I'm curious though, where did you get your Hispanic names from? Seems like they come from a database of funny, weird and probably offensive names. No one is called like that in Spanish-speaking countries, and yes, I know the names are supposed to be fake, but just comparing the Hispanic with the American ones, you can tell there's a big difference.
Whether or not is intentional, it could be even more helpful if you use "normal" names.
I’m from Argentina and also find that the names are very unusual. The usual names vary a lot by country, but I could not even recognize a most of them, and I had to retry about 10 times to get a realistic name.
The England/Wales names are weird too, lots of unusual first names.
Also, Wales seems way over-represented. It has a population that's tiny compared to England, but half the names generated seem to be Welsh names such as Cerys.
The first thing I noticed was that the vehicle field doesn't seem to be sensitive to the products available domestically for the person's address. And for bonus points, it could even consider demographics of the address versus price of the car (I got a Bentley while living in the middle of nowhere in the rust belt, which seems a bit suspicious).
I've got a colleague who's legal first name is "G". He's listed as "Gee" in about 50% of the places he needs to register (including some important places where there might be future legal ramifications of not using his proper legal name - I'm looking at _you_ Crazy Domains web/database programmer!)
Same for French names. Last names are okay, but first names aren't French at all for most of them.
I think maybe you had a list of all possible names, but you didn't take into account that some of them are more frequent than others ?
I tried the generator with gender:random; nameSet:England+Wales; country:UK.
All but one of the names is a forename of someone I know - that name is Scarlet and it appeared twice in my sample of 20. Only one of the names was 'foreign' though. In my area there would be several Asian names in that sample.
Yes, I was about to say the same. I tried a few "Hispanic" names and they looked absolutely weird.
On the other hand the notion of a "reasonable" Hispanic name can be quite broad. Names in Spain can be very different from names in Argentina (where Italian surnames are quite common) or from US (when it's not uncommon to have an English name with a Spanish last name)
Not sure about that, but other name set can have the same problems if they try to cover a language shared by lots of people (like Arabic)
Yeah I think that trying the bunch all Hispanic names under the same category is part of the problem. There is 400 million native speakers and they are spread over multiple continents, countries and cultures.
http://en.wikipedia.org/wiki/File:Map-Hispanophone_World.png
No surprise that names will widely vary from one place to the other.
If you are curious, you should always look at the FAQ first :). http://www.fakenamegenerator.com/faq.php . Look at the section: "What can I use the Fake Name Generator identities for?"
Same in Dutch-Belgium setting. None of the names sounded realistic and when van de/van der or variants were used it was always in lowercase which is a thing from the Netherlands. In Belgium it'll be in capital letters.
My thoughts exactly. After seeing the names it generated, along with the the addresses and so forth, I immediately started looking for the link to buy or download the code, database or whatever was on offer. To my disappointment, there was nothing available.
Colour me surprised too. Just 5 minutes ago I was searching for a false name generator. My heart leapt with fright on seeing my recent history at the top of HN.
Interesting to see this come up. A couple years ago I started and then mothballed a system for generating realistic fake people (which I called golems) using deterministic and reversible algorithms instead of random numbers (it has a bunch of advantages). Got to love heroku's free tier, cause it's actually still on line:
RIG (Random Identity Generator) is a free replacement for a shareware
program out there called 'fake'. It generates random, yet real-looking,
personal data. It is useful if you need to feed a name to a Web site,
BBS, or real person, and are too lazy to think of one yourself. Also,
if the Web site/BBS/person you are giving the information to tries to
cross-check the city, state, zip, or area code, it will check out.
$ rig
Adolph Cline
739 Anton Dr
Mentor, OH 44060
(216) xxx-xxxx
$
MacPorts is full of useless junk. MacPorts will install a new sandboxed python, ruby, perl, etc when just want to get wget (the built in versions are fine!). The libraries are all broken out where brew keeps them simple and a lot of the packages are old silly packages that no one on the Mac really cares about (every single GNOME library for example).
brew isn't just moving forward quickly, it has really just out right replaced MacPorts and anyone still using MacPorts is living in 2007.
MacPorts does not use system libraries for good reasons:
"There are several reasons why MacPorts uses its own libraries. It makes ports more consistent across different versions of Mac OS X. For example, if we can rely on openssl 1.0.0 from MacPorts, we don't have to test every port that needs ssl for every available openssl installation. Apple's software tends to break from time to time (e.g. openssl refuses to build with an old zlib, but for awhile Apple shipped the old headers of the vulnerable zlib version). Even if Apple's versions aren't broken, they're rarely up-to-date. Apple has a habit of not updating the libraries in Mac OS X until absolutely necessitated by a security vulnerability." [1]
I used all three at one point in time. What I love about homebrew is that the tedious work of maintaining a sane package repository is distributed over many shoulders (3429 at the moment) - and those who contribute most of the time do really care about the packages they are watching over.
I haven't found an easy way to add/update packages to finc or macports, but I could contribute to homebrew within 5 minutes.
I wish we would see more work distribution patterns like this in fields outside code/programming.
Once you start piling up some dependencies (vs. self-constained apps), I found homebrew more of a mess, while MacPorts more or less worked as I'd expect. The thing that drove me off Homebrew was dependency-hell with different versions of Python packages, while MacPorts pulled in the right versions of all the dependencies (it can even handle needing multiple versions of NLTK/SciPy/etc. installed in parallel, if different ports depend on them). This is perhaps the flipside of the reason many people prefer Homebrew: it tends to build against the base OSX stuff, while MacPorts tends to pull in its own parallel world of dependencies. IME the latter works more reliably, though it takes up more diskspace.
But, Debian's package management works better than either of them, so I've sort of been moving towards doing any kind of unixy work in a VM and treating OSX as just a desktop.
It's possible to end up in a different circle of dependency hell with macports, where it insists on installing 10 different versions of everything at great length, when you already had a perfectly good one. If I'm not a python or perl developer, I don't want to compile many different point releases of those language ecosystems just to run some little utility scripts. And then of course everything is connected to everything else, so you can end up trying to solve graph theory problems when you're meant to be working.
Battle-hardened *NIX admins rightfully laugh at this attitude, but for a lot of web developers who use their laptops as a "sharp tool" and do all the heavy lifting in linux vservers, it makes sense to have a slightly laxer approach to package management.
(and then there's the increasingly common cases of build scripts just being broken on macports, because it receives less and less community attention now. this can combine with the above dependency graph problems to produce situations where it's easier to just nuke /opt/local and start again.)
rig obviously supports only a handful of fields, if you need something in between rig and Fake Name Generator, picka may be an option https://github.com/antlong/picka
And if you're tired of manually dealing with AUR packages (or just lazy) you might want to look into yaourt [1]. It uses the same syntax as pacman but also searches and installs from AUR.
Same with Icelandic names. Also, the GPS coordinates were often out at sea or on a mountain. Using openStreetMap it should be easy to generate more plausible coordinates.
Try #1: "Dante Marcelo". Dante is a very strange name for an italian, I guess it is used in the US. Marcelo lacks an "l" (it is Marcello) so it sounds Spanish instead.
Try #2: "Berto Trentino". Trentino is realistic but Berto sounds a lot like an abbreviation of "Alberto", so not a real name even if I guess there are people actually named "Berto".
Try #3: "Pupetta Rizzo". Can't imagine somebody called "Pupetta", it is something you say to small children as "Little Doll" or alike.
Hey! I actually know a "Dante Marcelo"! (but both as given names, he's known as Marcelo). But he's from Argentina (where Marcelo is a relatively common name), not from Italy
I think its fair to say that scenarios where a service cancels a user account because his/her name was not "realistic sounding" are few and far between. (Assuming you're not entering names like "Doorknob Toothbrush")
Well, the site did generate Iva Lavallée as a French female name for me. I had never heard of someone called Iva, but it seems it's been given to exactly 100 persons in France since 1900.
Iva Lavallée sounds like He's gonna swallow it, really.
My guess is that it's just randomly selecting from a list of names with equal probability, rather than weighting them based on how common they are. So all of these long-tail names are coming up way more often than you'd find in a phone book.
Yeah, similar is when you get names from Canada and look at the cities.
The first one I get is from Weagamow, ON, which is a North Caribou Lake Indian reservation, with like 900 people in it.
The second I get is from Cornwall, ON, a small city of 45k people, the next is from Orangeville, ON another town of 30k people. Then Wawa, ON, a town of 3k people.
Meanwhile, Toronto, Vancouver, Montreal are all unrepresented.
I just got a person from Byemoor Alberta, a hamlet of 35 people. It says she lives on 1304, 90th Ave, but there's a 1 Ave, a 2 Ave, a 1 St, and a 1 St North. Definitely no 90th Ave.
Edit: Doing a HN search this was posted about a year ago. That is why I probably didn't post it. Should we start submitting reposts of old stuff? I probably missed "cool tools" from years ago.
Don't know where you live, but the Dutch government publishes spreadsheets with testdata. It contains names, middle names, addresses, domestic and foreign names, etc. No need to generate your own. I haven't had to use it yet and I didn't have a very good look, but I'd think it contains all sorts of exceptional cases which your system needs to handle. If it successfully processes that spreadsheet, it probably handles anything.
Hmm, in my experience, 85 yo Dutch women are unlikely to drive a Mazda Miata. Although maybe mine examiners are more thrill-seeking than the average 85 yo. Fun!
The occupations do seem the most likely to cause suspicion. Some of them are trades where you have to be licensed or registered to have that occupation, so a quick search would prove the person does not exist, or at least is lying about their job.
Then there's the automobile, which could also raise red flags. A Fiat Tempra driven in the USA would probably be suspicious to anyone who knows that Fiats probably weren't sold in the US in 1994, for example. And it might be weird for a clerk to be driving a 2010 Infinity...
Back in the BBS days there used to be programs written to auto-generate identities in bulk, for reasons i'm not aware of. They were designed to minimize scrutiny because there might be humans actually looking at the data you used, since there was less automation in terms of processing accounts back then.
I'm not sure why National Auto Parts would needs a Perianesthesia nurse who runs a website called SuicideLaws.com, but they seem to be paying him well enough to drive 1992 Ferrari 512 TR.
There's a problem with Belgium: it generates street names in flemish for cities in the french speaking part of the country. Makes it obvious the address is fake.
162 comments
[ 4.1 ms ] story [ 221 ms ] threadIf it could comb death certificates for names and demographic data rather than randomly generating it, that would be even better.
You'd probably make a fortune selling SaaS contracts to criminals in Russia and Washington.
[1]: http://www.fakenamegenerator.com/blog/2013/10/response-to-ac...
I'm curious though, where did you get your Hispanic names from? Seems like they come from a database of funny, weird and probably offensive names. No one is called like that in Spanish-speaking countries, and yes, I know the names are supposed to be fake, but just comparing the Hispanic with the American ones, you can tell there's a big difference.
Whether or not is intentional, it could be even more helpful if you use "normal" names.
Disclaimer: I'm Hispanic.
Also, Wales seems way over-represented. It has a population that's tiny compared to England, but half the names generated seem to be Welsh names such as Cerys.
You can be the first to own one!
Only a 1995 model though :-)
1996 Buick Roadmaster
which is about as likely as him riding around on a Unicorn.
Random name factoid: In England, J is a valid first name and A is a valid surname. So someone's full name could be J A.
I wonder how many databases would accept that.
All but one of the names is a forename of someone I know - that name is Scarlet and it appeared twice in my sample of 20. Only one of the names was 'foreign' though. In my area there would be several Asian names in that sample.
Not sure about that, but other name set can have the same problems if they try to cover a language shared by lots of people (like Arabic)
No surprise that names will widely vary from one place to the other.
Most of names sound Polish, but I never ever heard of anyone called that name.
Also lastnames seem to be missing Polish characters sometimes
Here are other resources that we found that were helpful:
- http://www.generatedata.com/
- http://databasetestdata.com
- http://randomuser.me (useful for frontends)
- http://gedis-studio.com/ (not free)
Also, TIL that Hipchat doesn't always pull all the links
Colour me surprised too. Just 5 minutes ago I was searching for a false name generator. My heart leapt with fright on seeing my recent history at the top of HN.
http://golems.herokuapp.com/person/random.json
If anyone's interested in learning more let me know.
But yes, arguably, `brew` is moving forward very quickly.
brew isn't just moving forward quickly, it has really just out right replaced MacPorts and anyone still using MacPorts is living in 2007.
"There are several reasons why MacPorts uses its own libraries. It makes ports more consistent across different versions of Mac OS X. For example, if we can rely on openssl 1.0.0 from MacPorts, we don't have to test every port that needs ssl for every available openssl installation. Apple's software tends to break from time to time (e.g. openssl refuses to build with an old zlib, but for awhile Apple shipped the old headers of the vulnerable zlib version). Even if Apple's versions aren't broken, they're rarely up-to-date. Apple has a habit of not updating the libraries in Mac OS X until absolutely necessitated by a security vulnerability." [1]
[1] https://trac.macports.org/wiki/FAQ#ownlibs
I haven't found an easy way to add/update packages to finc or macports, but I could contribute to homebrew within 5 minutes.
I wish we would see more work distribution patterns like this in fields outside code/programming.
But, Debian's package management works better than either of them, so I've sort of been moving towards doing any kind of unixy work in a VM and treating OSX as just a desktop.
Battle-hardened *NIX admins rightfully laugh at this attitude, but for a lot of web developers who use their laptops as a "sharp tool" and do all the heavy lifting in linux vservers, it makes sense to have a slightly laxer approach to package management.
(and then there's the increasingly common cases of build scripts just being broken on macports, because it receives less and less community attention now. this can combine with the above dependency graph problems to produce situations where it's easier to just nuke /opt/local and start again.)
If you absolutely can't wait, you can run this to install it:
Voilà.https://aur.archlinux.org/packages/rig/
[1] https://wiki.archlinux.org/index.php/yaourt
1. https://github.com/stympy/faker
Elves, on the other hand...
It's more of a fun read than anything useful, really
Try #1: "Dante Marcelo". Dante is a very strange name for an italian, I guess it is used in the US. Marcelo lacks an "l" (it is Marcello) so it sounds Spanish instead.
Try #2: "Berto Trentino". Trentino is realistic but Berto sounds a lot like an abbreviation of "Alberto", so not a real name even if I guess there are people actually named "Berto".
Try #3: "Pupetta Rizzo". Can't imagine somebody called "Pupetta", it is something you say to small children as "Little Doll" or alike.
Iva Lavallée sounds like He's gonna swallow it, really.
The first one I get is from Weagamow, ON, which is a North Caribou Lake Indian reservation, with like 900 people in it.
The second I get is from Cornwall, ON, a small city of 45k people, the next is from Orangeville, ON another town of 30k people. Then Wawa, ON, a town of 3k people.
Meanwhile, Toronto, Vancouver, Montreal are all unrepresented.
I just got a person from Byemoor Alberta, a hamlet of 35 people. It says she lives on 1304, 90th Ave, but there's a 1 Ave, a 2 Ave, a 1 St, and a 1 St North. Definitely no 90th Ave.
https://pypi.python.org/pypi/fake-factory/
The most useful tool besides this is http://10minutemail.com/
Edit: Doing a HN search this was posted about a year ago. That is why I probably didn't post it. Should we start submitting reposts of old stuff? I probably missed "cool tools" from years ago.
-Female
-Fitness instructor
-Weighing 205.9 lbs
-Standing 5 feet 1 inch
Individually, any of these things might be ok. Any three could even be possible. All four, however, just doesn't seem to work.
Then there's the automobile, which could also raise red flags. A Fiat Tempra driven in the USA would probably be suspicious to anyone who knows that Fiats probably weren't sold in the US in 1994, for example. And it might be weird for a clerk to be driving a 2010 Infinity...
Back in the BBS days there used to be programs written to auto-generate identities in bulk, for reasons i'm not aware of. They were designed to minimize scrutiny because there might be humans actually looking at the data you used, since there was less automation in terms of processing accounts back then.