I missed this when it was first published.
Friday, 30 December 2016
Saturday, 10 December 2016
Here I am giving a talk on translation, transclusion and linked data at the XML London conference earlier this year. Please enjoy, because I haven't dared to watch it ;-)
Tuesday, 6 December 2016
Interesting, especially when you get past the bumf at the start of the report and start browsing the maps. New Chinatown! I did not know.
The spatial construction of civic identities: A study of Manchester’s linguistic landscapes
Wednesday, 30 November 2016
Two years on (6,000 miles or so?) and this chainring is still going strong.
I seem to need to replace the whole drive train every 2000 miles or so
Monday, 21 November 2016
Tuesday, 1 November 2016
Monday, 31 October 2016
Sunday, 23 October 2016
Friday, 21 October 2016
Monday, 10 October 2016
Sunday, 9 October 2016
Sunday, 2 October 2016
RIP Prof Katharine Perera. Taught me sociolinguistics.
Saddened by the passing of Prof Katharine Perera, an inspiration to generations of students and staff @OfficialUoM @UoMSALC @UoMLEL pic.twitter.com/Admg7yX3rH
— Multilingual Manchester (@mlmanchester) October 2, 2016
Saturday, 24 September 2016
Monday, 19 September 2016
Sunday, 11 September 2016
Machine Translation
This is interesting, although it skips over some of the details.
Kind of raises the question though, why is machine translation still so rubbish?
Okay, so much more training data is needed. But maybe also machine translations enter an uncanny valley as they get better.
Machine Learning is Fun Part 5: Language Translation with Deep Learning and the Magic of Sequences
Thursday, 1 September 2016
Wednesday, 24 August 2016
Interesting article on shift in T/V usage, i.e. informal and formal 2nd person pronouns.
Across the western world the formal "you" is losing ground. From the archive: https://t.co/ffYqpA1keh pic.twitter.com/Mu6f4APLJj
— The Economist (@TheEconomist) August 23, 2016
In tech translations I noticed a slight shift towards T forms in Spanish over the years. I usually had to change it back to the V form for consistency. In German I noticed a shift from a clipped, telegraphic style towards a more conversational style, but still Sie not du.
Sunday, 14 August 2016
Vigenère
The Vigenère cypher. It turns up in Lauren Child's Ruby Redfort books, which daughter #2 is reading, and in Red Shift by Alan Garner, which I've just read. So I wrote this little library to help me encrypt and decrypt it. Probably won't be of much interest to anyone: the Vigenère cypher is almost useless for modern cryptography, and I've written it in XSLT, which most programmers seem to hate, but as a linguist I have a perverse liking for it.
janiveer/Vigenere
Saturday, 13 August 2016
Interesting. Twitter as a tool for linguistic analysis, again.
Text analysis of Trump's tweets confirms he writes only the (angrier) Android half
Friday, 12 August 2016
Edinburgh Bike Co-op Manchester Closing
Ah well.
Edinburgh Bike Co-op Manchester Closing
I haven't shopped there for ages to be honest. For parts and gear I go online. For expertise and fixes I'd go to the Bicycle Doctor or Harry Hall.
I think EBC occupied an interesting position. They sell goodish, mid-range off-the-peg bikes: not pile 'em high, sell 'em cheap, but not custom titanium frame either. There always seem to be plenty of places in Manchester selling overpriced road bikes. I guess the local market didn't support their approach.
Tuesday, 9 August 2016
Friday, 29 July 2016
Thursday, 28 July 2016
Friends of Manchester Peace Garden
Manchester Peace Garden. Okay so it was a bit daggy, just a few scrubby bushes outside the town hall extension, but also included a statue called the Messenger of Peace. (Mancunians also called it the Pigeon Feeder.) All removed to make way for the redevelopment of St Peter's Square.
This group is campaigning to get the Peace Gardens and the Messenger reinstated in the medieval quarter
— the space between the cathedral and Chet's. (Manchester must have more quarters than a 4-dimensional hypercube now.)
I don't call myself a pacifist, but I think it's a shame when quirky green spaces and public art disappear from our cities, so I hope this group succeeds.
Saturday, 23 July 2016
Mancky
Music, art, architecture of Manchester, mostly concentrating on the 1980s / 1990s.
I couldn't find an overall navigation page, you just have to follow the links and see where you end up, like a psychogeographical dérive.
See if you can find the bit about the rooftop housing estate on top of the Arndale Centre, abandoned after the bomb.
I love it, it's fantastic. I felt quite nostalgic for the Precinct Centre.
Interesting to see the page about Antwerp Mansion. Manchester Ceilidh considered this for a venue. We were genuinely concerned that the building would collapse on us.
Friday, 15 July 2016
A colleague of mine picked up this info at a public meeting in Leeds. May be useful / of interest.
Brexit: What EU citizens living in the UK need to know — Kingsley Napley
Thursday, 14 July 2016
This year's Tour has been absolutely bonkers. I think they're trying to compete with our politics.
Tour de France: Chris Froome forced to run up Mont Ventoux after crash
Monday, 11 July 2016
Tour Highlights
Yesterday evening catching up on the weekend's Tour highlights. Chris Froome's new descending style reminded me of my friend L bombing down the hill from Brinnington into Reddish Vale. I don't remember that she ever punched a guy wearing a chicken suit though.
Watching Froomey's aero tuck
reminded me that by getting off the saddle and hunching over in the cockpit, she somehow managed to get under the barrier at the bottom of that slope without slowing down :-)
Wednesday, 6 July 2016
Friday, 1 July 2016
Thursday, 30 June 2016
Monday, 13 June 2016
Saturday, 11 June 2016
Didsbury Festival
Just paraded through the village in the pouring rain dressed as a Jedi knight, accompanied by daughter №2 (an Ewok).
Most of daughter №2's classmates were Rey. There was much fan theorising over Rey's parentage.
Daughter №2's friends summarised the theories for me. There was a big division between camp 1: Rey is Leia's daughter, and camp 2: Rey is Luke's daughter. (They can't both be true, because eww.)
Rebuttal to camp 2: Jedi (i.e. Luke) had to remain celibate.
Rebuttal to the rebuttal: Luke changed the rules (... or broke them?) when he founded the new Jedi order!
Alternative rebuttal to the rebuttal: Luke gave up being a Jedi to become a father!
I think I've got that right. I don't know. When I say summarised
you have to imagine half a dozen 8-year-olds all talking at once.
Tuesday, 7 June 2016
Friday, 3 June 2016
Thursday, 2 June 2016
Satnav
I have joked that the emergency services wouldn't be able to find you these days without their satnavs. Doesn't seem so funny now.
Cyclist died after three ambulances could not find Olympic velodrome
I once dutifully phoned the fire brigade about a fire on the transpennine trail. Similarly, they couldn't find me. Glad I'm not phoning for an ambulance
, I thought :-|
Location accessible by road, and I described how to get there as clearly as I could, I just had no clue what the postcode was.
Friday, 27 May 2016
Pop
A long time ago, I had the idea for using pop tunes for ceilidh dances. This file on my PC was dated 2004. Don't judge me for my music.
Jigs
- Sweet Dreams My L.A. Ex (Rachel Stevens)
- Practically anything by Mark Bolan
- Panic (The Smiths)
Hornpipes
- Lost Weekend (Lloyd Cole)
- Sheila Take a Bow (The Smiths)
- Spirit in the Sky (Doctor and the Medics)
Rants
- The Safety Dance (Men Without Hats) [1]
Schottisches
- The Chauffeur / Blue Silver (Duran Duran)
Reels
- You Spin Me Round (Dead or Alive)
- Don't You Want Me Baby (Thomson Twins)
- Love Will Tear Us Apart (Joy Division)
- Lay All Your Love On Me (Abba)
- L'Amour (Erasure)
- Road to Nowhere (Talking Heads)
Polkas
- Half a Person (The Smiths)
- Torn (Ednaswap, covered by Natalie Imbruglia) [2]
Mazurka
- Asleep (The Smiths)
Some of those rhythms have to be squished or stretched a bit to fit the beat properly. Also I haven't considered tune lengths at all. So for English dances you'd have to pick the dance carefully, or alter the arrangement.
[1] Credit where it's due: doing a rant to Safety Dance
was Rhodri's idea. If memory serves, he claims to have taught himself to dance a rant step on the platform of Piccadilly station while listening to that song on his Walkman — a kind of primitive personal music player, kids.
[2] Torn
is perhaps more of a 2-time bourrée I think?
Scenes from the Commute II
What's going up where the Beeb used to be?
"Circle Square". Student accommodation, apparently: Vita Student | Circle Square
Demolition of the Precinct Centre.
There is a corner of Fallowfield that is forever 2002.
Thursday, 26 May 2016
Googacle verdict!
🚨 GOOGLE'S USE OF THE DECLARING CODE AND SSO OF APIS IS FAIR USE 🚨
— sarah jeong (@sarahjeong) May 26, 2016
Saturday, 21 May 2016
Snuck
I remember Simon and I getting very excited some years back about the rise of non-standard strong verbs in English. Here's an old blog post with a recent update.
“Snuck” sneaked in — Stan Carey
It doesn't mention drag-drug! One of my favourites.
I wonder about the influence of German, Dutch, and Yiddish speakers in America, learning English, but being influenced by similar words in their native languages. So tragen-trug giving drag-drug? I don't know about sneak-snuck, but there is a Norwegian verb snike (to sneak) whose past tense is sneik or snek. Wasn't there a lot of Scandinavian influence in the upper Midwest?
On the other hand, Wiktionary says sneak possibly comes from the Old English snīcan (with a long ī), a class I strong verb: snīcan / snāc / snicon / snicen.
Thursday, 19 May 2016
Saturday, 14 May 2016
Following Sarah Jeong's coverage of this on Twitter.
In Oracle v. Google, a Nerd Subculture Is on Trial
Although I don't like the way this article divides the world into nerds
and normals
.
It's noticeable how news breaks much more quickly over on Twitter, including this case.
Friday, 6 May 2016
XML London 2016
This is my second (and probably last) conference paper on the tools and techniques I developed at Stanley. It'll be almost exactly a year since I left!
XML London 2016

Wednesday, 4 May 2016
Sunday, 24 April 2016
Ponyo
Cooking a stir fry for tea and singing a little nonsense song to myself.
Daughter №2 looks up from her Raspberry Pi: Congratulations, Dad!
(sarcastic voice) the Americans have ruined date formatting, and now you've ruined the theme tune to Ponyo!
Ponyo, Ponyo, you're an edamame bean
You're a little girl and you're small and green
I said it was nonsense.
Thursday, 21 April 2016
In total across the distances, the jet dryer spread 20 times more viruses than the warm dryer and more than 190 times more than the paper towels.
Using a Dyson hand dryer is like setting off a viral bomb in a bathroom
Monday, 11 April 2016
Friday, 8 April 2016
Statistical Learning vs Memorisation
Fascinating stuff. People who don't understand Arabic, but have memorised the Qur'an by rote, are better at spotting grammatical errors in Arabic than Arabic language learners who have studied the grammatical rules and inflections in question.
Statistical Learning and the Qur’an — The Ling Space
Thoughts:
-
To me this implies that our brains have built-in pattern-matching and rule-inferring mechanisms that perform better than our
conscious
minds. -
The abstract doesn't say anything about individual learning styles. I hope this isn't used as the basis of one-size-fits-all teaching in the future.
-
Maybe 48 isn't a big enough sample?
Tuesday, 5 April 2016
Flecty
janiveer/Flecty
Monday, 4 April 2016
mFiL 2015 Slides with Notes
The slides from my talk at the Manchester Forum in Linguistics conference last year are available here:
http://www.slideshare.net/SimonDew/linguistic-markup-and-processing-of-transclusion-in-xml-documents-notes
Linguistic markup and processing of transclusion in XML documents (Notes)

Sunday, 20 March 2016
Interesting that both Google and Facebook are so certain that today is the start of spring
. Other definitions are available ;-)
Spring (season) — Wikipedia
Thursday, 17 March 2016
Re: privatising the Land Registry.
Budget 2016: The UK must take every opportunity to strengthen data infrastructure — The Open Data Institute
Tuesday, 15 March 2016
Google My Tracks
Google My Tracks is deprecated. Looking for alternatives.
I guess the obvious one is Strava, but I'm not particularly interested in uploading my ride data or comparing my times with others... :-|
I like to save ride data (occasionally), mainly to see if I'm getting any faster or slower over the years. Also Google My Tracks worked out stuff like gradients and elevation, which my bike computer doesn't do.
Sunday, 21 February 2016
Map stuff. The article doesn't make it clear if afford
means have any money left after rent
.
Mapping the Hourly Wage Needed to Rent a 2-Bedroom Apartment in Every U.S. State
Thursday, 18 February 2016
Friday, 12 February 2016
Saturday, 6 February 2016
Thursday, 4 February 2016
Friday, 22 January 2016
Up-Goer Five
Here's a description of my PACBook project written on the Up-Goer Five website, which only allows you to use (what it says are) the 1,000 most common words in English.
Sometimes a computer needs to put a new word into a story, like maybe the name of a group of people, or a type of thing that you have made. But when the computer does this, it needs to use the right form of the word. Also, the computer may need to change the form of some of the other words in the story to fit in with the new word. This can be a hard problem if the story is written in a different way of speaking, like they use in other lands. I have written a computer thing which knows how to put the right form of the word into a story. Also, I have told my computer how to change the form of other words in the story, if it needs to. You need to mark up the story first, and mark up all the forms of the words that the computer will put into the story. I can show you how to do this.
I'm slightly cheating, using mark up
as a single lexical item.
Club Tales
Reading The Notion Club Papers 📚 on the train, which leads me to think about the club tales
genre. You know, Tales from the White Hart (Clarke), Tales of the Black Widowers (Asimov), and so on... usually short stories, a bunch of blokes get together and talk, which forms a common framing device for the actual tales.
Is it always just men? Are there any examples featuring mixed groups, or all women? Is anyone writing a c21st version?
Trying to work out the rules / conventions of the genre. They're club tales; the framing characters all know each other and meet regularly. Guests are permitted. No one in the club actually tells the story, at least not uninterruptedly; it emerges through conversation. The stories may be something that (allegedly) happened to one of the characters or maybe happened to a friend of a friend. They're tall tales or mysteries or have a twist.
Nothing happens to the members of the club in the context of the framing story. They just talk. But they're strongly drawn characters.