From Speaking Into The Air To Making Products First Class Citizens With GS1 Digital Link: A Dialogue with Phil Archer
“Briefly, technologies such as the telegraph and radio refitted the old term “communication”, once used for any kind of physical transfer or transmission, into a new kind of quasi-physical connection across the obstacles of time and space. Thanks to electricity, communication could now take place regardless of impediments such as distance or embodiment.”
p. 5. In: Speaking into the Air A History of the Idea of Communication John Durham Peters
The knack for communication, writes Phil Archer, whom I want you to meet in this Dialogue, has been the common thread connecting his work at GS1 and W3C, threading a career encompassing online safety, broadcasting, teaching, linked data publishing and writing copy.
A treasure trove of webby ideas, lots of experience when it comes to the Web (and its standards) and a website full of archives, Phil’s work and writing spans the worlds data, archiving, culture, standards, and all that jazz. His key themes are interoperability through common terminology and URI persistence. Before joining GS1 – the standards body behind the barcode – in July 2017, Phil Archer has been part of W3C projects like Mobile Web Best Practices, Data on the Web Best Practices, Spatial Data on the Web Best Practices.
Phil was kind enough to talk to me about linked data, web basics and the love for communicating ideas, building consesus and development of globally-recognised standards.
Distance, disembodiment dissemination
Phil, what was the most fascinating during a live broadcast back in your radio days?
It’s always the audience reaction. The unexpected calls during the show, the interest, pleasure or displeasure in what you’re doing. For reasons that defy all sensibility, country music has an enormous following in the part of the world where I live so pretty much everyone on the station had to take a turn on doing the country show, whether you liked it or not. Not my thing at all but there is a community of people that absolutely love it and they were the most engaged audience I ever had.
Where do you think that invisible bond between the speaker and the audience lies?
Radio is a very personal medium, even more than podcasting, because it’s live. You sit in a studio with a mic in front of you, usually on your own, and talk… and people hear you. You always have in mind that you’re speaking to an individual person, not a group. So the bond lies in that belief that you’re talking to one person and their belief that you’re only talking to them (which given the audience figures may well have been closer to the truth than one likes to admit but that’s another story!)
Is there a parallel in terms of the radio language/scenario and the Web scenario?
[Teodora: At least for me, I can relate to being that one voice speaking into the air 🙂 when writing. Envision the people I am talking to, forms my ways of expression]
Yes. Although you hope a lot of people will read your words, you write with a single reader in mind. You want to give them a reason for continuing to read and to reward them in some way for doing so. That’s exactly the same thought process as being on the radio.
Linked data, seriously?
When did you first hear about the Semantic Web?
Twenty years ago in 2003 (may be slightly earlier). I was working for an online safety charity at the time promoting use of a metadata standard called PICS. W3C’s first published standard was PNG but a lot of that work had been done before the consortium was formed so PICS is generally reckoned to be the first set of standards fully developed at W3C. Almost no one used it, although it was in the major browsers of the day (Internet Explorer 3 onwards and Netscape 4.7). It was suggested to me that the Semantic Web was where the future of this kind of thing lay (the original RDF project was called PICS-NG) so I looked it up. It made sense to me immediately – the idea of encoding relationships between things, ideas and people and them all being connected via the Web.
I well remember sitting on a bench at our local beach (in Felixstowe) reading the draft RDF specs. I understood the Primer and could more or less make sense of the ideas in RDF schema (although pretty much every single term was new to me so it was a steep learning curve). I had no chance understanding the Abstract Syntax but knew that I needed to know more so I decided to pick on an author’s name at random and try to contact them. Purely by chance I chose the name Dan Brickley who agreed to see me shortly afterwards in Bristol. I walked into the room where he worked and there were Libby Miller, Dave Beckett and Max Froumentain all in one place. I was genuinely awestruck. As Dan – now one of my closest friends – can attest, I’m over it now.
How did you end up in W3C?
Working for that online safety charity and getting nowhere with PICS drove me to the conclusion that we needed an improved standard. As well as Dan Brickley, I’d met another Dan – Appelquist – then at Vodafone. Round about the same time, Dan Appelquist was setting up the Mobile Web Best Practices working group and there was the idea of labelling websites as being ‘mobileOK’. This was before the age of Responsive Web Design of course. We were into device description repositories, adapting content server side to match the screen size and capabilities and idiosyncrasies of individual devices. I convinced my boss we had to join W3C and I joined that Mobile Web Group as well as working on my own WG. More on POWDER shortly.
Later, W3C was looking for someone to teach mobile web development. I’d been made redundant from the safety charity and, as I had been a member of the mobile WG, and have an adult teaching qualification, I got the gig.
What did the W3C consortium work teach you? So many perspectives, standardization efforts, this is par excellence communication art, isn’t it?
A great deal. I’m very happy at GS1 and it’s very likely to be my last full time job (I’m 60) but the W3C job was the career-defining one.
Firstly, I learnt a lot about how the Web works. That happens when you’re privileged to be in the room with the very people who originally designed CSS, HTML Forms and Tables, XML and more. The complete failure of the POWDER work was also very instructive. We had big name company support in the WG, sure, and we had implementations. But the implementations weren’t by the big name companies and, in reality, no one wanted it. So you learn that standards are the consensus of the people who wrote it and who use it. If no one’s using it, it’s not a standard – it’s a work of fiction at best. I only ever heard of one person outside the WG who ever tried to use POWDER. It was billed as the replacement to PICS. Well, almost no one used PICS and no one used POWDER. The problem wasn’t the standards, it was simply that the underlying ideas were impractical in the real world.
That brings me to the importance of sharing code. From day 1, to become a W3C standard you had to have at least two independent implementations. That’s crucial, but actually it’s not enough. You need test suites, and, ideally, your standard should be proven in open source code. And you need people actually using it in production-grade systems (whether open source or not). That’s why GitHub is so powerful and the standards development process at W3C – now all on GitHub – has so much integrity.
By the end of my time there I was a gatekeeper for future standardization work in the general area of data. People would ask me for a new WG on their favourite subject. OK, I’d say, who wants it? Who’s going to implement this? And then the cruel twist in the tail: and how many of those people are W3C Members? Because, yeah, we all love the Web but, erm, there are bills to pay.
Standards development is either funded by the people who write the standards or by the people who read them. Somewhere, someone has to put money in the kitty.
W3C also taught me about the value of persistence.
Into that same vein, what’s the common thread between URI persistence and perseverance? 🙂
URI persistence is an absolute pet subject of mine.
Link rot is an undeniable fact of the Web. And it’s so unnecessary! URLs are persistent or ephemeral by policy choice. There is nothing inevitable about it. I’ve written about this a lot, see, for example, Danbri’s Bus and the Fate of FOAF, the Study on Persistent URIs I worked on (for W3C) at the European Commission, Twenty Years of W3C Mailing Lists, and, more recently, the Importance of Digital Persistence. The ability to link reliably to information over time, whether that information is in standards or policy documents, press releases or articles, is so valuable.
What’s the common thread with perseverance you ask? When I first joined GS1 in July 2017, the organization’s website had undergone a complete refresh not long before (I think the previous year). So, all the links to GS1 standards from elsewhere? All 404s. And there was no way to find previous versions of standards, let alone link to them.
Like most websites, gs1.org is entirely in the control of the marketing department. Their remit is one of attracting attention, of putting across the company’s messages and so on. That is completely at odds with the notion of publishing stable documents as reference points on the Web. So when I turned up talking about URI persistence, dated links and latest version links, I was not popular. I persevered for 5 years but I’m very pleased to say that now we have a separate repository for our standards and other persistent documents, complete with archives all managed under a rather familiar URI persistence policy. We now recommend that GS1 Digital Link URIs are minted on a subdomain of the brand owner’s website (if curious have a look at my QR code video).
Teodora: And my last question in that section is a throwback to you, from your own rhetorical one in one of your slide decks. So:
Linked data, seriously?
I remember going to Göteborg to deliver this talk and I remember using a lot of those images multiple times – but I admit I can’t recall the talk track.
But yes, I’m still serious about LD. OK, we can talk about knowledge graphs if people want to give it a modern twist but it goes back to TimBL’s original vision for the Web – connecting ideas, data and documents in what is essentially an entity-relationship diagram. The whole point about GS1 Digital Link is that it creates URIs for products (and batches, serial numbers, locations, assets etc) and the use of Linksets means that we can link each identified entity to multiple sources of info about it. Add in a GS1-conformant resolver and you can ask a ‘thing’ for its instruction manual, its nutritional info etc. all using a common API. It’s Linked Data with barcodes as the identifier.
GS1 Digital Link: Getting the Best of Both Worlds
Tell me about GS1 digital link, the portal bridging physical and cyber. What is it, why you care so much about it?
I care about GS1 Digital Link because it gives a URI to every product, every asset, every location and everything else identified with the GS1 system. That means we can hope to build a node on the Linked Open Data graph with those identifiers at its heart. From the position I’m in, I believe it is the best contribution that I can make to the Web.
That’s the high level, pretentious version. A more practical version is simply that we’re in the process of replacing the 50 year old barcode with one that still contains the same GS1-numbering system but in a way that is a link to online information and yet it still goes beep at the checkout. You can just scan the QR code with your phone’s camera and get a default response, or you can scan the same QR code with a specialist app and get a completely different experience – just like you can with barcodes today.
Then you get into the possibility of resolver services that can return the linkset for an item and, further still, redirect to specific types of information by asking for a defined link relation type.
If where you end up has a nice block of structured data in JSON-LD then we’re well into Linked Data territory.
I find those possibilities exciting and have no trouble spending large amounts of my time looking for new ways of sharing that excitement. The list of household name brands now adding GS1 Digital Link QR codes to their products is growing all the time (P&G, Pepsico & Puma are all recent examples).
Tell me about its connection to schema.org. And what it brings to e-commerce?
OK we’ve shifted slightly from GS1 Digital Link to the GS1 Web Vocabulary now. The two are linked and were originally planned to be a single project but, well, they ended up being separate.
Brands and retailers share a lot of data behind closed doors, often following GS1 standards. We have a large set of terms for describing stuff, but a lot of them are retailer-specific. Our ‘Global Data Dictionary’ is not Linked Data-friendly. The GS1 Web Vocabulary is an attempt to make terms from the GDD available as an extension to schema.org and to make it easier to open that data for eCommerce. The idea is that we have richer terms than schema.org has or will ever have so our Web Voc can be used for richer descriptions.
What’s held us back is that it’s really hard to prove the value. The Google tools generally report errors if you mix GS1 terms in with schema.org terms, even though they’re deliberately compatible. If we had time, we’d create SHACL files for the new schema validator that showed the two working together and generally take part in the whole structured data effort of which your work is such an important part. It’s always a question of resources, not willingness.
From our POV, schema.org is a bit of a mess when it comes to describing a product. Is it a schema:Product? Or schema:SomeProducts? And why can’t a brand describe its product without insisting on either an offer or a review? And so on… we’d really like to improve all that (as we see it).
Change control for the Web Vocabulary is pretty lightweight. It’s now handled by a Standards Maintenance Group in which I play a large role. So adding new terms or clarifying them is easy enough. We’ve just finished a major program to update a lot of our core standards (EPCIS, TDS & TDT) to work with JSON-LD. And the Web Voc is getting more attention than before so I’m hopeful that we may be able to devote some resources to it again soon.
What are the obstacles people face with implementation?
Google often reports errors – so it looks as if we’ve made mistakes. Also, the people we talk to are very rarely the people responsible for adding structured data to an eComm site. In your SEO work my guess is you rarely come across anyone who has ever heard of GS1. The two worlds are, sadly, very separate. As ever, it’s about offering the right tooling to the right people with the right value proposition – and we haven’t managed that successfully. I’d be delighted if that were to change.
What stops people from doing initiatives with that?
See previous – skill and incentive.
What do your daily activities include? I am asking because for me you are an epitome of bringing visions into actions, and I want to emulate 🙂
You’re too kind. I spend a lot of time answering email about GS1 Digital Link. The questions have evolved over time but they keep coming. Current priorities are an ISO standard that is the nuts and bolts of DL but without the GS1 bits. There are two aims to that:
- To create an ISO standard to underpin GS1 Digital Link (important for sectors like healthcare)
- To try and minimise the proliferation of different ways of encoding existing ID schemes in URIs.
The standard I’m drafting already has to handle two distinct methods. One, like GS1 Digital Link, puts the important IDs in the path, in the other, everything is in the query string. But everyone has their own way of doing it so… let’s see if we can increase interop by not inventing whole new ones every time.
I’m also updating the resolver specification (now to be called a GS1-Conformant resolver, not a ‘GS1 Digital Link resolver’ – politics, politics…)
Oh and the EU’s Digital Product Passport is more than a little project 🙂 A lot of colleagues across the EU are working on that so it’s not directly me as such, but I’m one of the people trying to bash out an architecture that is centred on the identity of the product, not the DPP, that uses a knowledge graph and does not necessarily need a blockchain. Quite why anyone thinks a blockchain is necessary for anything is beyond me.
And we’re trying to work out exactly what GS1 can and should do with Verifiable Credentials. They’re on a trajectory to become crucial for all sorts of things like product certification, invoices, cross-border docs and more. Where does GS1 fit into that?
I co-chair a W3C working group on RDF Canonicalization which, in part, is related to VCs. In terms of embedded structured data on websites – wrap that as a VC and now you can prove that the product description came from the brand owner and hasn’t been tampered with – potentially useful I’d say. And for marketplaces – onboarding new suppliers who come bearing credentials is useful. We made a video about that nearly 2 years ago.
Who’s gonna interview the interviewer?
Teodora: Phil, I would be grateful if you could spare several minutes to ask me a question or two. 🙂
Phil: You have a deep understanding of meaning, that is, the transfer of information from one person to another. Your work is about words, text and language but, from what I can see, you make little use of imagery. Is this deliberate or do you simply find it easier to express yourself without images? (background: I use words too, but I like to use images to emphasise or highlight talking points).
Teodora: I don’t use images as this is not my medium. I want to somehow leave enough space for interpretation and also avoid th emulti-modal media experience, only to focus on pure text. I understand a picture is worth 1000 words, but my textual hedge mazes are made of words, not images. 🙂
Phil: Similarly, do you find it easier to convey a set of ideas as a series of crafted paragraphs or as a story? (background: I get complimented on my ability to tell a story – but I don’t necessarily think of it that way. I get confused with some of the feedback!)
Teodora: Yes! Exactly. I don find me walking through the textual lands, leaving trails, paramount to the process of exchange, the process of my inner dialogue is intricately connected to the process of the inner dialogue I want to invite the reader to. Sometimes there’s a story, but sometime I don’t want to own the story, I want that to be abstract enough to be able to containing any story.
I am often told that my writing is aimed at a specialist audience with at least graduate level education. I am therefore asked to simplify text. Do you have the same request? Do you find it easy to express yourself in simpler-than-your-default language? (I usually shout that some things really are complicated, get over it… and then accede to the request).
My editor at Ontotext always tells me the same thing. Simplify! Cut! You said that twice! I don’t get this sentence! Hey, yet another sentence-paragraph. 🙂 I know simplicity should be shown at the end of complexity. And I also know that I AM NOT THE AUTHOR of my text. My culture, me predecessors, my readers ARE! So I let the lattice of words and thoughts weave itself trying to let enough space for breathing, interpretation, meaning-building or unbuilding. And I never underestimate the syntagmatic capacity (I just made that up, if it was a “corporate” text I would have had to think of something simpler, lighter. Thank God, it is not.) of my reader.
Phil: How does your writing style differ when writing in English and Bulgarian?
Teodora: I am a terrible writer in Bulgarian. It takes me ages to make a first ugly draft not so ugly. And you know, that’s so interesting about language, the way it arranges itself depending on whom you envision it to be perceived by. I love that emerging element, which you cannot get if you don’t imagine whom you are sharing with. How can we make something shared? How is that shared understanding fllying through all the communication barriers, language included? I really don’t know. What I know for sure is that I need to be 100 % eager to share, to help that meaning thread get out and further break into other threads, or join bigger ones.
Phil: Do you have a go-to person that you rely on not to constantly tell you how good you are but to tell you how you can do better? (you know you’re good, but who provides you with constructive help to be even better) I have a couple of people I can turn to for that and I value their opinion very highly.
Teodora: That’s an interesting question. I don’t have that go-to person. I am my biggest martyr. I am never good enough as I have assigned myself a crazy task: to explain one and the same thing in multimple contexts, to people of different backgrounds, even of different epistemological stances. This makes me always under fire :). But this is what I want to do. All stemmed from one and now it is time for things to get back to places.
Phil: What set you off on your journey? A love of literature?
Teodora: The love of intertextuality. And the way we connect, or don’t. I want to understand, touch, experience, maybe one day cross that invisible boundary between I and Thou using words and meaning that go beyond the usual chit-chat talk about the weather into worlds where understaning comes from and breeds interaction and action.
Instead of an Epilogue: Webbyness literacy advice from Phil
Tell me more about the best practices on the Web. If there was a “No walking on grass” sign, for our cherished beautiful Web, what would that read?
Separation of content and style.
That is, get the content right then, separately, consider the presentation style. There’s a reason why all the style attributes have long been dropped from HTML in favour of CSS – now a really powerful presentation toolbox. There really is no excuse for using tables for layout (actually, Responsive Web Design has more or less consigned that particular bad practice to the history books – thank you Ethan Marcotte).
No pictures of words (think accessibility and machine-processability)
And it’s still important to keep page weight down, not to have megabytes of JavaScript and keep your code both clean and valid. That’s what delivers the best user experience.
With that I will leave you to Phil Archer’s website and Twitter profile links, for your intertextual pleasure 🙂
For more Web thoughts and deeds, you might also want to dive in Surfing the Web Layer of Things, Technology and Thoughts: A Dialogue with Ruben Verborgh.