Council claims database pain forced it to drop apostrophes from street names A row in the UK has locals and council members at odds over apostrophes, and yes – this does actually have a tech angle. North Yorkshire Council recently decided to eliminate apostrophes from street names, because they and other punctuation marks simply don't get along well with geographical databases, apparently. This did …
I think - as the very last sentence in the article alluded - the point is that, particularly these days there should be no such limitations.
Oh, and good choice for the image illustrating the article, by the way. Personally I prefer the one in Harrogate to the one in York.
M.
I preferred the one on Cambridge Crescent, before the move a couple doors down into the Taylors building where it is now.
I'm probably biased, though ... it's where I had my first grown up date. I did like the new digs, too.
Hopefully "Tea at Bettys" hasn't become corrupted by the modern world. That would be sad.
Exactly.... I've never heard of anyone else having problems and working abroad there are loads of street names with "'" in them. Sounds like someone screwed up badly and is now trying to hide behind a misinterpretation of the rules... instead of confessing and fixing the problem.
I'm seeing "the designated name", which is not the same thing at the street sign.
I expect you have to pay people off when you change a name, though. Change it substantially.
An address that may be interesting and maybe confusing is Blaeberry Gardens, Edinburgh. That isn't a misprint here, but I expect it is in other places. In fact I think it's fifty-fifty whether a street with that official name gets its sign correctly spelled anyway.
The irony here is that the standard attempts to AVOID the name being altered. If only it stated ‘do not change the name, don’t add or remove contractions, don’t add or remove punctuation, don’t change the name by abbreviation.’ I’ll see if I can get the standard amended…
Maybe SAP could sell West Yorkshire Council a version with sensible input validation that isn’t just [A-Za-z]. I doubt the database is a csv file…. Although maybe that’s optimistic…
I think this sort of thing is egregious as it is casual cultural vandalism.
sensible input validation that isn’t just [A-Za-z]
I once had to work with an "artist" who didn't get this. His idea was to have a wall of moving message-board signs (96 of them), each one of which would show the title of an article from Wikipedia, chosen by following links on a "seed" article, and then links within the links, and so on.
Those moving message boards had character limitations of course so the artist sanitised the text sent to the boards (turns out that they had more characters than he thought, and that they also had a reasonably good in-built "translation" system so he could probably have got away with not bothering) but the real mistake he made was to store the sanitised page title as the link which he would next feed into his Wikipedia crawler instead of the actual URL. In many cases, this worked, but in more cases than you might imagine, even on English Wikipedia, this failed and his crawler ended up on disambiguation pages or search pages instead of back at the original page.
It was such a simple mistake - and I spotted it pretty much immediately I started looking at his code - but he utterly refused to acknowledge it and insisted until the "artwork" was dismantled that there must be a bug with Wikipedia, or a subtle bug in link handling by Processing and of course, as merely the lackey tasked with plugging the thing in and making sure it didn't burn the place down (and never having written anything in Processing before, though I was beginning to tinker with Arduino), my opinion was near worthless.
There were other problems with the artwork which I was allowed to look at. The boards were run from a single multi-drop RS485 line but instead of wiring it as a "bus" with "drops", which is what RS485 expects, it had been wired (effectively) "star", which was fine with a dozen or so boards attached, but once scaled up to the full number just stopped working. Rather than spending a week re-making that cable I think (if I remember correctly - it's well over 10 years ago now) that I solved that particular problem by using an RS485 splitter and being very careful about terminations.
M.
"Doesnt like certain characters" was my theory when we used a technoolgically wonderful password generator to give super secure temporary passwords (and finger strain) to our users, then they couldn't log in. I mean I don't know that they were pressing £ in a password and we were receiveing $ or anything like that - but we didn't have the problem if the password was "justs omele tters".
And yet many systems reject a password that doesn't contain £ or some other exotic symbol. Though I haven't heard of a system that considers some characters as "illegal" in the password field.
Perhaps you're meant to have the password app on your phone.
But if you can edit the random password value after generating it, then you can remove or substitute characters which you don't like. But that does make the password weaker.
And I keep mixing up g and q. Probably since elementary school.
For a while, I used the company's password generator, I may have been told to, which used punctuation marks freely. But then I only took letters from the generated password, and I made them lower case.
As a side effect, I needed more letters, to get the same security, than the typical generated password - because I wasn't using "! " $ % ^ &" and I was willing to let that be known.
So, I could generate two passwords, and use all the letters from both of them.
...I just reminded myself - setting up a new device - that some services don't allow a password with a space in it. But as very often happens, they don't tell you what can and can't be in the password, or they don't tell you properly. You find out. And so I've written out "thisi smyne wpasw" and it won't let me have it. At least, though, it doesn't let me set it - and then refuse to let me use it.
This post has been deleted by its author
> I'm sure it had spaces...
No, definitely hyphens, see e.g. https://flic.kr/p/VjZdQz
Such a common mistake - variously known as "the Scunthorpe Problem" or "the Clbuttic Mistake" - there's an entire Wikipedia article about it.
Thanks for the Wiki link.
> known as "the Scunthorpe Problem"
In days of bare-AScii chatrooms/forums, a Mr Richard Blick sold paints and decals, and a Mr Richard Smith sold transistors and jacks, and we talked online about their products and services. But they trademarked their nicknames, Dick, and forum censorships mangled their mother-given names beyond reasonable limits. (Blick sold-out but the company is going good after 113 years. Smith bailed on a declining market and sailed into a philanthropic sunrise.)
Before the incidents recited on that Wikipedia page, Compuserve's "Breast Cancer Survivors" group was shut-down, IIRC repeatedly until someone got a clue.
Before the incidents recited on that Wikipedia page, Compuserve's "Breast Cancer Survivors" group was shut-down, IIRC repeatedly until someone got a clue.
As a general rule, they still haven't got a clue. Despite claims of AI magic, most web filtering packages sold to schools/libraries/etc are generally the same keyword based dross they ever were - they've just got a better of exceptions and edge cases so they don't nobble Scunthorpe, Essex and Plymouth Hoe.
They're perfectly happy to automatically block domestic abuse support sites, women's aid, sexual health (including nhs.uk domains), etc, etc until challenged.
Other languages (eg Hungarian) would be able to pin down the right sort of o, oh, ho, eau, hoe, without all that messy insinuation and a massive dose of diacritics.
How on earth do you get across ... I'll try to spell it out for non native English speakers:
Ho! is an exhortation (basically an enthusiastic shout).
A hoe is a gardening implement for weeding and breaking up soil. It has a longer o sound than ho! A "hoe" is also a (en_US at least, not normally en_GB) term for a prostitute, probably riffing on "whore".
That ho! vs hoe confusion is delightful. Shake the spear wrote in Middle English and almost certainly used different vowel sounds than we use today. However, I suspect he would love that abuse of the lingo because that is what he did all the time. English is oh so malleable as required.
@JamesTGrant
Did you read the article? It is about the NORTH Yorkshire Council.
As someone who lives in the former W.R.C.C. I am unaware of a West Yorkshire Council. That was abolished decades ago and now I have to put up with being governed by the Bradford Metropolitan District Council..Jumped up fucking pricks who spend taxpayers money into oblivion. And don't get me started on that useless pile of shit that call themselves Keighley council.
"Consider yourself lucky. At least your metropolitan council is named after a place within it. I live in Kirklees, the council area, not the actual Kirklees which is in Calderdale."
It actually hard to see the logic with the naming of either of those. All of the West and South Yorkshire unitary authorities have one place in them which is by a considerable margin the largest, and with most of them the authority is named after that place - Sheffield, Rotherham, Doncaster, Barnsley, Wakefield, Leeds, Bradford. Logically, the other two should have been Huddersfield and Halifax rather than Kirkless and Calderdale.
"Logically, the other two should have been Huddersfield and Halifax rather than Kirkless and Calderdale."
I don't know about Calderdale vs Halifax but I doubt that Huddersfield actually outweighs that of the collection of other towns such as Dewsbury & Mirfield to any significant extent. What it does show is that the whole thing was an ill-conceived mash-up of two quite distinct areas - more than two when the rural areas are taken into account.
Nor is naming an area after Barnsley a good example. On that basis I can just about see Barnsley from where I live. It would have at least made more sense to have taken the ancient Agbrigg/Staincross boundary as the Barnsley boundary.
"I have to put up with being governed by the Bradford Metropolitan District Council..Jumped up fucking pricks who spend taxpayers money into oblivion."
Doesn't that apply to pretty much all local authorities?
I do remember that on crossing the border into BMDC territory in 2020, all the pavements had one-way arrows tied to their lamp posts - wonder how much that cost? In N Yorks we somehow coped without those.
I wasn't aware of the fuckwittery with apostrophes in road names though - their policy simply suggests that they are illiterate. Things have been going downhill in North Yorkshire since all the district councils were merged into one unitary authority - North Yorkshire is large (the largest English county) and has several distinct regions which have little in common with each other. From one of the places on the far west to one on the coast is a bit over a hundred miles; in much of England that would take you through several counties.
From Section 1, Annex D:
Standard data types used in BS 7666CharacterString: a sequence of alphanumeric characters
Integer: a whole number
Date: a date according to BS ISO 8601, in the form YYYYMMDD or YYYY-MM-DD
... and that's it, that's the only types of data it allows. Which seems kinda limiting, from our perspective, but it could be considered enough to ban apostrophes. As well as hyphens, colons, brackets etc.
Of course, a really strict reading could also be interpreted as prohibiting spaces. So it should be "StMarysWalk". I wonder how that would go down with NorthYorkshire council?
CharacterString: a sequence of alphanumeric characters
I quickly scanned BS 7666 and could not see a definition of 'alphanumeric'. But we live in modern times so I would hope that alphabetic would include everything that Unicode says is a letter.
I am sorely tempted to persuade my local council to name a new street after some foreign person, eg François Mitterrand to see how they cope with ç. My next would be, perhaps, some Indian or Chinese person.
You don't even need to go that far. Irish names (O'Neill (First Minister of NI Assembly), O'Leary, O'Briain to name a small few).They took our land, they took our lives, they'll never take our puncuation!
I work with a standard developed in the 1960s; it too doesn't allow punctuation, but it is nearing 60 years old. On the rare occasion I stray into modernity, string values have punctuation, escapted if necessary.
Ah, yes. The Indian S/W house that replaced their programmers every few months as their visas expired and we, on the receiving end, would be sent XML broken by an apostrophised Irish name because the new guy had fixed the bit of code he didn't understand and must therefore have been wrong.
There's one in Basingstoke.
I suppose the one at the southern end of the country is Alençon, fond de la patrie.
One of my eyelids is starting to twitch like the inspector off the Pink Panther movies at the mention of Unicode. The recent pandemic (little thing, hardly spoken about) demonstrated quite efficiently how unprepared certain public sector systems were to receiving data that wasn't A-Za-z when it all really kicked off.
I live and work in France. At work there is a clocking in machine. It cannot cope with displaying accented characters. "Céline" comes out as "C?line", "Amélie" as "Am?lie".
So this company, whoever it is, is making and selling stuff in France, a country with accents, that doesn't even attempt to show something logical (if it's a limitation of the LCD, then just show the unaccented character, that would be better than a question mark).
Somebody ought to be shot in the arse with a Blunderbuss for that nonsense.
Oh, and our timekeeping system uses decimal hours because, I guess, whatever script kiddie threw the system together couldn't handle base 60 maths. I kind of wonder exactly how badly we get screwed given that 60 doesn't evenly divide by 100. I mean, it might only be a minute, but if we're working around 220 days a year, well, that's nearly four hours that might be missing due to reality being bent to fit the machine rather than the machine really dealing with reality as it is.
Just like here, getting rid of the apostrophes because the machine can't cope. Warm up the Blunderbuss, there's a suffering to be dealt out.
"Oh, and our timekeeping system uses decimal hours because, I guess, whatever script kiddie threw the system together couldn't handle base 60 maths. I kind of wonder exactly how badly we get screwed given that 60 doesn't evenly divide by 100."
Sounds like the coder must have seen Superman III, and every part-cent for every rounded-down minute (of course the machine is going to round it down not in your favour) that you all aren't being paid for is finding its way to his bank account…
(Either that, or they were trying to reintroduce the French Republican Calendar?)
>> CharacterString: a sequence of alphanumeric characters
> to ban apostrophes. As well as hyphens, colons, brackets etc.
SPACES!
That definition strictly read bans SPACES!
(Ok, not the first person to point this out, but - SPACES! And it isn't even as if the N. Yorks folk speak *that* fast)
I haven't done any research on this, but my gut feeling is that spaces are more likely to cause parsing errors than apostrophes (or hyphens, commas or full stops) when fed to buggy parsing code.
If the intent of the ban is to molly-coddle some code that ought to be expunged from this Earth then a whitespace purge is surely the most urgent step.
I still have a diatribe armed and ready for anyone who sends me files with spaces in the filenames. The children were educated early on. Not that it often causes problems these days, but it does still cause occasional issues, and it's a right pain if I'm doing something at the command-line and can't rely on autocompletion to deal with the escaping for me...
M.
Curse this sans-serif font - I read that as "De ice", i.e. the process of removing frozen water. And wondered what on Earth the back story was for that one!
Not too far from me is the delightfully-named "Odge Gribble Lane". Named after a noteworthy local resident, but it never fails to bring a smile to my face.
I live on a street with an umlaut in the name. Admittedly, it's in Germany, so I suppose there's some sort of excuse for it... but I am a Yorkshireman, and I can assure that my late father would have been writing rude (though very polite) letters to the council were he still a a position to do so.
There's a Fanny Street in Cardiff too. If you look around that area you will see Minny Street, Letty Street, Flora Street, Gladys Street, Catherine Street, May Street and such. Obviously named after someone's daughters or mum and aunts or whatever. I suppose during the 19th century when these places were undergoing very rapid expansion, you had to think of something! Not terribly far away, the Adamsdown area of the city, quite apart from very obviously being named after some bloke or other (but what is the origin of the nearby "Splott"?) has System Street leading on to Sun Street, Planet Street, Constellation, Comet, Meteor, Eclipse and Orbit Streets; Metal Street leading on to Gold, Silver, Copper, Lead, Iron, Tin and Zinc Streets alongside Sapphire, Emerald, Ruby, Topaz, Diamond, Pearl Street and probably a bunch more.
Street names is one of the things which make UK (I suppose, European in general) towns and cities interesting - certainly more so than grid-planned cities with "First Avenue" and such like!
M.
Just wondering whether Slutshole Lane in Norfolk has been "adjusted" to modern standards?
<time passes ...>
(a dwarf appears fingering his etymological dictionary and wearing a quizzical expression>
Luckily, according to the historic maps, it started as Sluts Lane then was changed to Slutshole Lane at some point in the 18/19th century. So presumably the incumbents originally lived and/or worked there but didn't have title on the land. So we're safe, Sluts and Slutshole are perfectly good names without apostrophes!
Even more fun in bilingual regions / countries. Google Maps seems to select the language to show at random independently for each section of the same road.
In Spain a moderate number of roads are named after the dates of significant events. Or, at least, events which the councillors at the time deemed significant. I have no idea which of the many events of Spanish history which occurred on the 2nd of April is behind the naming of the street C/ 2 de abril in my city.
Backside Lane, Doncaster, South Yorkshire, DN4
Dick Place, Edinburgh, Scotland, EH9
Assloss Road, Kilmarnock, Scotland, KA3
Semicock Road, Ballymoney, Northern Ireland, BT53
Trailcock Road, Carrickfergus, Northern Ireland, BT38
Slack Bottom, Hebden Bridge, West Yorkshire, HX7
Lickers Lane, Prescot, Lancashire, L35
Spanker Lane, Belper, Derbyshire, DE56
Grope Lane, Shrewsbury, Shropshire, SY1
Hardon Road, Wolverhampton, West Midlands, WV4
Minge Lane, Upton-upon-Severn, Worcestershire WR8
Fanny Street, Cardiff, Wales CF24
Court Cocking, St. Ives, Cornwall, TR26
Slaparse Lane, Exeter, Devon, EX52
Cock-A-Dobby, Sandhurst, Berkshire, GU47
Crotch Crescent, Oxford, Oxfordshire, OX3
Titty Ho, Raunds, Northamptonshire, NN9
Hooker Road, Norwich, Norfolk, NR7
Butthole Lane, Shepshed, Leicestershire, LE12
Willey Lane, Newthorpe, Nottinghamshire, NG16
"So you looked for them, then"
The last, of course, brings to mind the classic line of cricket commentary "The bowler's Holding, the batsman's Willey".
The villages of Overton, Middlestown and Netherton between Grange Moor in West Yorkshire were originally Upper, Middle and Lower Shitlington.
The difficulty with the council's argument, which seems to revolve around correctly identifying a street in an emergency, is that the equivalence between "St" and "Saint" is a lot harder to allow for in a text search than simply disregarding punctuation marks.
In any case, the definitive source of wisdom regarding apostrophes is well known to lie not with BS7666 but with greengrocer's.
> which seems to revolve around correctly identifying a street in an emergency
That would suggest the character set limit is as defined by the established phonetic alphabet and morse code….
Personally don’t see the problem (no apostrophes), given what we are talking about is signpost English.
I've just been looking at some places I know and all I can see is glorious consistent inconsistency! In Manchester St Werburgh's Road is spelled on maps and the tram stop with the apostrophe. But the street signs don't use it. The next road, St Brannocks – also a Saxon "saint" – is blissfully ignorant of apostrophes in all forms!
I can understand not using apostrophes in street signs that are all in capitals: it's easy to overlook them and there are similar rules for diacritics in other languages: French road signs don't have accents, for example; German road signs retain the umlauts but infuriatingly use different sized fonts, rather than abbreviations with apostrophes, for long names. Given the time you have to read road signs, consistency of forms is key for quick reading, which is why many Germans try and learn the route they're planning before the set off.
Then we can look at the history of the apostrophe: it wasn't used with the genitive to denote possession rather than the plural until fairly recently. The other Germanic languages that use the s-suffix for the genitive struggle with the apostrophe, though they have few words that use "s" for the plural so that there are few chances of confusion, say, between "Stefans" and "Stefans"… Though this doesn't make them immune from some other fairly odd decisions: Swedish uses the colon instead of the apostrophe in abbreviations so that "TV's" (or more simply "TVs") is written as "TV:s".
The other problem with apostrophes is that, even when present, they're not necessarily correct.
I think that's a good argument for omitting them (on street signs).
We should remember that punctuation and diacritics are supposed to be an aid to reading, specifically prose. I don't know how it's done in Braille or signing but there's no difference in spoken language with or without the apostrophe. Then I remember that the split infinitive brigade (SIB) will never understand the difference between and precedence of descriptive and prescriptive language rules.
I remember an MP tweeting that she had just been to the local "Farmer's Market".
She was berated by pedants pointing out the errant apostrophe... she pointed out that this was a car park off the Finchley Rd (north London) and it was possible a farmer had been present, but couldn't be sure
"French road signs don't have accents, for example"
Tell me you haven't been to France without telling me you haven't been to France...
One of the two main programmes I work on uses "St Marys..." and similar - except a few years ago some bright spark renamed all the "St" road names in the DB as "St." and the address search won't return any results unless you include the full stop. But whoever made the change forgot that there are a couple of blocks of flats starting with "St ****" and those remain dot-free.
Confused? You will be.
Thankfully the other programme uses boolean searching and doesn't care whether you type the full stop or not.
"that the equivalence between "St" and "Saint" is a lot harder to allow for in a text search"
Given that these sorts of abbreviations have been in us for hundreds of years, shouldn't any half competent setup look for "Saint" and "St."?
It's like, we would refer to "Dr. Zhivago" as Doctor Zhivago, not Duhr Zhivago (or however you'd say "Dr").
These are known knowns.
What is it with computer programs that can't figure out that they're working with data, not with code ?
I've known the Bobby Tables joke for ages now and I still can't understand how that happens. If I type DROP TABLE USERLIST here, El Reg is not suddenly going to find that table gone. This is data, not code.
It would have to be '); DROP TABLE USERLIST
But even that wouldn't work, of course, because El Reg isn't that stupid. (And even if it were, and even if a table by that name does exist, I'm pretty sure it would have dependencies that would cause the command to fail anyway.) But is anyone, any more? I mean, I know that sort of thing used to happen back in the earlies, but surely we've all outgrown that stage by now? Haven't we?
'Twas elsewhere on this site - the story about Nokia, in fact - that I read earlier today a comment about how early Android versions would interpret typed text as system commands. The example given was that typing "reboot" in a note or contact name would, indeed, reboot the phone.
So confusion between data and code is definitely not a solved problem. One has to imagine that the sort of programmer that would make that mistake would also stick their socks up their nose and wear their shoes on their hands whilst licking the walls and dribbling, but...
I worked for BT when they owned Cellnet, and Cellnet had the idea of providing a web-to-SMS service, giving registered customers the ability to send a small number of messages per day. (Obviously, limited to prevent spamming.)
I didn't work on that, but the group next door did. It was due to go live on Monday, and on the Friday afternoon they sent an email to colleagues "This is about to be released. Do you want to have a go and see how it works?" Well, I knew they were using Perl for CGI (that dates it) so I tried a few messages with special characters to see what arrived on my phone. I was able to work out how to get the code to execute a shell command which emailed me the password file. (Not very interesting. It was a test server with no actual users.)
As I passed their office on the way to the pub, the place was a hive of activity. Apparently, they were at it until late that night.
In the early days of GSM I used to dail-in to some kind of text-based system (modem and terminal software) to send SMSes to my boss when he wouldn't have wanted to have to deal with a phonecall. Pretty certain that was Cellnet, I don't think Vodafone offered the service and Orange and One2One hadn't arrived in our neck of the woods yet. Very simple, no authentication at all from what I remember, just "enter the phone number", "enter your message" and it rejected messages longer than 160 characters. Would probably have been trivial to use the system for nefarious purposes, but we were all better-behaved in those days, weren't we?
M.
SQL injection relies on the fact that the data is- or can be- embedded within the SQL command (i.e. the code) itself and exploits the fact that quote characters are used to delimit string data and keep it separate from the executable code surrounding it. However, if the contents of that "data" string are being supplied dynamically by an untrusted user (e.g. via a web form where someone is expected to have typed in their name) and isn't properly "sanitised", it might itself contain quotes. These would prematurely end the data string and result in the rest of the supplied "data" string being treated as executable SQL commands- something a malicious user can exploit.
This happens if you're creating SQL statements by pasting strings together, (e.g.) taking someone's name from a web form embedding that within an SQL query which looks up someone's data from the database system. So if someone types in "Jim", our web app creates the following command to fetch the data:-
SELECT * FROM users WHERE name = 'Jim'"
However, if a malicious user types in (literally):-
a';DROP TABLE users; SELECT * FROM userinfo WHERE 't' = 't
The resulting string (with embedded "data" string in bold) will be
SELECT * FROM users WHERE name = 'a';DROP TABLE users; SELECT * FROM userinfo WHERE 't' = 't'
However, because the supplied (embedded) data includes extra quotes and other characters (including the semicolon which indicates the end of a statement), when this string is passed to the DB, it will see it as three separate commands:-
And, well, you can guess what happens when it runs that "DROP TABLE users" in the middle...!
One can "sanitise" user-supplied input to escape or omit such characters, but that's still risky- i.e. prone to error or oversight. This is why creating SQL queries this way (i.e. via string concatentation) is widely seen as antiquated and unsafe and (e.g.) parameterised queries are much preferred instead.
More detail and an example here.
This road sign always used to amuse me: Bronte Parsonagë. These days they've lost the diaeresis on Brontë.
AFAIK diaresis is the symbol of two dots above a letter and is thus equivalent to the umlaut. I mean, really, that's just the preference for Greco-Latin names in analysis, which is why we end with shit like ptermigan. Whether the symbol denotes a vowel shift as in the umlaut in German or Swedish, or the opposite as in English, French and Dutch, well, who cares. It's worth noting that in both cases the symbol is redundant as "ä" == "ae" (and is written so when the character isn't to hand) and "poet" manages two syllables without any additioal markings.
Add the fact that almost all address forms in S/W assume that the address will be in a city however incorrect that may be geographically and historically.
The link includes the town of Street, Somerset as a counter example. I once worked with a database where an address was listed as a number at High Street, Somerset. When on holiday in Somerset I checked. The business was where I anticipated: in High Street, Street, Somerset.
According to local legend, the HIgh Street is the only street in Street. All the others are roads, avenues etc. This was supposedly to avoid confusion. Judging by the addresses on some of the post we got, the local postmen were well-used to deciphering dodgy addresses.
address recognition
Officially the town where I live is Caerphilly in the county of Caerphilly so "Caerphilly, Caerphilly" would be correct (or probably more correctly, "Caerphilly, Caerphilly County Borough Council"). The Welsh spelling, "Caerffili", is also legitimate*. I'm sure I've mentioned this here before, but back in the early days of address databases they were woefully incomplete and just for the heck of it I manually filled an address as "Caerphilly, Caerffili". To this day at least one address database retains that (well, I'm claiming it as mine anyway).
M.
*In fact most villages, towns and cities in Wales have two legitimate names. Some are more-or-less translations, some are simply transliterations (blame Victorian cartographers / census takers), some have the same origin but different takes on it while others have completely different origins;
I find this sort of thing fascinating (I dare say it's the same in Scotland and Ireland)...
.
A comment above about France reminded me: I think that the Welsh language uses accents, and even has two "unique" ones (w-circumflex and y-circumflex). Do any Welsh placenames use these? Given the protected status under law of the Welsh language (in Wales), there's scope for all sorts of fun.
Without pausing to think about it too long, Ynys Môn springs immediately to mind. I don't think there are that many but I'll probably think of half a dozen others as soon as the edit window is closed for this post :-)
(edit): there are several places with "dwr" in their name (meaning water / lake) but only a few seem to spell it with the correct accent; dŵr. I'm guessing there may also be several with "tŷ". This word, meaning "house" or "dwelling" is often one of the first accented words young children are taught as the "roof" on the y is easy to remember (cos houses have roofs). St David's in West Wales is known as Tŷ Ddewi (David's House) in Welsh, though Openstreetmap doesn't seem to have this alternate. Oh, and OSM also spells it "St Davids" without the apostrophe!
M.
You beat me to it.
From The UK Gov Index of Place Names in Great Britain, download the index, unzip it, grep out all the usual characters, and you are just left with ! and ô
Westward Ho! and Ynys Môn
River river is quite common, I believe it's true for Thames/Tame/Tamar and some others. Makes sense too because locally we tend to refer to the various bodies of water just by their names: the Thames, the Mersey; the Tyne, etc. In days gone by, it would possibly have been the only river people would need to talk about.
Only 5 or 6km from Torpenhow is Aspatria, which has its own problem. The locals pronounce it "Spyat-ree", posh folk "As-spay-tria", but the recoded message on Northern Rail trains says "As-spat-ria".
"Let's not mention Torpenhow Hill"
Good idea. It doesn't actually exist.
https://www.thefreelibrary.com/The+debunking+of+Torpenhow+Hill.-a098250320
On the other hand, Pendle Hill does ... "Pen" means hill. Pen Hill became Pen 'il, became Pendle ... Which became Pendle Hill, because, well, because it's British. So basically, we now have the rather imaginatively named Hill Hill Hill.
Chester is straight from the Latin for castle – it was a a Roman port – to English just like nearby Personchester…
Not so sure about Wales but Scotland now has quite a lot of road signs in Gaelic for places with Saxon (East coast) or Norse (North East) origin which I find not only confusing but patronising.
Not a punctuation thing but I have to phone a building society later because their website insists that I enter the first, third and twelfth character of my password.
The acceptability criteria for passwords on their site includes that it must be between eight and twelve characters long. Mine is currently eight.
It's ok, he doesn't really mean "password" in the way we mean it (but I'm willing to bet that the website calls it that!).
If you really haven't come across this system yet, is a hang over from moving to the web from 'phone banking, where you would (still do) have to tell the person a subset of the "secret phrase" so that the phone operator couldn't learn it all and type it in when they aren't in a recorded call with you.
Think of it as 8 to 12 totally independent, but very short, passwords, (hopefully!) stored accordingly.
Yes, but the secret phrase/PIN/password, whatever you want to call it, is stored somewhere in plain text, even if it isn't displayed in full to the customer service agent on the phone. Are you seriously arguing that's good practice?
edit: Yes, I've come across this system often.
Bit late, but as I have been impugned:
> stored somewhere in plain text ... Are you seriously arguing that's good practice?
Well, no. Ahem,
>> passwords, (hopefully!) stored accordingly
And what is appropriate storage for a password? Hmm? Give up? Well, Doctor Syntax has been good enough to supply a few answers in my absence from here, for which I thank him.
And the letters you supply aren't decrypted and shown to the phone operator (again, unless the system is as appallingly designed as you believe it has to be): the operator just types your answers into their terminal and presses "check". The system then does the password-style processing[1] and comes back yay or nay. If you get it right then, yes, *you* have just told the operator that subset of your code, but if you missed one then the operator (again, hopefully) can't even tell which one (or two or ...) was incorrect.
[1] hopefully (that word again) you know that passwords aren't checked by comparing the plaintext, whether by decrypting the contents of the password database or by, shudder, storing the plaintext in the database.
It should accept a null value for the twelfth character, however, to force the input checker to recognise you have entered a digit in each field, you might have to enter a character and then delete it.
I have a slightly different problem, just accidentally discovered the limit to one financial site, it accepts the setting of a 14 character password, but only actually stores and uses the first 12 characters…
This seems to be a fairly common problem (code copy/paste?), I've met it some years ago on a merchant website, which would happily let me set a 10-character password, but subsequently only accept the first 8 characters of it (choked if I gave it all 10). Message to support went unanswered.
A variation on this is requiring the use of special characters, but actually silently not accepting some of those, so an "a§b+c#d" password is stored as "ab+cd". Endless hours of fun!
You're a serious masochist if you use § as a character in your password! ;-)
I'm not sure if anyone can type that unless they have a non-US Mac keyboard, or is a dab hand with the <compose> key (which might not be available on an actual OS login screen, if you need to type it there)…
> You're a serious masochist if you use § as a character in your password! ;-)
That's precisely what makes it interesting: Very rarely used, so less likely to be guessed. As for the keyboard, who cares: I use a password manager and copy/paste them...
There is no brute forcing a password like "ꝷ‽ƝzA8ͲꞂ¶ʨﬔᵺs73▒œᴔῷ╟ʓ₩ʯkΦdb∩Ӝᴚꝷ1┬┴Qr‽Ɲ‰"... *evil snigger*
My bank has an automated system to use it for authentication when you phone them, before they connect you to an operator. The bank also switched from numeric-only passwords to alphanumeric ones. But, and it's a big one, although you can enter numbers by dialling, for letters you have to use voice recognition, and the voice recognition fails spectacularly for me. The last time I tried it I made six attempts on the phone and then gave up and went across town to visit "my" branch.
I suggest they check out the definition of PAF (postcode address format) which what the UK Post Office uses for address matching. Apostrophe is allowed, if their address matching software can’t handle it, it’s a bug. If they can’t work out how to store in a database it is incompetence.
Yes, PAF does allow apostrophes as in this example:
1 O'Neill Crescent
Barrack Road
MAGHERAFELT
BT45 6QY
However interestingly, for the street quoted in the article the apostrophe is not present (you can check on the RM Postcode finder here: https://www.royalmail.com/find-a-postcode
1 St. Marys Walk
HARROGATE
HG2 0LW
So I suspect the Royal Mail PAF team have a pretty relaxed attitude to this.
Even more interestingly, I actually used to live on St Mary's Walk in Harrogate (and I don't care whether the apostrophe is used or not personally).
Do not start me on websites which let you select your flat (apartment) address from a list in the PAF form "42/3 Station Road" [1], and then move on to the confirmation screen where they then insist "Address cannot contain / character"!
Seriously, probably a good quarter of the population of Scotland [2] would like to kill you, if we ever find out who you are…
[1] It's bad enough that in many cases the flat is more commonly known as (say) "42 (1/1) Station Road": if the Royal Mail was going to devise a standardised (ah, yes) address format, it would have been nice if they had forcibly twisted the arms of councils, land registries, etc, to actually use it!
[2] or other parts of the UK where flats are common…
You can get riled up by one thing, be in a cold fury about another, resignedly sigh over a third, become apoplectic over a fourth, work in grim silence over a fifth.
And channel proper anger at a social injustice into the driver for quietly, physically, doing something to aid, whilst at the same time being left with nothing but vocal disapproval over the grinding stupidity and petty incompetence of the ignoramus whose feeble grasp on power extends to street signs and no further.
Those with a good grasp of the language and the ability to coin a quotable response are not of necessity unable to respond to the conditions around them.
> Double negative.
If you are referring to
>> are not of necessity unable to respond
then consider what would happen if we attempt to simply replace the "two negatives" with positives - "are of necessity able to respond" - and notice that this does not carry the same meaning. Because of the modifier "of necessity" that lies between the two original negatives.
You could attempt to forage for a way to replace that modifier and retain the positives, but by then you will have moved the sentence drastically away from the original form and making it unnecessarily difficult to recognise it as objecting to the original comment's implication.
Syntactically there's no difference in English but serifs are easier to read in small print. Also, typefaces have to cope with different conventions in different languages and include the relevant glyphs. However, I would like to see keyboards provide electric shocks for the abuse of ticks and backticks, designed for accents, for anything else, including shell script! This is unfortunately very common in Germany: Charly`s Bar
What I find most annoying are so called smart quote systems that insist on changing "this is a string with double quotes" into “this is now broken with different left and right quote characters”. Makes providing program code examples impossible for simple cutting and pasting.
One would hope that the council keeps their lamp posts sufficiently maintained that the act of hanging a hanging basket (likely from a pre-existing bracket) does not cause an electrical fault.
The more likely safety issue with the hanging baskets is old chains or fatigued/rusty brackets and baskets and the whole darn thing ending up landing on someone.
More importantly, I would hope it was not possible to get an electric shock off a lamp post; or should we be telling children not to touch metal and grounded lamp posts…
Suspect the safety training, requires the person, wishing to attach stuff to a lamp post, to earth the lamp post before commencing work….
New name: St Mary%27s Way
Aaaarrghhh, that gave me horrible flashbacks. Where I work, our corporate IT is woeful at handling links between email, file explorer, instant messaging, our internal social media etc. Cutting 'n' pasting a link from one to another invariably adds superfluous prefixes like "file" and countless slashes of both forward and backward types. Often, all spaces are converted to "%20" which of course the receiving system won't understand. Our IT rules say to not use spaces in file names. One system always insists on opening links in a web browser, whatever they are links to. Sigh.
And our IT people wonder why so many of us email files to each other instead of sharing links.
the apostrophe has been their stock in trade from the advent of literacy. :)
Arkwright once explained to Granville that the properly misplaced (grocer's) apostrophe was certain to lure the passing pedant onto the premises and into the mercenary clutches of an experienced grocer and from whom the poor victim would only escape with a lighter purse.
This apostrophe would cause non-locals a bit of puzzlement given that it replaces 5 letters not to mention the missing 'U' from 'BURY':
https://www.google.co.uk/maps/@53.6449441,-1.7758415,3a,37.5y,171.44h,79.34t/data=!3m6!1e1!3m4!1s6j2bb7p8he-sGAhCV59-6w!2e0!7i16384!8i8192?hl=en&entry=ttu
Implemented in 1890, the BGN's Geographic Names Information System (GNIS) "discourages" the use of the possessive apostrophe in place names, with only five exceptions granted since enforcement began: Martha's Vineyard, Massachusetts; Ike's Point, New Jersey; John E's Pond, Rhode Island; Carlos Elmer's Joshua View, Arizona; and Clark's Mountain, Oregon.
I wonder how O'Hare airport escaped the GNIS' slimy clutches.
No, surprisingly, the BGN was established in 1890 and has been going since then. See the links in the story.
C.
> use of the word percent instead of %.
Hey, yeah!!! The local newpaper's proofreader retired, and for years the copy was full of escaped entities. Any time they hit a French letter (here next to Quebec logger country, that can happen a lot). I have not noted it in a few years. Maybe patched? Or maybe the paper is just so shrunken that they hit a foreign character only every blue moon?
Middlesbrough was in the North Riding once upon a time. Roadsigns and white paint on roads usually abbreviates this to Middlesbro'. I think I may have seen Midd'bro as well. (Not been up there in a while.)
Will their database be instructing the white letter painters to use Middbro now?
Edit: I think I may have seen M'boro as well.
Whilst that's acute thought, losing the accents in French would be a grave matter.
whatever the "hat" and accent underneath a letter are called
FYI - Circumflex and Cedilla (pronounced "sedilla"), respectively.
Thank you Mrs Flitcroft for all those GCSE French lessons so many years ago :)
"Remove the punctation from the data base but retain it on the printed road signs. A bit of common sense, or is that asking too much from the council mind?"
That would confuse poscode and delivery systems!
Just keep the apostrophes everywhere. It's not hard to design systems which can cope with it - CRM databases have to cope with it as there are plenty of surnames wich contain an apostrophe.
CRM databases have to cope with it as there are plenty of surnames wich contain an apostrophe
Unless they make people change their names. This happened to a lot of emigrants to the US in the Ellis Island days, and there wasn't any sort of a system of rules it just depended on whatever the official checking them in felt like. Supposedly Donald Trump's family was Drumpf in Germany.
That's a bit of an urban legend:
https://www.smithsonianmag.com/smithsonian-institution/ask-smithsonian-did-ellis-island-officials-really-change-names-immigrants-180961544/
"“If anything, Ellis Island officials were known to correct mistakes in passenger lists,” says Philip Sutton, a librarian in the Milstein Division of United States History, Local History and Genealogy, at the New York Public Library, in a blog post delving into the name change mythology.
More commonly, immigrants themselves would change their names, either to sound more American, or to melt into the immigrant community, where they were going to live, says Sutton. If name changes happened with any frequency on Ellis Island, it was not noted in any contemporaneous newspaper accounts or in recollections from inspectors, Sutton says."
Regardless of what Sutton states, I know for a fact that all eight of my great grandparents had their surname changed by the folks on Ellis Island. A few also had their given name changed (or "adjusted"). Family lore says that the officials couldn't read their paperwork, nor understand what they were saying, so they just guessed to keep the line[0] moving. All were emigrating from the north of Finland. All arrived on separate boats, over a period of a couple decades in the mid-late 1800s, and somehow managed to find their way to the north coast of California (fishing and timber).
Fortunately, all of them made certain to pass along who they really were to their sprog, and down through the years. Rather amazingly, the family still has all of their original paperwork, and as of the late 1970s we are now back in touch with most of our kin in "the old country" (some branches went missing over the years; seems that living in close proximity to the Soviets had that effect on many families).
[0] British readership: For "line" in this context, please instead read "queue". Ta.
English really is an lazy language :-) Oops sorry those last three characters are probably banned. So imagine that this software company was delivering their product to France, Germany, The Netherlands, Spain, Portugal, Italy, or any of the countless other countries where accents and other characters are just normal everyday occurrences, are they going to simply tell Johnny foreigner to play by English rules? Coding for characters other than A-Z really isn't much to ask.
Isn't that sign for the bit of land which belongs to old Tractor Tim? The farm with the rainwater runoff risk of reducing ground to mud? Warning you that the side road Tractor's Turning is Tractor's tract where tractors turning in torrents are treacherous
Not to be confused with the Lorry's Loading, where despite the labelled loads laying 'longside the layby by Laurie's Lading, the signwriter just got it wrong.
I often see signs like this that look wrong, but I like to imagine how they could be right. "Tractor's Turning" could be a warning of a single tractor that is turning. Niche, but possible.
A town I used to live in had a convenience shop called "EAT'S AND TREATS". A friend once explained how he knew the owner (a lady called Eileen Allen Turner) and that despite appearances, the grammar in the shop name was correct.
Sometimes there can be too much punctuation.
Thankfully it was eventually fixed.
""This restricts the use of punctuation marks and special characters to avoid potential problems when searching the databases as these characters have specific meanings in computer systems," the spokesperson added. "
What bollocks - as the article alluded to with it's reference to little Bobby Tables. Do they think that a string following a dollar will be converted to a local variable, or one forllowing % becomes a format specifier?
As for the UI level, there's nothing stopping the apostrophe being stripped from both user input and database output when doing a search-match.
They were probably refering to their elections management database, not their GIS database. There are several which just stuff anything in the fields without escaping characters. In Oxford, every ' is replaced by a grave accent ` . So there is a St Mary`s ward in their database. I wrote a Perl script to fix that (and the strange address fields) in CSVs of their electoral register.
I took on a new client recently. They had a custom PHP app written by an old guy who was retiring. Among the jewels he left was a page containing a form that proudly stated:
Do not use apostrophe's [Sic]
I didn't know it was possible to despair for two different reasons simultaneously, but I managed it.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
