Before we met at university, Laura suffered with ME as she entered sixth form. She missed out on so much, as her friends, who saw her frequently and kept her involved of course, nevertheless started to go out and socialize while she was immobile at home. After her recovery, as soon as I got to know her, I could see she was making up for lost time. According to her, she was the one always organizing activities for her friends at university. She expressed a certain frustration at the natural disinclination of others to make fun plans, but this was couched in a very real and wry self-awareness that her drive was unusual and often comic.

We started our life together, and almost everything fun I can tell you about that life almost certainly started with her. We hitched to Paris and slept in bin-bags under the Eiffel Tower, drinking apparently the only vile bottle of wine for sale in France. We walked part of the Ridgeway and drank a pint among the Avebury stones. We bought our first house. We got cats. We made each others’ wedding rings. We moved to the remote countryside. We learned to sail. We got our first dog. We took a cruise across the North Sea. All of these were Laura’s ideas.

It was Laura’s idea to buy a share in a narrowboat. Over twenty years on, that ownership is sadly coming to an end against our wishes. I can’t believe it will have outlasted her, but it means she got to enjoy it until the end. It was one of the best things we ever did, as holiday after holiday was an absolute joy.

Once the children were born, it was Laura who organized activities time and again for them both. Alys now has her grade 6 in tap dancing, thanks to Laura starting her in lessons when she was four years old. She has been volunteering at Kielder Observatory since she was fourteen, and is applying to study astrophysics at university. Barnaby loves learning the piano and performing with the Bellingham and District Dramatic Society. He will hopefully start volunteering with birds of prey once he is old enough, after Laura set the ball rolling.

We went to music festivals, we had holidays abroad, and we had a fabulously packed London trip just last year, seeing Shoot From the Hip, The Play That Goes Wrong, and Matilda. We camped and walked, and went to Riverdance and the Strictly Come Dancing live show.

Sometimes Laura’s planning was overambitious. Sometimes she was frustrated at the lack of time to do what she wanted. She was famously quite astonishingly terrible at estimating how long any given task would take. We had holidays of which we missed the first two days (out of seven) because of her passionate hatred of packing.

Laura supported me through the deaths of my uncle, my dad, my sister, and my brother. She took Mum to Chinese watercolour classes after Dad died, and was there through Mum’s increasing dementia and eventual move to residential care. She was hit so hard by her mum’s death, and looking back, perhaps never really recovered from it. None the less, she arranged holidays to Istanbul and Toledo for us and her dad.

We went back to the old town on Rhodes last summer, and revisited many places with the children which delighted us when we first went in 2000. The day before we were due to have a week on the narrowboat in February, the doctor told her that she simply must not go. I took the doctor’s advice, and absolutely insisted that Laura stay safe with her dad, within reach of medical care and support. You will be unsurprised to hear that my insistence fell on deaf ears, and Laura enjoyed what turned out to be her last holiday on Willow, again revisiting favourite places from the last twenty years.

In the last two weeks before she died, Laura went to stay with her dad to take care of him through post-op recuperation. She went with him to buy the plants to go in the pots laid with her mum’s ashes. She lost all strength. I knew things were more serious than they had ever been as she stopped reading, and was sofa-bound, comforted by “crap telly”, as we always described it. But she still heard Alys perform a solo piano piece at her school concert. She still spoke lucidly and made plans with her manager at work.

Laura has gone, and the hole she leaves in our lives is incalculable. I knew we could lose her, but the suddenness has left us baffled. If I can do a tenth of what she did with us, I’ll be proud. And she’d be cross that I’d not done more, as she always did.

We met at university, studying Anglo-Saxon, Norse, and Celtic. Laura loved Anglo-Saxon history and palaeography. Her love of manuscripts started earlier. She described looking at a reproduction of the Lindisfarne Gospels with her grandmother. Laura said “It’s so old!”. Her grandmother replied “I don’t like it because it’s old, but because it’s beautiful.”

In our last year at university, Laura revived the Troubadours, the university medieval music society. She broadened it out from a single hurdy-gurdy player to a fabulous group of enthusiastic amateurs. It was important to her to welcome everyone, in contrast to some university music groups. It was a highlight of our time at university, and we remained in touch with several members.

When Laura moved up to Newcastle after we graduated, her first job was at Past Times. She made great use of her staff discount, albeit on less than the minimum wage, a couple of years before such a minimum wage existed in this country.

Laura then spent a couple of years in the funding sector, and during that time published her article Some Anglo-Saxon Cuthbert Liturgica: The Manuscript Evidence. This prefigured her association with Bede, and built on her work with early music, specifically neumatic notation.

Laura then landed the job of Museum and Education Manager at Bede’s World in Jarrow. She was thrown in at the deep end, promptly being required to tell some of the specialist craftsmen who were building the reconstructed Anglo-Saxon village that their work was finished, and their contracts would not be renewed. It affected her deeply, as a young southern woman, newly appointed, affecting the livelihood of incredibly skilled northern craftsmen.

Laura then did something which probably defined her professional character. She knew that the Lottery-funded huge capital project to build an enormous new museum was very close to its proposed grand opening. However, she strongly believed that the interpretation for the permanent exhibition was not good enough. Laura understood the implications of this, and yet told truth to power, in arguing to the board that the opening should be postponed. They were persuaded by Laura’s authority, and the opening was put back.

Laura then threw herself into every aspect of getting the museum ready for opening. She edited, rewrote, and argued every interpretation panel. For her, the cardinal importance was to show what an extraordinary place of worldwide significance Wearmouth-Jarrow was. She railed against designers’ obsession with Vikings, arguing with characteristic dryness that concentrating on the people who destroyed monasteries was probably not preferable to celebrating those who lived and worked within them.

After a successful opening, Laura worked tirelessly to promote Bede’s World’s primary raison d’être: to educate and inspire people about Anglo-Saxon England and Bede. Her drive for authenticity was what motivated her. In her time at Bede’s World, she edited a new guide book, wrote and collaborated on temporary exhibitions, organized academic lectures, and was the voice of authenticity for the reconstructed Anglo-Saxon farm.

Laura’s academic side did not preclude the everyday logistics and realpolitik of working for an independent museum. She built partnerships and repaired damaged relationships. Her funding knowledge was a huge benefit to the museum. She helped commission artwork - always authentic to the museum’s core purpose - and brought international scholars, curators, musicians, and authors to Jarrow to give talks and performances. She smiled as our dog Max barked at John Prescott.

Laura was humbled to work with scholars such as Michelle Brown and Professor Rosemary Cramp. Professor Cramp’s death hit Laura harder than she expected it to. She would utterly dismiss the comparison, but her standing inside a collapsed septic tank had a similar vibe to Professor Cramp weeding the Bede’s World garden.

Laura naturally became centrally involved in the gathering momentum behind the bid for World Heritage Status for Wearmouth-Jarrow. When a full-time job was created to work on the bid, she knew she had to apply, though she was torn in two at the thought of leaving Bede’s World. Several intense years of partnership management, academic research and writing, and community building followed.

The World Heritage Status bid was rejected. This hit Laura hard. For it to fail after so much work, and for such unfair reasons, was a bitter pill to swallow. For her to have left Bede’s World in order to work towards it was very upsetting.

Laura’s professional involvement with Anglo-Saxon history and palaeography largely ceased after that. Bede’s World subsequently closed down. While this is hardly a disinterested opinion, perhaps it would have survived with such a passionately informed historian and multidisciplinarian as Laura back with the museum. Who knows?

Laura’s love for history, and Anglo-Saxon in particular, was lifelong. My belief is that the long long unpaid hours she worked above and beyond to get the museum to a quality fit for opening took its toll on her. She learned a pattern of overwork and stress, and this carried on through the rest of her career. But I don’t think she would have missed it for the world.

Laura and I met at university, and she agreed, when we graduated, to move up to Newcastle from her native Hertfordshire. From the very beginning, work was incredibly important to her. She defined herself by it.

Laura started out with a shop job in Eldon Square - do people remember Past Times? She also did a little office cleaning work, and served at The Bay Horse in Dinnington. She was known as the Cockney Barmaid, and she and the locals found their mutual incomprehension of each others’ accents extremely funny.

Laura’s next job was for English Heritage at Bessie Surtees’ House on the Quayside in Newcastle. Having moved from ersatz retail heritage to the real thing, she sought out opportunities beyond the administrative. Together with her colleague and friend Mary, she put on a living-history event for schoolchildren, and she worked at Corbridge Roman site as curatorial assistant. She quickly grew to know and respect many people at English Heritage and in the wider North-Eastern heritage sector.

Laura then moved into funding advice at the National Lottery Charities Board behind the Live Theatre, close to Swan House Roundabout, and then North Tyneside VODA at Linskill in North Shields. I think it was VODA which really cemented her love for her adopted North East. She loved helping out small organisations such as the Tynemouth Model Boat Club and many, many others. During this time, she also started a law course, qualifying in criminal, contract, and I think property law.

Then came the beginning of her career in heritage, perhaps foreshadowed by Past Times, when she amazingly landed the job of Museum and Education Manager at Bede’s World. The museum had received a multimillion-pound capital grant (remember them?) to turn the small Bede Monastery Museum into a major attraction. She threw herself into the interpretation for the permanent exhibition, and early on had to recommend the postponement of the grand opening, not an easy thing to have to say in a new job.

We married in 2001. And around this time, Laura successfully completed an MA in Museum Studies. She later had her prime professional focus and commitment acknowledged in her appointment as Curator at Bede’s World, later stepping up to Acting Director.

The plan was then devised, and the partnership formed, to bid for World Heritage Status for Wearmouth-Jarrow. Laura was appointed jointly by Sunderland and North Tyneside councils to work on the bid. The bid was ultimately unsuccessful. A lack of support and belief from a key senior individual at DCMS combined with an academically inept application assessment guaranteed that the bid could not succeed. The assessor’s report described the standing Anglo-Saxon remains at Wearmouth-Jarrow as being less significant than those at Lindisfarne. There are no standing Anglo-Saxon remains at Lindisfarne. Despite this, the Wearmouth-Jarrow board decided not to appeal. Laura strongly disagreed with this decision.

It was during Laura’s time at Wearmouth-Jarrow that Alys arrived, as if she didn’t have enough on her plate already. Alys’s achievements made her prouder than Alys probably knows.

Laura then joined Northumberland National Park Authority, where I worked, to build the funding case for The Sill, a new visitor centre on Hadrian’s Wall. This marrying of her love of heritage and the countryside with her knowledge of funding and the law was an inspiration for her, and she played a big role in achieving success. After the failure of the Wearmouth-Jarrow project, this was incredibly important to her. I was proud to play a small part in the fitting out of the Sill to achieve the vision of which she had been a huge part.

A short stint at South Tynedale Heritage Railway followed, and around this time, Barnaby joined us. It was an extremely hard pregnancy, but Laura kept working. Barnaby’s apparently bottomless kindness and humour have been the hugest joy to Laura. I’m not sure she would have made it this far without Alys and him.

Then followed Laura’s decade at the Arts Council. During her time there, she helped such organizations as Beamish, Woodhorn, Bowes Museum, Sunderland Museums, and of course her beloved Jarrow Hall. Much like her time at VODA, she loved being involved in the cultural heritage scene of the North East and the rest of the country, and her role as Relationships Manager dealing specifically with Museums meant that she’d successfully continued her professional life in heritage until the end.

This piece concentrates on Laura’s work. It was so unbelievably important to her. I have always been, and will always remain, in total awe of what she achieved.

During my life, I’ve seen many fabulous players and managers fail to win at Newcastle United: Gazza, Quinn and McGhee, Kelly and Peacock, Beardsley, Cole, Keegan, Big Les and Shearer, Sir Bobby. All legends to us all. But Burn, Isak, and Howe have simply done what I genuinely thought was impossible. While supporting the Toon, I’ve lost my dad, my sister, and my brother. I’ve got married and had two children. I’m a great uncle, for God’s sake. But Big Dan Burn’s header will never fade.

Everyone’s talking about the Dan Burn story, as is right. But what about the goal itself? Trippier’s corner was astonishing. Howe said post-match that set pieces hadn’t gone well in training, and you can imagine why. Tripps lined up for a typical right-footer’s inswinger, but instead cut across the ball, slicing it severely to bend it away from goal. To do that with such accuracy to find, not BDB in general, but the specific spot five yards in front and a foot above DBD’s head, is simply astonishing.

Big Dan then had the platform to do so, but to get that header so perfect is one of the greatest things I’ve seen. For a major final, I’m thinking of Keith Houchen in the Eighties. To be brutally honest, I feel the keeper might have had the time to take a another step to his right before the dive, and maybe he could have got it. But there’s the microcosm of the whole match - Liverpool were simply poor. And that header deserved to take a Wembley lead.

The second goal is getting less attention, but it was such a team goal, and had so many stories behind it, that I love it just as much. Barnes had a lot to live up to, standing in for the criminally idiotic Gordon, and he has often gone missing in matches. Not this time. He held the line to take Tino’s pass, and then returned it round the corner with perfect weight. Tino, who had an absolute blinder all match, then channelled Keith Gillespie against Barcelona, crossing at right-angles to his direction of travel, and falling over in the process. With his weak foot! I can say with certainty that I didn’t have a Murphy knock-down on my card - that header, in its own way, was as impressive as Burn’s goal.

Isak then did what I don’t think I can remember any Newcastle striker ever doing. He made a point of showing that the offside flag correctly shown for his disallowed goal moments before was simply not going to stop him getting his Wembley goal. Again, I just can’t think of many Newcastle strikers who could have kept a half volley like that down, but Isak is quite simply the best striker I’ve ever seen play for us. And after all the plaudits from his other team-mates and the crowd, a silly handshake and the hugest of smiles between Murphy and Isak, perhaps the partnership of the season.

It all went as sour as it ever could with Isak as the summer unfolded. He will be justly vilified when he comes to our place. But he was part of the greatest moment in my supporting life.

A longstanding client requested a minimum viable product (MVP) for updating some of their data via a web front end.

Designing an MVP dictates the quickest development path possible to a deliverable solution. The web front end itself was the MVP, as the database already existed. Not touching the database would give us more time to work on this front-end GUI.

However, there was another stronger reason not to touch the existing database: it is live, the public querying it every day via website and app. It therefore made no sense to try to re-engineer live data for an MVP.

A final reason for not touching the existing database is that it stores the data in question in related tables. While useful for storing richly structured data, this is not ideal for geographical representation, and especially not for spatial use (in other words, on a web map). GIS works with features, expecting them to have a flat list of properties, not related tables of attributes.

My colleague Matt mentioned a technique he had used on the another project for the same client: updatable views. This technique had great potential for this MVP.

Databases store data in tables. Many systems query those tables directly, retrieving the data and presenting it to the user. However, databases also support views. Views don’t have to be of single tables, but can be populated by complex SQL statements. They are often therefore used to gather data from many tables in a highly relational database into a human-readable form. In our case, they solve the problem of converting relational data in the database into flattened properties for the webmap features.

Database views are traditionally read-only. Their primary purpose is the presentation of data to the client. But Matt had mentioned updatable views. How is this done?

This project’s database is PostgreSQL. While at a broad concept level, views are read-only, PostgreSQL does allow views to be updated: passing the updated values to the table on which the view is based. However, there are severe limitations on which views can handle updates in this way. Our views did not fall within these limitations, so we have to do some more work to get this up and running.

Database triggers are “listeners” which can be assigned to fire when specific events occur. For example, a trigger can be created which fires when a specific table’s data is updated. A trigger function is called, which carries out whatever ancillary function is required – perhaps inserting a row into another table.

A view can have triggers, and although a view is read-only, it can still have a trigger which fires when an UPDATE query is made against it.

This fact opens up a whole model of abstraction within the database itself: store the data in tables, create views, present the views’ data to the client, and query the view for create/review/update/delete (CRUD) operations. Add triggers on the views to call functions to update the tables. This abstraction is extremely valuable in our use-case: we don’t want to touch the underlying data structure at all, but still want to be able to update the underlying data.

So how does it work? For each view which you want to be able to update, create a trigger for each CRUD operation you want to support:

• ON BEFORE INSERT
• ON BEFORE UPDATE
• ON BEFORE DELETE

We set these triggers to fire BEFORE the events above because none of those SQL operations can succeed on a view: you can’t INSERT, UPDATE, or DELETE something which is read-only. Each trigger calls a respective trigger function which carries out an INSERT, UPDATE, or DELETE on the related tables where the view’s data is stored.

But how do we get the new data which the trigger function needs to update the tables? We coded the client to carry out the following pseudocode:

UPDATE view SET view.field = CHANGED_VALUE;

This fires the ON BEFORE UPDATE trigger, which calls our custom UPDATE function:

UPDATE table SET table.field = CHANGED_VALUE;

But how do we pass CHANGED_VALUE from the original query into the triggered function? In PSQL (PostgreSQL’s SQL dialect), we use the NEW keyword:

UPDATE table SET table.field = NEW.field;

The NEW object has all the same properties as the original UPDATE had columns, with the values assigned in the original query.

And that’s it. We can make our view as complex as we want (I could write a whole post about using CROSSTAB to destructure related tables), and then restructure the flat input and update whatever underlying tables we want. The table structure is never touched.

Thanks to Matt for sharing this valuable technique.

Viking Raiders was a long time ago. That wonderful, innocent age of gameplay alone, and writing games in Basic, is what gave me the career I have today. A few forays into PC gaming followed: Civilization, one of the Star Wars ones, Myst (completed), and Riven (incomplete) were followed by years of nothing, until a little phone gaming (well, really an extension of my middle-aged enthusiasm for pool by way of 8 Ball).

The gift from a very kind colleague of an XBox One opened the door a crack. Leaving others to their Roblox, I started out with Red Dead Redemption, which had intrigued me since finding out about it when it launched during my non-gaming years.

Red Dead Redemption is simply miraculous. It was my first experience of a game both vast in scale and minute in detail. The realism of the trackways through a turn-of-the-century Western landscape join with a tangible plot and rendering which is good enough to create a reality which compels. I’m only at the start, but I could watch it for hours.

Red Dead Redemption is hardly suitable fayre for the brood, however. By chance, my youngest and I stumbled on Unravel. This game is a natural successor to Myst in its beautiful rendering, disguising simple physical gameplay. Four factors elevate it way beyond its precursor.

The visuals are simply beautiful: lush, detailed, small in scale but big in realization. The real-world physics remind me of the peerless Thrust. The complexity of the puzzles verges on the fiendish.

But everything is trumped by the profoundly evocative atmosphere. Scandi hygge pairs with a gently compelling soundtrack to tell an achingly nostalgic story, almost melancholic in its depiction of past childhood.

So that’s me, at least for the moment. I can’t see myself wanting any other games until I’ve completed these two. While each has a sequel, only Red Dead 2 seems to have a decent reputation. So I think I’ll be set for some time. Though the lure of Hadrian’s Wall in Assassin’s Creed is beguiling.

I read The Weirdstone of Brisingamen when I was ten. Like many others, fantasy was a passion for a time, and, along with inevitable Tolkien, The Weirdstone stuck with me for some years. I remember little about it now, other than a sense of darkness.

Many years later, I was lucky to attend a talk given by Garner. It was a storytelling event held in the reconstructed Anglo-Saxon hall of Thirlings at Bede’s World in Jarrow. My older brother had been a Garner fan for years, but all I had was that vague childhood half-memory.

Garner’s talk was like nothing I had ever seen or heard. It told of his experiences writing his imminent short novel Thursbitch. I’ve never before or since heard someone hold the room so completely, as he dug us down into the earth of his new book.

After reading Thursbitch, I sought out all of Garner’s fiction. The Weirdstone didn’t stand up wonderfully well to an adult reader, and the other two children’s books were similarly unmemorable. The Stone Book Quartet gave a glimpse of what was to come, in its peerless evocation of childhood.

Then came The Owl Service. I confess almost to having my breath taken away while reading this short book for the first time. The Welsh valleys act as a protagonist, perhaps emphasized to this English reader in the contrast drawn between its middle-class English and working-class Welsh characters. Class, and cultural and intellectual hatred, soak every character, and I had my first glimpse of Garner’s horrific perception of his childhood and adolescent self, and his unwavering certainty in his own intellectual superiority. It’s a tough read, but a compelling one.

The Owl Service is rightly remembered, of course, for how it sinks its landscape and protagonists through the fabric of history and nature. This is Garner’s genius. Even in this early work, he creates echoes between ages, made concrete in elemental man-made objects. His is the liminal world between the natural and the unexplained. And that threshold is one of horror.

The Owl Service is everything which Jacqueline Wilson fails to achieve. It’s a children’s book about misery and anger. Red Shift takes these themes to another level for a slightly older audience.

Red Shift strips out the bucolic setting of The Owl Service, rejecting its beauty for the industrial hinterland of Garner’s beloved Cheshire. The themes of class and intellectual snobbery infect one of the three narratives which run in parallel and intersect in this disturbing, upsetting, and raging work.

Each strand distances itself from the reader quite deliberately in its characters’ language. Modern-day irritating intellectualism, seventeenth-century Cheshire dialect, and US military jargon in the mouths of Roman legionaries all unsettle.

Red Shift’s multiple narratives allow Garner, quite astonishingly, to place massively greater emphasis on the echoes and touches between ages, again channelled through the ancient man-made. However, because of the explicit narratives in different periods of history, this totemic relationship is set free to become vastly more impressionistic. The conscious uncertainty amplifies his central theme of inter-generational historic emotional resonance to an astonishing degree, and the book chills with dread.

The splitting of the narrative into three allows and requires Garner’s overweaning evocation of place to sharpen and localize from The Owl Service’s Welsh Valleys to Red Shift’s Mow Cop. Again, the historic and man-made act as a conduit between ages, in both the land and the ancient axe-head, which inherits the mantle from The Owl Service’s megalith and spear-head.

Red Shift is despair. “Not really now not any more.”

As described, I read Thursbitch, last in this undeclared trilogy, before the other two. I was still buzzing with dread and fascination from Garner’s mesmerizing talk.

Thurbitch pares Red Shift’s three parallel narratives down to two. The present-day protagonists have continued to age with Garner, and death is constant. While hateful snobbery persists, Garner’s characters recede, and his land looms ever larger. The Owl Service had the iconic Welsh valleys. Red Shift had a local curiosity and landmark. Thursbitch gives its name to the novel, but is otherwise unremarked, to the best of my knowledge. The landscape’s lack of historic or monumental significance allows the novel’s setting to drop all cultural baggage and creep its own quietly screaming path. The horror of The Owl Service’s myths, and of Red Shift’s history, becomes purely of Thurbitch’s earth.

Again, the ancient man-made bridges the narratives and the ages. The seventeenth-century Cheshire dialect, crisp and lilting in Red Shift, becomes rhythmical and beguiling in Thursbitch. While the former’s characters battled in extremis, the latter’s chthonic archetypes mutter and lilt, “think on”.

Treacle Walker was up for last year’s Booker. To me, it was the faintest echo of the unflinching horror of this most English trilogy. If Garner had won, in my mind, it would have been in recognition of these unforgettable books.

Walk and do. Walk and do. Walk and do till all is done.

My esteemed colleague Leigh prompted us to embrace the colour green to raise awareness of World Mental Health Day this Monday just past. This resonated with me. While currently owning no green clothing other than my socks, I thought writing, not wearing, might contribute.

I don’t know where it started, but green’s always been my joy. Perhaps Dad’s insistence that green was his lucky colour initially hooked me. He would wear a green tie on occasions when he hoped for luck, along with green cufflinks. I tried that with underwear on Newcastle matchdays. It worked as well as could be hoped.

These days, a couple of tangential greens make me happy. We were invited into Greenhaugh Primary School to hear the children read and sing for the first time since lockdown. I’ve written about Greenhaugh before, and while all of our mental health undoubtedly suffered in the fight to save it, the fight was worthwhile, and this visit lifted my spirits.

I used to play football extremely badly. Moving into the middle of nowhere twenty years ago (not to mention anno Domini) made this harder, and it fell by the wayside. I didn’t realize how much I missed participating in a competitive game until I was asked to join the mighty Blackcock pool team. Perennially bottom of the league, the green baize nevertheless calls me, and I love the game.

But it’s green in its common meaning which lies at the heart of the matter, of course. Someone once commented that one’s happiest childhood memories are never indoors. Camping and narrowboating certainly rank among mine.

Although I grew up a city boy, my wife persuaded me to move into the green yonder. While we worried that we both might struggle, in truth we’ve never looked back.

She had form. A short break as students had us walking a portion of the Ridgeway, from Swindon to Avebury via Uffington. Sunburn and blisters didn’t dilute the joy of a slow journey through the English countryside.

We doubled down on narrowboating, and have been on the canals several weeks a year for over fifteen years. Landscape and nature are balm. Seeing HS2 carving swathes through the South Oxford summit was one of my saddest memories.

When I commuted, a random tweet made me count up the number of waterways I crossed each day. That commute was of course a burden, but it was a beautiful one, almost running the entire course of the North Tyne. Getting Max the dog soon after we moved meant four daily walks, and inspired another.

Many years of the Big Garden Birdwatch partially scratched the perennial itch for collection and documentation (I still delight in nuthatches, siskins, and the very occasional long-tailed tit), but I was happiest to compile a list of every wildflower I could see in the verges of our single-track road.

Mum said she once heard a professor say on the radio that she firmly believed that if you take an interest in the natural world, it will hold your interest for the rest of your life. Maturin would agree.

qgis2web has now been downloaded from the official QGIS plugins repository over half a million times. It’s the sixth most downloaded QGIS plugin of all time. This is a short piece about its history.

qgis2leaf and qgis-ol3

In 2014, Riccardo Klinger created qgis2leaf, which exported QGIS projects to Leaflet webmaps. Around the same time, Victor Olaya created qgis-ol3, which did the same, but exported to OpenLayers webmaps. I got involved with qgis2leaf, as we needed it for work (adding support for WFS, as I recall). I’d never coded in Python or used Qt before, nor used git, but as languages and source control systems go, they’re easier than many.

Riccardo and I continued to work on qgis2leaf intensively, adding considerably to its functionality. It was a really exciting time for me. Learning to code in Python (albeit not at all well) was fine, but what I’d never really experienced before, as an amateur coder, was having the public use code I had written. It was a big step, but a good one, having to deal with issues reported by users I had never met, and never would.

qgis2web is born

In 2015, Paolo Cavallini, now chair of the QGIS Project Steering Committee, got in touch with Victor, Riccardo, and me to suggest that it would help users if qgis2leaf and qgi-ol3 were merged. Both did essentially the same thing, but using a different Javascript webmapping library. We all agreed, and work started on qgis2web. Huge thanks are due to Paolo for the suggestion and support.

qgis-ol3 had a GUI with more functionality than qgis2leaf. Specifically, it had a preview window. This was important, as there was a huge benefit in migrating qgis2leaf to the same GUI. It meant that the proposed merge had advantages for qgis2leaf (as was), as well as for end users. This certainly motivated me, as did Victor’s elegant Python in qgis-ol3.

I lifted all the qgis2leaf code, placed it alongside the qgis-ol3 code, and slightly altered the GUI to allow users to switch between the two. It worked, but really was in essence two plugins side-by-side, with all the duplication that implies.

Improvements and tests

Over time, common functions from the two original plugins were deduplicated, and the code became a little less unwieldy. However, I worried about releasing buggy versions–it did happen once or twice. I had become aware of automated testing, via a service called Travis, hooked into Github. The idea is that, every time a change is made to the source code, a suite of tests is run to ensure that no regressions have been caused.

I thought for some time that this would be an incredibly useful service for qgis2web. With so many variables (layer types, renderers, styles, options), it was already impossible to be rigorous in checking manually that nothing was broken.

With a lot of help from Akbar Gumbira, we finally got it all up and running. It fundamentally changed the reliability of qgis2web across versions, even though wrangling code and reference output files can sometimes be frustrating.

Collaboration and support

Paolo remained a big supporter of qgis2web. As part of a Faunalia contract, he commissioned Nyall Dawson to work on the plugin. This was brilliant news, but the idea of a QGIS core dev looking at my code was also somewhat daunting. Sure enough, Nyall fundamentally refactored the entire plugin, decoupling the GUI from the export process. His work remains to this day, and the plugin is the stronger for it. I must also give credit to both Jürgen Fischer and Matteo Ghetta for their help during this period of development.

It must be said that this work did expose some inherent limitations of qgis2web for some projects. At the end of the day, QGIS, Leaflet, and OpenLayers are very different pieces of software, and the area of functional overlap between the three is significantly smaller than the functionality of each individually. If there were one constant frustration in working on a piece of software used widely by strangers, it would be the struggle to manage expectations, especially given that many users have strong GIS knowledge, but less strong knowledge of the web and its inherent performance limitations.

QGIS3

The changes from QGIS2 to QGIS3 were enormous. Python was upgraded from v2 to v3, and Qt from v4 to v5. In addition, so many things in the QGIS API changed. It was a massive job to produce a compatible version of qgis2web, and it would have been basically impossible for me without the Travis testing environment. However, it all got done, and the concomitant tidying and refactoring of the source code was a massive collateral benefit.

There have been two major version upgrades to OpenLayers and one to Leaflet during qgis2web’s life, and another OpenLayers one is imminent. The work never stops.

Burnout

After the big push to make a QGIS3-compatible version, I did slow down. The eighty-twenty rule kicked in, with remaining issues being either less significant or harder to solve. My attention was also elsewhere (but still using QGIS!). As a result, development slowed down significantly.

Things have picked up again, but I doubt I’ll work as intensively on it again. It’s frustrating, but we’re all human.

Significant help

Many other people have contributed to qgis2web. I specifically want to mention a few here, and apologies to those I miss out through forgetfulness.

Luca Casagrande coauthored qgis-ol3 with Victor. He has continued to support qgis2web, including fixing some problems which totally had me stumped. He also worked on qgis2web with Paolo at a QGIS developer’s meeting.

Nathan Woodrow wrote a module from scratch to convert QGIS expressions into Javascript, and hence make them available in qgis2web. It’s a beautiful bit of code, all the more impressive since I get the impression that Nathan doesn’t really work with Javascript.

Thomas Gratier wrote an OpenLayers plugin from scratch to allow searching by layer attributes. Handling the diversity of possible search sources was a challenge, which he overcame effortlessly.

Per Liedman wrote a Leaflet plugin from scratch to support multiple symbol layers. He also helped a great deal in implementing Leaflet.VectorGrid to support vector tiles.

Iván Sánchez Ortega has been a huge help also in Leaflet.VectorGrid and more generally in Leaflet and vector tile matters.

Ola Kovalev made significant contributions including developing OpenLayers clustering from scratch. Ola is notable in being happy to learn and then implement tests alongside her work, going through the same occasionally frustrating learning process that I had.

Paul Dos Santos implemented overall map styling from scratch. One of my frustrations is not having taken his work further–a fairly simple task, but one which never seems to get to the top of the list.

Martin Boos wrote and maintains the QGIS plugin Vector Tiles Reader, without which none of the vector-tile functionality of qgis2web would be possible.

Ed Boesenberg developed multi-unit measurement in OpenLayers, and also deserves massive thanks for his constant help in answering queries on the qgis2web repo.

New maps

For the first time since it was created, qgis2web has a new output format. Mapbox GL JS excels at rendering vector tiles, and has many other key advantages over Leaflet and OpenLayers. However, it should not be used indiscrimiately, as the other two libraries still have their advantages. In addition, the currently implemented feature-set for Mapbox GL JS in qgis2web is much, much smaller than that implemented for Leaflet or OpenLayers. This will improve over time, but is unlikely to be quick. I hope to write a separate piece investigating the core differences between the three libraries, but that’s a job for another day.

Huge thanks are due once again to Victor Olaya, who has developed bridge-style at GeoCat. It does all the style conversion work, and I can see qgis2web offloading its OpenLayers and Leaflet style conversion code to it in the future. Other projects will then be able to make use of the routines.

A quick note, by way of clarification. Mapbox, the company, provide many services. Mapbox GL JS is one of their products, and is completely open source. qgis2web only exports standalone Mapbox GL JS maps. It doesn’t use any Mapbox services–no hosting, no styles, no data. It’s key to refer to the export format as Mapbox GL JS, to keep it distinct from Mapbox the company and its ecosystem of hosted services. Many thanks to Vladimir Agafonkin for putting me in touch with Kathleen Lu at Mapbox, to ensure that Mapbox are happy with the approach.

Where next?

Developing more functionality for the new Mapbox GL JS exporter is top of the list for me, as qgis2web has reached some level of maturity with its Leaflet and OpenLayers exports (though one day we might get layer groups working properly). There will no doubt be many issues with Mapbox GL JS exports, so handling those will probably also take priority.

More important than these, however, is making qgis2web’s capabilities more readily transparent to the user. We need massively improved error reporting, with an emphasis on feedback on unsupported QGIS features in an export. Alongside this, a compatibility table would also help users decide which export format to choose. With the launch of the new Mapbox GL JS exporter, this has now become much more critical.

And documentation, obviously.

Bellwether

Beyond the code itself, I’m extremely proud of qgis2web for a couple of other reasons. One is having turned code into software–in other words, supporting users, implementing testing, and also promotion, which is a big part of the project. Documentation still lags horribly, but that’s a badge of honour for an open-source project.

Above all, though, I’m proud that qgis2web represents consistent growth in a sector I love–open-source, provider-agnostic software. To my delight, there has been no real plateau in growth:

Despite the recent Google Maps price hike, the trend away from non-provider-based webmaps sadly continues. qgis2web tries its best to help those who want to remain truly independent and in control.

Walking further out of town on the road where I work, you climb a hill to some woods. You can avoid the main road and find more interesting ways to get there, but my walk without Max today only really started after the climb. Limited lunch time meant we used to drive up each day, until my guilt at unnecessary driving pushed us to alternative routes, and this walk was consigned to our past together.

An unprepossessing start, parking in a muddy narrow layby, and tiptoeing nervously along a short stretch of fast rural road with no pavement, leads to a wide track. Straight as an arrow, it runs at right-angles to the road, into the forest. At the junction is a gloriously overgrown cottage, with improbable bamboo hedge.

I’ve never liked forests much. The trees obscure the land and the sun, and one’s horizon is shortened. Pragmatism won the day, though, as tree cover has its benefits in the English climate. Max was black and tan, so as well as protection from inevitable rain, some shade was also welcome on our summer walks.

While the wide track marches on into the distance, various openings in the stone wall to the right grant access into the forest itself. One can pick out any number of routes, but the functionally redundant parallel course to the track always suited us best.

Where this almost-path rejoins the main track, a wider forest path slopes down to the right. Taking various measures to avoid incursions of water, one passes what I can only imagine are four redwoods. What these giants are doing among this otherwise commonplace English plantation is a mystery, and a beguiling one.

The path narrows as it skirts the private grounds of the large house at the heart of the woods, and then slopes further downhill. This path seems ancient – a wide holloway, again very straight, ending at a tiny stream.

I paused here, as here is compensation for the lack of view so far. Through a fence is open countryside, of the most beautiful rolling English kind. Max and I would always pause here, I to smile as my heart was lifted by the view, he impatiently. I never found out what horizon was revealed, and this view remains to this day precious to me, as a shared memory with him, silent and content. Today it’s just me.

Across the stream, the path flattens out, as it gently curves round to the left. This is more traditional forest pathwork, not the ageless sunken path we just descended. But to the right, a fork slopes sharply down.

We must have walked on past this fork hundreds of times, but one day, taking a break from over-cheerful enthusiasm from a software trainer, we turned right. The gradient rapidly increasing, the path became narrower and narrower, steeper and rougher, descending a conifer-root stairway, and nearly meriting the sit-and-slide approach.

As a younger man with a young mongrel, evidently largely collie, I scampered down this path. For years after, I couldn’t find it again. But three years ago, Max and I being older and more nervous, the ground slippery with falling rain, and with my nine-month-old son strapped to my chest, the precipitous nature of this descent was all too apparent.

I wasn’t going to turn back, though. Max and I had not been here for maybe ten years, and I wanted him to see it again. Sentimental thoughts of sharing it with my son, and even of this being a boys’ trip out, made me slither down.

My wife and I do everything together, with our family. These weekday lunchtime walks with Max were my time without her, without talk. Max and I would cover the ground, and see where our journey took us. I wonder if one day I’ll take my daughter here. She’d love it, but life gets in the way, and there is a little part of me that wants to keep this walk to myself. Me and Max. Today it’s just me.

The worst of the sheer path negotiated, this break from IT training yielded the most amazing surprise. An unexpected river, with yellow sandstone cliffs, and a sandbank shore in the bend. This felt like our own discovery. Today was only the third time I’ve ever been there, and it still takes my breath away. I jumped onto an island stone to be surrounded by the flow, and saw a tiny fish dart away.

Water makes me think of my dad. He was in love with the sea, but moderated his passion to rivers and canals in order to devise family holidays my mum would not flat-out refuse to join out of terror. Today would have been his 89th birthday. I had planned to go down to the abbey to sit for a moment in my perplexed atheism and remember his faith and love of church music. This felt more like common ground.

Back up—climbing easier than descending—I rejoined the long left-hand bend of the main path. Each of the three times I’ve done this walk, this long section has tested my mettle. I’ve no map (ironically). I don’t want a map. The first time, I wanted to see where paths took me, and each subsequent time, each separated by years, I long to see whether something in my sense of direction keeps me right.

Though devoid of landscape, nature entertains me today: a squirrel, a robin, and hundreds of thousands of wood ants and their anthills half my height are the distraction I need to settle my nerves.

Eventually, my faith is paid off. The continued gentle curve to the left felt like the right direction, and rather splendidly here we are, at the far end of that first arrow-straight track. Intriguingly it’s a crossroads. I’m not sure I can ever foresee an occasion when I’ll have the time to go straight on or turn right. Left it is.

I’ve been worrying about my age and health a little recently, for the first time. Being closer to fifty than forty, coupled with an increased rate of consumption of my personal vices (fat, salt, chocolate, and alcohol) has led to a feeling of slowing down. I’ve become aware in the last couple of weeks that I now make a noise when bending over, straightening up, or standing up out of a chair. I think I’m safe finally to identify as middle-aged.

For years and years, I defined male middle age as “that quality which compels a man to describe the death of former pets to younger women and girls”. This walk without Max confirms it.

To combat my lack of condition, I pick up the pace. I’ve always run up and down stairs, and had also become upset to realize that I’d unthinkingly stopped. So, the pace is upped when I can, and while the drink will remain, perhaps I can stave off the inevitable.

A little way along this unwavering mile, the manmade glory of this walk rears up to the left. A magnificence of Scots baronial has crossed the border. How can a house need so many chimneys, and so tall? And who doesn’t love a turret? Presumably this is the house whose owner planted the redwoods on the far side of the grounds.

Pound back to the main road. Turn right down the hill. While Max and I would by now be in the car, both drivers and walkers are now granted the most magnificent view of this Northumberland market town. No castle, but an abbey and old gaol offset the white-elephant empty hospital and inevitable vast supermarket.

Nothing else remains. Nearly back at work, the far side of the road offers the steep muddy footpath which starts what became Max’s and my later walk together. Its close forest perspectives are occasionally interrupted by beautiful views of pasture, and intriguing cliffs and ponds possibly bear witness to historic quarrying.

This walk with Max was what told me he was dying. He was slowing down, and I wanted him to be able to walk on the flat. I also wanted him perhaps to revisit a walk of his youth, in the days when he and I got to know one another. We got out of the car, and he stood still. He stood still.

The vet told us he was in a great deal of pain. He was gone by the end of the week.

We have Tixall and Kinver now. But somehow I don’t want to take them on this walk. On this map without a map. My walk without Max.