 911’s. Gary’s company, GK Restorations, was one of my computer clients. In late 2006, Gary and I began dating and we were talking about getting married. Sadly, he passed away unexpectedly in June 2009 from a heart condition. I continued working with GK Restorations and Gary’s estate in order to help Gary’s mother (he was an only child) and son in the aftermath of his passing with the probate process, work with the clients and employees of GK Restorations and liquidate Gary’s unusual collection of old boats, planes, trains and automobiles. I had a vertical learning curve about the Porsche restoration business like climbing a rock wall, but during this past 10 year odyssey I have had a lot of fun with Porsches – and lots of stories about the world of these collectible cars. Of course DRIVING the cars is a blast and there are a lot of reasons why people are fascinated with them.
I painted ‘Hatha Petey’ in 2007-2008 when Gary was still alive and we had a lot of fun with it. The description about the car on this website is pretty thorough: http://www.margaretwarren.us/2014/04/29/porsche-art-car
I call it “Hatha Petey’ because it was once owned by a man named ‘Pete’ and because I painted on its front – the morning with a rising sun and on the rear, night time with a full moon.
In Sanskrit, ‘Ha’ means Sun and ‘Ta’ means moon and the Hatha expression in Yoga means a yoking together of the duality of opposing energies (yin/yang, physical/mental, masculine/feminine). So – I like to think of the art and the car ‘shape’ itself (the Porsche 356 (when it is not being called a ‘bathtub’) as the feminine energy and of course, the masculine energy from the automotive ‘drive’ associated with torque and horsepower.
Each of the 9 panels on the car has a theme. The four corners: the two front fenders, and the rear quarter panels represent each of the four forces: earth, air, fire and water.
There are many references on the car to the lyrics of two of my favorite albums; Jimi Hendrix: ‘Axis: Bold as Love’ and ‘Electric Ladyland’. But I had also immersed myself in the music of jazz artists of the era while I was painting the car. I spent a summer with the car parked by a pool listening to Jimi, Miles Davis and Thelonious Monk while I was painting.
The car was born in 1965, and in 1965 there was a famous gig for the Miles Davis quintet at a club called the Plugged Nickel in Chicago. The story goes that before the gig, the drummer, Tony Williams, had challenged the band to play virtually nothing from their current repertoire for their 2 night run at the club. He suggested they play jazz standards instead but in a radically differently way (in an anti-jazz style). It was one of the events in jazz history that marked a departure from the traditional jazz of the time to newer styles – such as jazz-rock fusion. It was quite controversial at the time.
I found this 1965 connection compelling because I was intentionally painting my car in this completely whimsical, abstract-expressionist, radically different way – even from the way other car artists work. The usual restoration work of these cars goes in the complete opposite direction: minute details of historical correctness and perfection to the nut and bolt level. In a lot of Porsche circles, these cars are regarded as sacred, just as jazz standards were considered sacred to jazz purists. Someone once published in a online automotive blog that they absolutely detested my car. I thought that was delightful!
But for some fun, take a look at the BMW art car series. BMW definitely gets the concept of cars as canvas a lot better than Porsches.
I have taken the car to a lot of shows over the years and during the summer of 2016, a lengthy article about the car was published in the German, Porsche Scene Live magazine [Titled: Not the Janis Car (as in Janis Joplin’s famous painted Porsche 356 which I seem to get compared to a lot even though the cars are very very different in many ways]. I actually think I am getting ready to sell my car this year. :-) I think it may be time to put it out there and risk having a ‘purist’ want to buy it and repaint it. Impermanence, right?
Margaret Warren’s Quick Favourites:
Favourite thing that you could not describe formally but wanted to?
This was my favorite question!
There is nothing I love more than trying to figure out how to formalize really intricate concepts!
In ImageSnippets, we have many advanced features for experimenting with complicated descriptions and a process for doing it. We can create entities when we need to – perhaps for things that might never exist in a public corpora (such as a local restaurant.) But we can also do more complicated things like create subclasses from any other existing class, and we have a ‘which’ statement convention that lets you string together triples, so you can describe objects in the image more completely. With the ‘which’ statement we often temporarily ‘mint’ relations as placeholders. This process is part of a method for ontology engineering that we are developing in which we want to get as much data as is naturally possible at the ‘information recording point’, the moment when users are trying to capture their thoughts in a formalized way – even though some of those relations may not form strictly correct RDF. Our goal is to continue to develop easier ways for users to naturally express themselves and so the use of the temporary URI’s allows us to examine the kinds of things users would like to try to say.
So – here is an example of something very hard and why. Suppose we want to describe a photo of a house where someone we know used to live as a child.
In the interface, I might say something like:
[this image] depicts (a) [house] which *isWhereOnceLived* Pat Hayes
Where: ‘isWhereOnceLived’ would be a URI relation placeholder. It can sort of work as a data structure, but it is truly awkward, not complete and it doesn’t easily lend itself to stringing out other information such as the address, the years he lived there and at what ages. The idea in the ontology work that Pat and I are doing, is to work towards a kind of pidgin language between humans and a machine driven process that can easily be reproduced, used somewhat consistently by humans and reasoned about by machines. But in order to triple tag something like this concept more completely, you would need to have, perhaps an ontology of all the places Pat has lived; when he was born and the addresses where he lived during what years and then you could simply attach Pat’s name and the address to the image snippet and link that to his personal ontology and then, ideally, a query for: where did Pat live when he was 12, would find this image. It’s not impossible – and might now work for more famous people like Mozart or George Washington. And Facebook surely has enough data to make these connections :-) but for us, now, with ImageSnippets, it’s difficult to knit this information together.
We can say, the image depicts a house which hasAnAddress 123 Main Street, Somewhere, Florida, 32561 and then we could query the things in the system related to that address, but without the links to Pat and his history, those data points (the picture of the house and it’s address) would lack quite a lot of meaning.
There are many other more interesting things to talk about that are challenging to formalize. Another that comes to mind is shadows and reflections.
There is a tendency to want to describe a shadow, for example, as an object…though actually, it’s not. It’s really a visual phenomenon occuring when light is blocked by an object.
So, it’s basically OK (in ImageSnippets) to just talk about the shadow as if it were object by itself (even though that is not ‘logically correct.)
[This Image] depicts (a) shadow
And that is fine until you also want to talk about the real object that is creating the shadow, especially when you may or may not be able to see the source object in the image.
So then you find yourself trying to say some variation of:
[This image[ depicts (a) [shadow] which usesPictorially [Tim]
[This image] usesPictorially (a) [shadow] which isOf [Tim]
[This image] depicts [Tim] which hasArtisticElement [shadow]
None of these usages are exactly ‘wrong’ in the system. As part of this research, we group search results based on the relations – so if you search for shadow in ImageSnippets right now – you can see the variety of ways in which we have formalized shadows so far. It’s not terribly ugly, but it’s definitely one topic we feel we *might* be able to axiomatize more clearly. There is some neat work being done by Roberto Casati on shadows we are currently studying. This link is particularly fascinating: http://www.shadowes.org/home/?page_id=161
Favourite ontology element from your ontology?
I think I might have to pick 2 :-) ‘looksLike’ and ‘usesPictorially’
LooksLike – was identified very early on when we were designing LIO. It was such an obvious relation because people already use it quite easily. Our property just means that – it’s some ‘thing’ that is not that ‘thing’, but that it looksLike that first ‘thing’. Just go to google images and type: Looks Like xxxxxxxx (word of your choice; the word ‘horse’ is pretty revealing).
But I think my favorite element is: usesPictorially —- it has been the last element we included in the current incarnation of LIO and I think we added it around 2013.
The distinction between usesPictorially and the other properties is a subtle difference in
the way one would say they ‘normally’ see something real in the world depicted and
then those items that seem to be visually represented in an unusual way – perhaps in an
artistic sense, or maybe just photographed in an unconventional way. This distinction
most often includes artistic renderings of objects in the form of paintings or
Illustrations – particularly abstractions, but it often goes beyond that into abstract photos of everyday objects or things represented from unusual angles. Although it sounds complicated at first and reads a bit awkwardly in pidgin human RDF, almost every new annotator learns this property very quickly.
With each property, we thought carefully about how to use a ‘word’ for that property. We could have easily named them: prop1, prop2, prop3 or foo1, foo2 or foo3, etc. and assigned meaning to each of some made up term, but instead we tried to use words that would be familiar, but could come to have their own meaning for annotators.
As with all of the properties, there will be ambiguous edge cases, but what we are always looking for, are properties that will have an ease of human use and demonstrate a utility in search results.
For a visual understanding of all of the properties in LIO, you can use this:
Visual Guide to The Properties – http://www.imagesnippets.com/ArtSpeak/help/properties.html
Favourite feature (metadata item) about Porsches ;)
This is hard, because there are a LOT of neat historical things to talk about. Off the top of my head, this came up:
There was a color you could get on the Porsche 356 Pre-A cars (built between 1950 and 1953) called: Pascha Red. It’s the deep chocolate brown/maroon red. It’s obviously named for the color of Pascha eggs – originally dyed from natural ingredients like onion skins during the Paschal (Passover/Easter time). Eggs dyed this color at Easter go all the way back to early Christians in Mesopotamia. From my research, no one seems to know why they were dyed that way. Today, in the US, we aren’t as familiar with that particular color and that name for Easter eggs (though one of the companies you get Easter egg dye from is called, Pascha I think) but I just found it fascinating that in the early 1950’s, someone at the Porsche company would create a color and name it Pascha Red.
Who’s Gonna Interview the Interviewer?
And now, Margaret, I have this section Who’s gonna interview the interviewer? Where I would be grateful if you could spare several minutes to ask me a question or two. :)
Margaret: This one might be a bit obvious – but what led YOU to the Semantic Web ocean? I think ‘we’ are a special group of visionaries (those of us who are philosophically interested in the semantic web) but I am especially intrigued because your background as a philologist, you certainly came into the ocean from a different entry point I think. Obviously, most people we meet are computer science academics and programmers or perhaps librarians or curators, but I remember reading some of your first writings on Google Plus quite a few years ago and I was definitely intrigued with your perspective and wondered how you found yourself swimming in these waters!
My entry point was an SEO book: Google Semantic Search. And the quantum portal was my curiosity about one time seeing the semantic web, as is, and another time seeing it with capital letters. Why? So, I started researching. And the concept, and Tim Berners-Lee’s book enchanted me. And I started writing and exploring. Then, Milena Yanklova of Ontotext found me via a hashtag!!! #linkeddata. And that got me into the nitty-gritty of implementing semantic web technologies, and reading even more. Isn’t that mezmerizing and mysterious a way?
Margaret: Do you have a favorite word (set of words)? A poem? What is it or what are they?
I do. One of many. Still. These are Bukowski’s words. I love them because they bring me silence. And from the depths of silence you can feel your words pulsating and ripe to give the gift of giving to the reader. Not rushing. Waiting for the right wave of sound and cognition to bring you the riples you need. So, the words are:
“The important thing is the obvious thing that no one is saying…
Margaret: How do you see these words connected to the semantic web?
I don’t. I see them as a vibrating potential somewhere between the nodes. Waiting to be surfaced upon a query.
Margaret: Have you linked them to a dataset like wikidata or dbpedia before?
I have had an amazing experience with Andrea Volpini who is my cyber alter ego kind of. Back in 2017 we played with The WordLift tool and he and his team built and we did that: Crafting Texts in the Age of Fragmented Reader Experience – a visual heavily linked representation of my words in the article I wrote. And isn’t that article a beautiful thing to watch and contemplate.
Margaret: How do you find images to go with your content writing? What kind of mental steps do take to lead to a querying approach do you use?
I.e. Sometimes we just search for text strings…such as: II hear a song and type in the lyrics. OR I am looking for a product or a brand, so I type that in…but how about something more esoteric…like images that communicate ‘long range thinking’ ? Do you have a vision of what might best communicate a particular concept and how do you go about mentally weaving that to a query? I actually have a bit of a difficult time with this [despite my being able to describe the images formally] – so I am always curious how people come up with amazing images to illustrate complex abstract concepts and wonder how to make this process better with semantic web techniques.
I don’t have a vision in advance. I can’t think it up all. So I let myself flow, or astray in the informational currents around me.
But to answer your question, usually I always search for an abstract that is related to my idea or words and thoughts. For example, this amazing image I found to illustrate the excerpt from my book Linked Data and the Good Text.
I ended up with it thorugh a digital magazine and an article about Frances Yates.
What I know is that I usually end up with cool images when I collide two domains. I had met the concept of De Umbris Idearum and searched for it in Google Images. :)
WOW – what an image and resource. I will find it and put it in ImageSnippets!
And this is where this part of our adventure with Margaret ends. With her never-ending passion for making information auras that last.
See you in my next Dialogue, or maybe in my new adventure: The Intertextual Animal Video Series!