Hello world, and welcome to my corner of the web. This is where I write words about what I'm working on, and post photographs of things I've seen.
I'm a software engineer at the Wikimedia Foundation,
and so of course my personal website is a wiki (running on MediaWiki).
In my spare time I volunteer with WikiClubWest to work on Wikimedia projects, mostly around my family's genealogy and local Western Australian history (especially to do with Fremantle).
I try to keep up with issues on all the things I maintain (but usually fail), as well as listing the software that I use.
I try to find time to work in my workshop on various woodworking projects. Recently, that's been focused on building a metalworking bench, and will soon be about a set campaign-style drawers that's in the works. I've a good-sized workshop because I don't have a car.
Travel features in my life, not because I really hugely want to go elsewhere but because I just do — and also because then I can do some interesting mapping on OpenStreetMap, and take photos for Wikimedia Commons. Sometimes I ride my bike to get there, or walk, but more often it's planes, trains and ferries.
Calibre have added an AI thing they're calling "Create your own adventure": a writing game, or "interactive storytelling with an AI managing the world you create."
This is a little different from the usual run of calibre features. It started out as a way to test the AI backends in calibre but turned into something fun to use. As always with AI features its completely optional and not even loaded unless you explicitly run it.
Calibre already has a reputation for being a pretty strange bit of software, but it's well-liked by lots of open source people because it is really good at converting ebooks to various formats, and managing large libraries of books, and moving them on to ereader devices. I'm not quite sure anyone's wanting it to generate AI slop.
Personally I only really use it (or rather the built-in ebook-editor tool) for debugging EPUBs.
The 2026 ALIA West conference was held on a Thursday and Friday in September 2026.
I'm not a librarian, and was attending as a Wikimedian
partly to help with the idea of Wikifying a Conference.
The venue was the State Library of WA,
and it's good that the café there opens before the library itself does,
because we got an extra hour of talking time.
(It's a clique but so often true that conferences are about talking to people more than listening to presentations!)
It was good to meet the WA Military Digital Library folk.
There were two of them, representing a team of about ten.
WAMDL uses MediaWiki and it was great to hear more about what they think of it (they like it of course).
It turns out they're running from a self-hosted server at the home of one of the organisers!
There's a possibility that we'll get organised and start a chat channel for Australian history wiki people, because there are about five such wikis currently operating (plus a few defunct ones). Time for an Aussie MediaWiki users' group!
The Welcome to Country was by Freda Ogilvie, a retired school principal from the north west.
Trish Ryder's keynote about the 1940s Carrolup artworks was very moving.
The John Curtin Gallery is using the term 'once known' rather than 'unknown'
to refer to children who's names weren't recorded alongside their art
(it's also the name of the exhibition that's opening soon at the Cloisters).
An autistic LIS student gave a lively talk about what it's like to be neurodivergent and attend conferences. It sounds like the library world is fairly supportive — perhaps there's some sort of overlap between librarians and neurodivergence?
Curtin have set up a mature age student club, run by the library.
I think they need a wiki club (although really I'd rather Notre Dame had one, because they're closer to me).
Two metadata nerds from UWA presented about UWA Collected.
I spoke to one of them, Jonah Highman (a self-confessed metadata nerd),
and it does sound like some sort of UWA-Wikimedia communication could be fun.
The Two Centuries of Chinese Heritage in WA project looks terrific (and huge).
I found a letter that I added to my family archives: archiveswiki:Letter to H. Gregory 20 October 1920.
UWA Collected uses the Recollect software, which seems to be growing in popularity these days.
FamilyHistoryWA presented about their on-going ghost town project.
They're a good bunch for planning (like perhaps many organisations run by retired nerds).
Two years of thinking before they created their first database record; they've now got 140,000, although none are public yet.
They're only including people relating to any of the 950 "ghost communities" that they're tracking.
I'm not sure how much interlinking they're doing to other sites such as Wikidata or WikiTree.
There were two ghost communities on Rottnest! I'm guessing one was the Army camp, not sure about the other.
Some currently-disappearing towns are being included too (like Sandstone).
The are creating records of every person, and leaving disambiuation up to future researchers.
Alice Woods and Gnangarra talked about Wikimedia Australia and the partner projects and other things (and I even spoke briefly about MediaWiki, and Noongarpedia!). In none of the sessions was there much time for questions, so it's hard to know what the library folk think of wiki stuff. Hopefully they're keen.
There was plenty more talked about, all very interesting, but I was trying to listen more than get distracted with note-taking.
I don't think I'll use it (I've got my own bloggy wiki, of course), but it's nice to see MediaWiki experiments for things. They seem to be using DynamicPageList4 for the listing of blog posts (why are we up to rewrite number four with that?), and perhaps some custom code for creating user pages as soon as accounts are registered.
The AI agent breakout in Germany was detailed in a report shared exclusively with Reuters, opens new tab by a group of researchers including Sydney Von Arx, CEO of AI safety nonprofit Nightingale, and Cormac Slade Byrd, a quantitative trader-turned AI researcher. They uncovered the activity in late August while scouring the internet for signs of unauthorized AI-agent behavior, they told Reuters.
The pair said they found more than 15,000 edits carried out by AI agents on a German-language wiki site, DseWiki, that is geared toward programmers and accepts communal edits along the lines of Wikipedia.
This new phrasing that seems to be taking hold is pretty annoying. AI doesn't "escape containment" or "break out", it's just a set of computer programmes running on computers, doing stuff that people want it to do.
I hope the DseWiki manages to recover. It looks like they're using UseMod or some other old-school wiki engine. I think on Freopedia we'll have to be sticking to the current approach to new accounts of basically requiring people to turn up at the Buffalo Club in person! The clankers are too good at getting past any CAPTCHAs and things. Even requiring email validation isn't always enough (although that does reduce the rate to at least manageable levels).
Thankfully the person ID is retained, and it's just the tree ID that's changed. The trailing text (sarah-constance-leake in the above) can still be anything, but whereas it used to be able to be left off completely it now has to be something (at least one character).
And if you prefer to install and run your own content management software or create your one yourself, you can do that too: https://indieweb.org/CMS
If you do create one, watch out for spending more time coding than writing.
Or code away! It’s your site, code, write, mix it up as you wish.
It's an old clique I guess, that programmers spend more time tinkering with their blog than actually writing anything on it. That's certainly true for me. But then, I think I spend more time writing on other places these days, mostly non-chronological ones (they feel a bit more useful). But picking a CMS and getting a site online used to be great fun! I'm not sure that it is as exciting, on the modern corporate internet. The indie internet still exists though! Everyone should come and do whatever they want on it.
W.A. Greens leader Dr Brad Pettitt put a reel on Facebook, pointing out the danger of vehicles intruding into the shared cyclist/pedestrian path, that runs along the railway line at Victoria Quay.
I pointed this out years ago, and a couple of times since, and Fremantle Ports promised to put stoppers on the road, but then changed their minds, claiming it would become difficult for motorists to park. What utter nonsense!
The shared paths is very popular and can be extremely busy with passengers of cruise ships, so vehicles sticking out onto the path are creating a serious safety hazard, that Fremantle Ports really needs to address.
But then that same day, Fremantle Ports said they'll do something!
It took years, and about four blog posts here on Freoview, to finally get Fremantle Ports to commit to safer parking at Victoria Quay, so that the shared pedestrian and cyclist path along the railway line, will no longer be blocked by the backs or fronts of vehicles.
Fremantle Ports just contacted me to say that a contractor has been engaged to install wheel stops. That work will commence in October.
The wharf construction works are still ongoing at the ferry berth.
It feels like these fences have been up for a long time, but I guess pulling everything apart and replacing beams and maybe even piles is a pretty big job. They do seem to be reusing lots of the original deck beams, putting them on top of huge new steel bearers.
I guess the ferry terminal stuff that goes on top will take less time, once there's an actual surface to build on again.
After a coffee (and steaming hot bread and butter pudding) I hopped on the bus and headed out. I got off at Strickland, and probably should've got off one earlier but forgot what I was doing. Anyway, it's a nice spot to start walking because no one ever seems to get off there and so there's a feeling of solitude as soon as the bus has gone around the bend.
I have various wikis, into which I dump stuff as I think of it, or digitise it, or find it, or whatever.
It's mostly a good feeling of eventual organisation,
like things can be safe for now and at some point put in their right places.
By date, or media type, or geographic location.
Then, all the wikis get backed up (to three different storage systems in three different places, one of which is not managed by me),
and the process of backing up helps to remind me of things that need organising.
It's nice watching the rclone progress.
I'm slightly worried sometimes about what'll happen to them in the end,
that some poor sod will be left with some gigabytes of files and a few large XML files,
and may well just give up and throw it all away.
Possibly a Git repo full of Markdown would be better (certainly would be more fashionable),
but that doesn't deal with the images and PDFs and whatnot.
I've written before about websites as collecting endeavours. That there's some good aspects to features like MediaWiki's insistence on not being able to embed foreign images. Keeping things together within one place, and pulling in things that we find elsewhere, can build a more coherent and resilient website. Having copies/duplicates/multiples of things, spread out and maintained by completely separate people or organisations, increases the chances that things will survive. The web is an ephemeral place.
Some sites, such as many of those on the fediverse, don't like archiving. They have very clever systems of interchange that does result in multiple copies being hosted in different places, but they also have good systems of those copies being strongly linked — so they can all be deleted if the original is. In the more disperse indieweb world that's not really how it's done: some RSS feed readers, for example, will keep all copies (following some retention policy perhaps, but that might be "keep forever" — that's certainly how I do it for a bunch of sites I follow).
This is an image loaded via InstantCommons (a thing I uploaded recently to Commons).
MediaWiki does allow embedding foreign images in one specific way: if they come from Wikimedia Commons. This is called InstantCommons, and it's a very cool feature. It pulls in the image and its metadata, and links to the thumbnail hosted on Wikimedia's servers. This has a few issues, such as the image being able to be changed at any point without the local wiki knowing, and the fact that it makes the local wiki's collection dependent on an external service. It also puts quite a lot of load on the Wikimedia servers, which is probably fine if it's for a purpose that's aligned with that of Commons (that's a tricky thing to figure out of course).
A better system would download the thumbnail to the wiki's own storage, and serve it from there. But I also have been wondering about a system of copying the image (and its metadata) into the local wiki, to remove any dependency on the external service. That uses even fewer of Wikimedia's resources, and makes the wiki self-contained. The trick to it I think would be to make it quick and easy to pull the image in (and perhaps it could work with any website), and also maybe to view any changes that had been made since it'd been pulled in.
Then the site would be more of a standalone collection — perhaps annoyingly indivisible.
The (nerd) world seems to be much more interested in ActivityPub than RSS these days,
but RSS feels so much simpler and also in a way more powerful.
It's sad that even podcasts are abandoning it.
It's slightly ironic that I came across the post in via my RSS feed reader (via the IndieNews feed), where it didn't have any body and just looked like this:
I still hope to implement Webmentions on my own blog one of these days.
Interesting that Miraheze are looking to disable the DisplayTitle extension
because of performance concerns.
I've recently been removing it from some wikis, mainly because people have gotten confused about what a page's title actually is.
There are also a bunch of places that used to make use of DisplayTitle that now have templates
that basically follow redirects to figure out titles, and it's more obvious what's going on when those are used.
