LAEP2016: Weds pm

Liveblog notes from Wednesday afternoon, 16 March 2016, at the a two-day expert workshop on “The implications and opportunities of learning analytics for European educational policy”, at the Allard Pierson Museum, Amsterdam, organised by the LAEP project and the LACE project.

Foresight exercise

Small group: Learner profile 4 – Marie Martin; Educator profile 4 – Remy Depardieu.

Concern for Marie about coping with the course. If LA could provide reassurance, based on previous students with her background and how they coped, that would be useful. Maybe recommend crash course, update on technology. Reputation and learning about interaction, but managing those courses requires digital courses in the first place. Also social media skills to manage the reputation. Maybe invest in software that would help her do the course but would hide her identity. Like Tor for biometric tracking. Black market eyeballs! A LA blocker. How can she get good analytics? Blocks her identity, but not her activity. Having other people doing it for you?! We get quite dark thinking about selling her eyeballs to other people. She wants to be able to take the course but not be identified. Identity assurance, to avoid anonymous bullying. You need policy measures, to prevent sale of valuable identity information. She should be able to pay extra to not have her data sold on. Policymakers perspective, we should have a strong framework. Verifying your identity makes sense, to prevent anonymous bullying and cheating, but it doesn’t have to be public. Maybe they can’t make it public unless it’s authorised. What are the main barriers? This biometric registration, for a leisure course – what if she wanted to take a course about her sexuality. It’s different if you are going to be certified or increase your pay grade.

It’s about privacy, and consent. This is about data. The other parts are easy – she can build up skills through a series of Khan Academy courses, or something. But attitudes towards sharing course participation? If you’re going to take a course like this, would I want a group course, or a personal? You want it just in time, just enough, personal. With a group, you have to take them in to account. She wants a one-to-one with Remy. There seems to be a mismatch between what she wants as a learner and what’s available from the school/commercial system. They’re not being customer-friendly. Maybe they have a business model to sell the information on to others, e.g. advertisers. Examples of very intimate tracking. Posture prompts, all sorts of crazy smart gadgets.

Policy changes? Register privately. Consent to it. As a consumer, if she doesn’t like the implications, she should just not register. There’s bound to be a company that will let you register privately.

Plenary feedback

Learner profile 3: Thomas Müller

It’s not good for Thomas.

Disruption for the household is coming through gaming. Stealing resources from the boy next door.

The scenario of unemployment for drivers is not far-fetched, with self-driving vehicles.

Here we don’t see data sources for LA. His career, market-wise, interesting to know whether to go in to this new market, but is it too late? An average person has 7.3 jobs over their lives, likely more in future. Draw on the idea of giving the data back to the learner, career advice. There’s a trajectory. Big data could match and gather data around similar people with this profile, and the decisions those people had made, how that impacts on their choices and outcomes. Give them wise advice. Maybe getting in to an industry when it’s already peaking may not be a smart move. In Maastricht, the mines closed, the Government made investment for a university.

I love this example best! He wants to become a beer brewer, look at LEA’s Box Facebook page, I wrote some articles about LA for guitar players, chefs, painters. What can tech do to support him. Statistical data are so far away from things this guy wants to learn. It’s about taste, fine skills to make a good beer. Also about C21st skills like making good PR to sell your product. The microbrewery scene is all about your commercial appearance. We have a mixture of skills, competences not necessarily related to your subject matter, and extremely difficult to be treated by statistical information and data, like taste. This is a great example of how to support people who want to become beer brewers, wine makers. In Austria, we have schools that teach people how to make good wine. We can learn a lot from that approach. Tricky thing for next couple of years is bringing big data together with creativity, C21st skills, overarching need to be successful in this world. Empowering learners to be successful in our society.

Job market intelligence, disconnection between education and the job market. Redesigning the curriculum is the area to explore here. Hinting at some automated system to suggest career to follow. How do we avoid the SatNav problem – road is blocked, everyone go that way, but it’s blocked. Add some jitter to the advice. Also have to understand what the job is. Universities, there is no limit on university staff. Beer brewers there’s a finite limit. Link to existing quality frameworks – vocational training quality frameworks, a reference we can use and this should be linked to, instead of reinventing it. Who is not on a standard career track? [a handful – suspect most people didn’t hear] Outcome analysis linked to the job market, you’ll always suggest the paths lots of people have done. But most interesting are people who’ve followed a different path.

Teacher profile 3: Maria Koch

Working on training for midwives.

Two aspects. Using LA to look at background knowledge, aptitude. Which 10% are likely to have an aptitude for it. An approach that might be feasible, not sure it’s the right thing to do, because she doesn’t have the expert knowledge, the idea is to use monitoring and video capture of expert midwives in their practice. Matching learner behaviour to expert behaviour is a role for LA. Barriers, lack of expert knowledge is one. How to conceptualise the role of different kinds of education and training. Engaging taxi drivers, maybe if they need a job. Privacy is an aspect, medical practice in real life, how to manage that? VR may be a way, simulate. Risk of lawsuits with intensive monitoring, used as evidence against them if things go wrong. Policy changes, some consideration about policies to allow or require monitoring to go ahead and be appropriately QAed. Maybe indemnify against lawsuits from the capture process.

Similar feedback from another group. Particular aspect around quality assurance. Course development, short time, lack of domain expertise.

Learner profile 4: Marie Martin

Post-human impressionist painter. Wants to learn combination of cooking with 3D printer, but has reservations about tech and concerned about need to give biometric data which might embarrass her and damage her artist reputation.

In terms of learning the technology, LA would help build tools, figure out her skillset, bring up to speed with the technology. Buying it isn’t a problem, doesn’t have to steal it from her grandchildren. The company that she wants to take the course with, she doesn’t want to share her information, she should be able to take a course direct with Remy (the educator) so investigating idea to hire an identity to take the course for her. Like World Of Warcraft players hire people from other countries to build up their character, now. LA Blocker (TM). In a way it’s cheating? It’s like she’s wearing a mask, Remy wouldn’t see Marie, but Bob. There’s an arms race with that. She’s willing to pay extra for premium service, why is this not possible? I’ll pay extra to have full attention. Is it just a financial problem? No, she has power as the consumer, can just find another course. Maybe it’s a market solution, maybe some regulation about what has to be public, or consent.

E.g. schools with base level analytics, or higher-quality. Should there be taxation on the higher to fund the lower? A degree of elitism here, should we have redistribution. You pay for extra services, get an advantage, that propagates across generations. If it’s commercial enterprises that charge a premium, they pay taxes and VAT on their profits, not a problem. If universities, it’s budget allocation. That’s their own freedom of choice. I don’t think you should ‘punish’ universities on the choices they make freely. This is a course driven by personal interest, she doesn’t need to be certified, there’s a different element of policy here. Also policy element, does she understand what the company will do with her data. It should be possible to be anonymous.

Educator profile 4: Remy Depardieu

Well established teacher. Had problems with incident exploding on YouTube. Problem was failure to pick up the impact of institutional change, which should have been picked up. Maybe you need a computer to say he drops off when his learners are more diverse [but maybe you should have noticed that anyway]. Maybe worth keeping an eye out matching the time of organisational change in the analytics.

Our group focused on the internationalisation of education. Already happening. Coursera, MOOC platforms, brings the problem of cultural context. This is a problem that you can try to tackle. Two approaches, one is to avoid it at all, you just generalise, take off all the edges, one size fits all. Interesting for cuisine, already not one size fits all. The other is to take it in to account. I say something, I don’t know how you hear it. LA system could help the teacher with. Trying to bring over this message, take care, there are cultural dimensions in what you’re saying.

Metaphor for standards. It has to do with every learning module you are trying to make. Here everything is in the public domain, you can see a love-hate model. Is that happening now? Here it was an incident, triggered debate. Can the system inform Remy he has a student with a creative response.

We have standardised computerised tests for your driving licence. Because someone ticks the box, yes I will stick to the speed limits, is it real? I know I have to tick that there. I can drive whatever I want. There’s a legal issue there. You consent to obeying the laws where you pass your tests. In this case, creative side of things, it’s different. This student is being provocative. It’s a course on French cuisine. We’re assuming this is leisure-based course, but should be a way for a teacher to respond to this, support cross-cultural work. Company loses reputation, maybe a refund.

Can analytics be so smart? Not even that would be enough. You have to develop algorithms that smart. Behaviour and how it relates – is it like persistence? Keeping on a solution that won’t help them with their goals. That’s different from this one example. It depends on the learning outcome, is it specific, or more broad. If it’s just cooking, Sprite’s fine, if it’s specifically this sort of Coq au Vin, it isn’t.

Foresight exercise 2

Small group:  Learner profile 7 – Giovanni Zanardi; Educator profile 7 – Laura Botticelli.

We’re back to zero with pen and paper, below zero, huge pushback on use of technology in schools. We have to start with policies. My education friends say the teacher is doing analytics anyway, just not with the sophisticated data. Teachers have the rich data in their heads, they know the background of their students, their parents, their interests. There is some analytics in there. Like for K12, not later? Most of these cases, the decisions about using the tools, it’s based on individual events with high impact, and beliefs about whether they are good, not on data about the efficacy. Focus on true knowledge! How can students take PISA (?) tests? I guess they can’t. We randomise, get data out of everyone.

Looking at educator. She doesn’t want to [use LA], she likes chalk and talk. To what extent are they allowed to do what they want? There’s a national curriculum. But the ways of working may not be standardised, may just be learning goals, but the teacher can decide how those goals are achieved. I expect they would say about digital skills … but not here.

In the learner profile, he must be doing things in his free time. Even nowadays we have a backup, can restore to the previous days. This is a political decision. We need policy. Strong EC decision here, to anchor curriculums in core European competencies, including digital ones, in every subject. This could be against competition laws. From research perspective, a great natural experiment on learning tools. But you can’t collect data on the schools? Do it with pen and paper. But then it’s a different outcome. Why are we talking curriculum? The personalised learning, they’ve gone back from that. Like Poland rolling back the reforms. Reduce number of years, reduce the number of dropouts. Any evidence of improved PISA scores, because they don’t drop out? 15yos take it. Then 16yo are now out. Suggestion to impose a common European curriculum on the member states.

If we can show that students are unemployable, can use that to follow up. Wait until they graduate, see if they get jobs. Doesn’t say the workplace has no technology. Wait until the companies start complaining. That may take a while. They are developing replacements. Like Steiner, Montessori learning. Adapt to corporate surroundings with technology. Maybe they can be the innovative, out of the box thinkers. Not enough energy to power our devices. Risking the education of a generation based on a thought that there won’t be a replacement for power? Too high a risk! Do you ask for an operation without anaesthetics? No! Why leave our students without a chance.

[Diversion discussion about Brexit!]

There’s an institute campaigning against ICT. More serious is he’s at risk of exclusion because of his device. Thinking they may be sent out of school just for wearing a device, it’s kind of a discrimination. Many schools have practice fo collecting cellphones, they can’t do it if it’s implanted, you have to send them home. We already don’t let people go to school with a scarf. It’s discrimination.

What about analytics? We don’t have data! Have EU try to convince leaders of Italy that their policy is wrong. Perhaps more subtly. Track student mobility, in tertiary education. Immigration, tracking employability of Italian graduates across Europe. Maybe people will move, people with kids move away to get a better education, a brain drain. It’s already the case in southern Italy. People move north, to the cities.

What do we do about this institute? They might have a point. [laughter]

Plenary feedback

Learner profile 5: Åsa Anderson, Educator profile 5: Kristofer Palm

8yo being ranked swiftly. Many things wrong with this. Not too little use of LA, but too much. Too much measuring, leads to stress on children. Very behaviourist, quick responses, very granular. Maybe broad feedback is Ok. Comparison issue, normative measures are problematical. Notion of, what’s it measuring – what you can measure, rather than measuring the right thing. In this case, people selling a product, wizardry, needing to make sure that user is bringing a sense of pedagogy, what the child needs. She’s cheating!

Is help from a friend cheating? Getting questions about using social media in class – that’s called living now! Real challenge if you organisation relies on IP, if they are chatting on Twitter. What do you recruit employees for? Recruiting me and everyone who follows and responds to me online?

This scenario, white gleaming classroom with chrome. It’s antiseptic. Bleached white wood too. It’s cool and interactive? Question about diversity again. More efficient sausage factory, make more sausages that are the same. How do you maintain creativity if very specific goals. This is what marketing companies are selling to us, you’re skeptical. We’re not skeptical, we’re antagonistic – we could do this, we don’t want to. IBM are doing this already.

The teacher is wondering if they’re doing what they’re trained for. Most would like to have better control over classroom, especially new teachers. Some type of support. Not sure they would like to have the motion-tracking support of knowing this child is bored or not, on or off tasks. Sometimes good to let them off task a bit, important part of being a teacher. That freedom is a bit lacking. As a teacher, I can see you talking to me. This is an erosion of trust, your role is to trust your colleague. The student is gaming the system, but maybe that’s Ok. Inside education you can do that more than outside. Maybe she’s not really cheating.

Learner profile 6 – Natalia Eglitis, Employer profile 6 – Lina Xin

Natalia wants to go abroad, does she have the chance. Is this LA, or just personal choice? Do we have to have policy about this, or is it just what you do? If you share it with FB, that’s your choice. If you share it, employees know it, it’s also a disadvantage.

You have this on LinkedIn with voting. But maybe in the future you will want to see evidence. Here she might be good theoretically in building bridges, but maybe we should see the data. More and more our students will have to prove they are capable citizens. Stanford doing online student transcripts with linked evidence, that’s quite fun.

It’s a matter of personal choice to apply, doubt we have a role as teachers. Why did she not just go to Portugal?

Why would you want them to speak Chinese if they have an implant/device that can do it for you? It’s not just about speaking. They don’t have to learn the language, it’s not so important, there’s many things you can do beside talk.

Like influx in Sweden of refugees, big market for language skills for them. If not sure how many are asylum seekers, how many will have to be moved back.

Learner profile 7 – Giovanni Zanardi, Educator profile 7 – Laura Botticelli

In this world, there’s retraction from use of technology in schools. Two incidents. Legislation response.

Our discussion said, Ok, can’t really understand how they retract so fast, so believe this is a political movement to something else. The reports from the institute is leading us to believe there is a political thing behind this with subsidiary agents working against use of technology in school system. Reflected in frustration of the learner, he would like to use tech but not allowed. Device implanted in his hand, he may be expelled. Raises interesting questions on discrimination, exclusion of students. One thing is to collect cellphones from students. Another thing is to send them home based on what they are wearing. Link to wearing hijabs and the debate there. It is definitely controversial. For the educator, brilliant profile, lot of sentiments teachers have today. She is relieved not to use tech in her instruction any more. Some problems, Brexit victims, trying to reach them, other places there are programs that are better to catch up with problems like this. She’s retracting to comfortable position of chalk and talk curriculum.

How to crack this one? All educators are analysing their pupils, very richly, dense data from human interaction. We’re lacking overview. THinking supranational, recommendations from European level, tracking how Italian education system is performing now. Are the outcomes from the Italian system, students graduated, are they still attracted on the European employment level? Of course, if one goes that far, you could impose central European curriculum, common curriculum, including technology use.

It’s a provocative approach. Say if in 2027 PISA say no difference? There’s won’t be PISA, they can’t do that. Say they make exception?

The teacher and parent perspective is great. We can’t believe that technology is extracted at all, only the education system. This could be reality in some places in 2025. In eastern Hungary, don’t even have windows, that would be an improvement. The context will be different. I’d guess 80% of teachers would love it. If they were used to it, they would not love it.

Informal use of it? Could Brexit parents let them use tech at home? Is all of Italy offline? No, only schools. The teacher could go to them and say, maybe you should try to learn Italian at home.

[tea break]

Learner profile 8: Elena Fechter, Educator profile 8: Ernst Bild

Conversation about trust and trusted third party, role for Commission to look at those. Philosophical conversation: many questions.

Issue, we’re too much talking about elites, an elite picture of our learners, parents want to access data of my 25yo just passing final exams (!). I think, most people are so smart, already elite. I’m one of the average guy. I do not know any 15yo boy or girl interested if they’re in the top 3% of their peers. That is not a reality for me in my environment. The arguments I see appear to be too much driven from elite, societal perspective. Conversation, you must be above the average – that’s the way of thinking – half of them must be below the average. The median. (!) How to translate our love of big data, we have them, partially – what can we learn from them? I’m pessimistic that only this approach, focusing on elite, come from number 3 to number 2 or 1, on the one hand, and from this charming big data approach, to a very realistic learner-centred approach, in real-life settings. All of those scenarios, they scare me a bit. It’s from a bad science fiction movie.

Why’s it bad? The more data we have, the more precious people feel about their data. Elena wants to own it?

It’s a dystopia – it’s not meeting either of their needs. People obscuring their own presence, that goes on today. People are not always going to be forthcoming with information if they don’t trust the environment. That’s not just a future problem. Not having the information might be realistic.

We’re talking about learning analytics, analysing data. That is only one part. We must be very cautious about what we personally think we can do with these analysis.

Should EU prevent students providing data?

There should be a minimal subset of data, to bring up the quality of education.

Learner profile 9 – Juan Hernandez-Santiago; Manager profile 9 – Marianne Salome-Hernandez

Spain established an awesome system. Open access database, it’s visible for everyone. The manager of the university has a great one. Silicon Valley model of outsourcing, online coaches from abroad, feedback.

The issue was, even though they have a great system, you can’t add something from outside the system. Even though an open system, it’s closed to the realities of Juan’s opportunities. CVs no longer exist. We don’t know how to solve his big issue.

For Marianne, she should look at more quality. Question is how trustworthy are her measures, approval scores. Bit of lack of trust in this data. Maybe needs to see how to get more quality and trust in to the system. Make sure that her selling point is to make the brand of the university, not just cheap.

LA has a big flaw here, hasn’t predicted there won’t be Alonso in 10 y.

The right hand side, Oviedo is not so good a university right now. If the data said so? [unclear] Satisfaction is high, so the throughput is good, high quality. If you’re selecting a certain group of students, perhaps they’re paying a lot, of course, they are very happy. If this system works, the data is reliable. If that many are satisfied, you’re doing something wrong. You want people to make mistakes to learn. Some people will not be satisfied, that’s a byproduct of learning. It’s not how many mistakes you made, it’s you have a degree, lets you be well paid. But this isn’t saying that.

Maybe thanks to LA and going global, many people in Chile and Peru don’t know the reputation for Oviedo. It’ll never happen. Why not? JISC national infrastructure, self-declared stuff from students to bring their own data. Bring diversity, do analytics over that diversity, will solve those problems.

Validation of foresight exercise

Each participant has three stars: if you like a scenario, add a star at the top; if you dislike it, add star at the bottom.

[Lots of star sticking.]

Plenary: Potential for European policy related to learning analytics

Imagine you meet the minister for education at the airport, have one minute to convince them what action should happen now, and wish list.

The minister will say, don’t ask me for money. You can ask anything, but no money.

Teacher union perspective about the dynamic in the community. You need a reference model.

In this example. Ministry of education, I met our minister in the summer last year. At a conference of teachers. E-education summer summit. I put on the wish list, we should shift teacher education, still driven by people of my age or older, to a level where media competence, the ideas of learning analytics, and perils, dangers, data mining – all this is treated in a way next generation teachers can use the solutions we have, and will have in a couple of years from now. So they can use them effectively for the benefit of their students. More comprehensive, complete, realistic teacher training. There are already efforts to do that. This doesn’t cost money, there’s teacher education anyway. It must be modernised.

On the now list – clear statement from privacy commissioners about the benefits and controls to protect learners, teachers, society.

Two things. First, orchestration of grants around a reference model, so we don’t have replication. Need to start being more efficient. That’s an immediate thing. [Can you say, you’re wasting money?] Sure you can. They will pay attention. Can say, by assessing project based on number of LOC, you are ensuring lots of LOC written. You have projects ending, particularly LACE, if you don’t continue the Evidence Hub immediately, you’ll be inefficient because it’s not visible. It’s important to continue the LACE project.

On evidence, the risk we have is LA takes us down an objective view of education, promotes the objectively measurable things. One is on, thinking about C21st skills. Curious, plenty of people writing the scenarios are engaged with this, but they didn’t come through. Gather evidence for C21st skills relationship to LA will be a powerful tool of realisation of LA that does not favour a particular view of education. It could be done now, it’s on the agenda…

Evaluation. Another challenge, thinking about answering the question, was the LA successful? If we focus on attainment, we miss relationship between LA and process. Framing the evidence about success around process, not just headline results that mask benefits and disbenefits. If we say do PISA different, focus on what students are doing in the classroom, without extra cost, that’s a heck of an exercise. But teaching only happens where process happens in the classroom is. The final destination doesn’t help the teacher to act. Have to think about action.

Cost benefit. If I’m investing 100k in a solution, at school level, what are the consequences of that expenditure. We could have a mandate, 2% of ICT budget should be spent on LA. That could be a recommendation – do we want to do that? I don’t know it’s mature enough to say that should be an aim.

A Kickstarter type thing for LA companies. It’s almost asking for money. European approach, different stakeholders, companies, SMEs that want to build on LA to create opportunities for innovation. Biggest issue is what we do when the budget ends. SMEs are supposed to take over, it’s not so easy because of the way the funding works. Not an MIT Media Lab model, but where private/public opportunity beyond the grants. It doesn’t have to be connected to research. It can be a detriment to connect to people who think too deeply about things. Partly innovation is about disconnection, self-belief. It’s a different type of grant, bit like Innovation in Action in H2020, but not as formal. Crowdsourced, more market need would be good. Not sure how to formulate it. So e.g. crowdfunding, EU tops it up, then gets something back if it works. There are existing loan schemes like that. It doesn’t need to be crowdfunding. Take H2020, small innovative companies, plan on 1 page. Make application approach like that. Not 100 page document, why not 1-pager, 3 milestones, 10 letters of recommendation. Would save process costs.

More comprehensive approach, more scientifically based, identify the success cases. Bring this to the reference model. Solid foundations to reference models. Need to identify which are the success cases. Taking in to account cultural differences, curriculum. Maybe not just one methodology Perhaps we should have more stronger international coupling. Easy to deploy in your institution. Could be the LACE Evidence Hub, or more general? Methodologies is a key aspect.

One problem, we discuss LA nicely here. What’s going on outside? People know very little about this area. What ministries, politicians can do, is spread the word. Ministries can try to delegate this task to local communities, in schools you should start discussing it. In 2025, these people will be living in that world. If we do not start it now, it’s already too late. An ambassador? It could be. Part of the national curriculum somehow. Include in the teacher’s curriculum. Raising awareness, organise conferences. The field is not major enough. Collecting good practices is still difficult.

Technology standards and architecture. Implicit in orchestration of grants. Several years since we lost the workshop, no standardisation track in Europe that’s open access. Problem was that it was a route to take outcomes in to pseudo-standardisation, not drawing from practice. But now are creating profiles for standards, JISC work in UK, xAPI in NL. Put them in to practice for analytics, what will work from capture, data science, to users. No mechanism to create those common profile standards so people can use them across the market. Open access forum to allow standards to be created from practice.

Technology, interoperability is important. But is that a solution? Heterogeneous teams. Practice. What happens if you put IT people in the school, they make UML diagram of how they work. Innovative pedagogy. Heterogeneous teams. Maybe we could suggest, in the NL, teacher has EUR3000 for education. Take that money, put it in to a challenge or focus, put technicians in to schools to work out, we need technology, but we need pedagogy driving that. This combined approach needs to be addressed.

We are overstepping territory. Creating solution, also creating the problem. nobody else is. I would ask the minister, what is on your agenda, your long term agenda. What do you want to change. Then we know their priorities, we can create that, use it to create solutions.

They will say – I want to be top one in PISA. Any solution for that I will take.

Gender bias, youth mobility, inclusion, early school leaving – that can all be supported. Follow-on question: how will LA help solve those problems.

Data privacy, ethical standards. Openness. Transparency. LA is part of a hierarchy of data-driven decision-making.

I don’t love data. We need to be smarter. Not tomorrow, in 2y, but in 10y. Need to find smarter solutions than having statistics, like Google Flu Trends. Tried to estimate flu traces from Google searches. That is a technical problem but little to do with human learning. It’s so very different. Assignment to classes. There’s a big big danger that if we have an imperfect set of data gathered through whichever channels, we are biasing people to make decisions based on imperfect data. A big danger I see. We need big data for that, ways to treat it, smarter than Google Flu Trends.

Voting!

Top three action – innovative pedagogy, evidence hub, and ethics.

Top three wish list: Teacher education, decide which problems we are solving – third is data privacy, ethical standards, open and transparent.

Next steps

Will send out a draft report for comments.

One further thing. To see how LA can support systemic change. Education ministers, they want change, the bit trouble is inertia. Bringing change is difficult, many stakeholders. The question could be, how can LA and tech that provides evidence, how can that support systemic change. If you do research, research on that. Then we need money. But that’s how you justify asking for money!

Thanks to all, safe journeys home.

–
This work by Doug Clow is copyright but licenced under a Creative Commons BY Licence.
No further permission needed to reuse or remix (with attribution), but it’s nice to be notified if you do use it.