|
http://www.kalzumeus.com/2010/06/17/falsehoods-programmers-believe-about-names/quote:I have never seen a computer system which handles names properly and doubt one exists, anywhere.
|
# ? Oct 25, 2012 11:53 |
|
|
# ? May 21, 2024 07:32 |
|
My rule for names is that if I can't read your name enough to even attempt to pronounce it, then I don't care if my system breaks on it. e: pretty much the same rules that would apply if I met someone in person.
|
# ? Oct 25, 2012 12:04 |
|
The Gripper posted:My rule for names is that if I can't read your name enough to even attempt to pronounce it, then I don't care if my system breaks on it. Similar article: Falsehoods programmers believe about time and part 2. e: the SA forums are gross if you use unicode Malloc Voidstar fucked around with this message at 12:27 on Oct 25, 2012 |
# ? Oct 25, 2012 12:24 |
|
Are you endorsing this article, or pointing it out for derision, or just presenting it as food for thought? Admittedly it might be very different to work on a system that's supposed to store and use people's real names (as opposed to their sign-in names for a service), but admonishing the assumption that people's names can be expressed in unicode, without giving suggestions as to a way to avoid making such assumptions, just reads like a load of self-important wank rather than serious discussion. I mean, the author points out a number of genuine pitfalls but I get the impression that he thinks that if someone comes along and tells you that their name is , or , your system should say "sure, no problem!" His list of peeves is of no use without some suggestion as to how to avoid them, given that some of them would be all but impossible practically to avoid in a working system. The time articles linked by Aleksei have much the same issue. I would be interested in reading about why it is a misconception, as it is presented in part 2, that every year has either 365 or 366 days, but no explanatory link is provided.
|
# ? Oct 25, 2012 15:12 |
|
Hammerite posted:The time articles linked by Aleksei have much the same issue. I would be interested in reading about why it is a misconception, as it is presented in part 2, that every year has either 365 or 366 days, but no explanatory link is provided. The switch from julian to gregorian. code:
|
# ? Oct 25, 2012 15:16 |
|
Hammerite posted:Are you endorsing this article, or pointing it out for derision, or just presenting it as food for thought? Admittedly it might be very different to work on a system that's supposed to store and use people's real names (as opposed to their sign-in names for a service), but admonishing the assumption that people's names can be expressed in unicode, without giving suggestions as to a way to avoid making such assumptions, just reads like a load of self-important wank rather than serious discussion. I mean, the author points out a number of genuine pitfalls but I get the impression that he thinks that if someone comes along and tells you that their name is , or , your system should say "sure, no problem!" quote:All of these assumptions are wrong. Try to make less of them next time you write a system which touches names. See sites that don't accept apostrophes or hyphens, or apply length requirements.
|
# ? Oct 25, 2012 15:24 |
Hi, my name is 37F89DA9-612C-498C-AFA5-AB9DF762E55F, but my friends call me 37.
|
|
# ? Oct 25, 2012 15:34 |
|
PHP code:
|
# ? Oct 25, 2012 15:45 |
|
Hammerite posted:Are you endorsing this article, or pointing it out for derision, or just presenting it as food for thought? Admittedly it might be very different to work on a system that's supposed to store and use people's real names (as opposed to their sign-in names for a service), but admonishing the assumption that people's names can be expressed in unicode, without giving suggestions as to a way to avoid making such assumptions, just reads like a load of self-important wank rather than serious discussion. He is, in fact, talking about real names and not account names. A lot of systems expect your real name in addition to your account name (or will attempt to derive your account name from your real name, or will just assign you an account number or something). And a surprising number of those will choke and die if your name has an apostrophe in it, or a hyphen, or non-ASCII characters of any sort. Or it'll work fine, but changing your name later is either outright impossible or causes other things to break. And so on. Yeah, it would be nice to have more information on how to avoid that issues, but it's a rant, not a lesson.
|
# ? Oct 25, 2012 16:12 |
|
ToxicFrog posted:He is, in fact, talking about real names and not account names. A lot of systems expect your real name in addition to your account name (or will attempt to derive your account name from your real name, or will just assign you an account number or something). And a surprising number of those will choke and die if your name has an apostrophe in it, or a hyphen, or non-ASCII characters of any sort. Or it'll work fine, but changing your name later is either outright impossible or causes other things to break. And so on. If I legally get my name changed to Jewel I have a lot of problems too because sites like to say "OH NO YOUR NAME HAS 'JEW' IN IT, WE CAN'T ACCEPT THIS". A lot of sites currently do this even with my name as a Username, I've gotten banned from BigPond's gamearena TF2 servers because Jewel's an "offensive name" (I had to contact them to get it whitelisted, took a month), and I couldn't register to a few other services with the name either for the same reason. Never ever put a restriction on usernames, just let users report offensive ones if it's in an environment where it matters.
|
# ? Oct 25, 2012 16:26 |
|
Excuse me but I come from a long line of Bonerhitlers,
|
# ? Oct 25, 2012 16:43 |
|
Plus there's weird barely-known countries like Australia, where if you're open and notorious with using an assumed name it's considered legal, and there's no formal process required to use it wherever you want. e; the only caveat is a few government agencies require written proof to issue you certain documents, but that's not necessary for the name to be legal. If I openly called myself Jewbag McNegrofingers then I could have that name considered as my legal name, despite it being dumb and racisty. I'm pretty sure you'd get no objections to banning me from your service for it though, since you're not obligated to provide me any services and I'd obviously be a worthless person. The Gripper fucked around with this message at 16:47 on Oct 25, 2012 |
# ? Oct 25, 2012 16:44 |
|
Aleksei Vasiliev posted:Similar article: Falsehoods programmers believe about time and part 2. Similarly, http://qntm.org/gay
|
# ? Oct 25, 2012 17:01 |
|
tef posted:Similarly, http://qntm.org/gay There's a proposed genealogy data exchange format being developed by FamilySearch, a mormon company, and (surprisingly?) it will support gay marriages The old standard (Gedcom) has hardcoded HUSB/WIFE fields; GedcomX Couple relationship types will allow any two persons. https://github.com/FamilySearch/gedcomx/blob/master/specifications/conceptual-model-specification.md
|
# ? Oct 25, 2012 17:54 |
|
Manslaughter posted:Hi, my name is 37F89DA9-612C-498C-AFA5-AB9DF762E55F, but my friends call me 37. Any relation to Tom Lehrer's friend Hen3ry?
|
# ? Oct 25, 2012 18:00 |
|
Golbez posted:
Every time I try to justify to myself that PHP, while terrible, is at least perfectly usable if you're aware of some common problems, something like this comes along. It's not even that it's different. It's that it keeps changing back and forth for newer versions.
|
# ? Oct 25, 2012 18:11 |
|
code:
|
# ? Oct 25, 2012 18:37 |
|
Golbez posted:
Seems like it might be a bug that's been introduced in 5.1.3. I don't see why anyone would ever want to do stuff like that in PHP unless it was a practical joke, though. There's always the possibility that somewhere deep in the bowels of the PHP documentation, there's a note saying you can't actually define constants using reserved words too. "Nope, guys. Working as intended!" - Either way, you can't really call it a coding horror if you're trying to alter something like the truthiness of TRUE at runtime, and you get unexpected results. That kind of nonsense might have worked in days gone by, but we now we have evolved sensibilities, man! I'm thinking that TRUE and FALSE aren't actually constants, and calling "define("TRUE", $some_val);" creates a constant that will always be trumped by TRUE in the interpreter. Note: I'm not defending PHP in general, I'm defending it in this specific case. There are legitimate horrors in that language, but we can't go around calling things horrors just because they don't behave the way we expect them to.
|
# ? Oct 25, 2012 18:40 |
I believe "true" is a keyword in PHP, and keywords are case insensitive. So in version 5.1.3 defines might have been made to resolve after keywords are checked for.
|
|
# ? Oct 25, 2012 19:11 |
|
Aleksei Vasiliev posted:No: I feel it is wrong to call these people programmers. Although there is also a very good chance they are programmers but their boss is a micromanaging idiot. However the unicode thing is not an assumption, it is the point where you throw up your hands and go "look, you're just going to have to meet me in the middle here, wtf do I put in the email template 'Dear _____,' and no; I'm not putting an image there."
|
# ? Oct 25, 2012 19:41 |
|
Golbez posted:
Stuff like this is going to happen as PHP devs try to make PHP behave more consistently. Here is the entry for it in the changelog for 5.1.3 quote:* Eliminated run-time constant fetching for TRUE, FALSE and NULL I do not see this as a Coding Horror, more like it's trying to correct a previous Coding Horror. Considering this change was made to PHP more than 6 years ago and it's doing something that no sane coder would do, I don't think it's that big of a deal.
|
# ? Oct 25, 2012 19:42 |
|
Zamujasa posted:
According to the linked page it doesn't? It gives one result for versions 5.1.2 and earlier, another result for versions 5.1.3 and later
|
# ? Oct 25, 2012 19:55 |
|
Carthag posted:There's a proposed genealogy data exchange format being developed by FamilySearch, a mormon company, and (surprisingly?) it will support gay marriages You'd think they'd generalize it to handle polygamy?
|
# ? Oct 25, 2012 20:10 |
|
nielsm posted:I believe "true" is a keyword in PHP, and keywords are case insensitive. So in version 5.1.3 defines might have been made to resolve after keywords are checked for. No, if you try to code:
So the horror here is that it's allowing you to make a variable definition, not notifying you that it won't work, and then ignoring it. The example itself was stupid (redefining true as false), but I can think of a semi legitimate reason to make a define that would say, define TRUE as a string 'true', or as 1, or as a number of things to deal with things like input parsing. I'm not saying that it's a good thing to do, but I'm sure that someone, somewhere, would think that it's a good idea, and when that happens, PHP should either let them do it, or be consistent with notifying them that it won't work.
|
# ? Oct 25, 2012 20:15 |
|
Horrors all the way down https://crypto.stanford.edu/~dabo/pubs/abstracts/ssl-client-bugs.html
|
# ? Oct 25, 2012 20:17 |
|
ohgodwhat posted:You'd think they'd generalize it to handle polygamy? Yeah actually that'd make sense. Though I guess they'd just model a polygamous person by having them in several concurrent marriages. Depends on the exact nature of the marriage I guess, I mean if it's an "equal" one where everybody is married to everybody, or it's one dude with 5 wives. I should open an issue on github
|
# ? Oct 25, 2012 20:22 |
|
The Gripper posted:My rule for names is that if I can't read your name enough to even attempt to pronounce it, then I don't care if my system breaks on it. Not a fan of the Chinese, huh
|
# ? Oct 25, 2012 22:32 |
|
The Gripper posted:My rule for names is that if I can't read your name enough to even attempt to pronounce it, then I don't care if my system breaks on it. How often are you reading someone's name in person? Don't you just repeat their name after they introduce themselves? Do you live in a conference centre?
|
# ? Oct 25, 2012 23:48 |
|
I spend most of my free time at AA, reading nametags.
|
# ? Oct 25, 2012 23:51 |
|
The Gripper posted:I spend most of my free time at AA, reading nametags. And I'm sure the frequency of members using unpronounceable symbols to identify themselves is a problem that destroys meetings.
|
# ? Oct 26, 2012 01:50 |
|
Carthag posted:Yeah actually that'd make sense. Though I guess they'd just model a polygamous person by having them in several concurrent marriages. Depends on the exact nature of the marriage I guess, I mean if it's an "equal" one where everybody is married to everybody, or it's one dude with 5 wives. I can't wait until people have to learn graph theory to understand marriages.
|
# ? Oct 26, 2012 02:40 |
|
yaoi prophet posted:I can't wait until people have to learn graph theory to understand marriages. Is that when the Deep South allows homosexual polygamy? "Well, I'm married to my grandma, nephew, sister, father and myself (and my pig but he don't count)."
|
# ? Oct 26, 2012 04:36 |
|
yaoi prophet posted:I can't wait until people have to learn graph theory to understand marriages. classic: http://stackoverflow.com/questions/6163683/cycles-in-family-tree-software
|
# ? Oct 26, 2012 04:48 |
|
God of Mischief posted:Is that when the Deep South allows homosexual polygamy? I suppose then the rule will be that one must be able to construct a spanning tree for each family. Edit: crap, Deus Rex kinda beat me to it.
|
# ? Oct 26, 2012 04:50 |
|
1337JiveTurkey posted:The switch from julian to gregorian. That was only the switch in the UK & subsidiaries, and I think they were late because the Gregorian calendar was papist. (and didn't places Greece switch really late, like early 20th century?) Otto Skorzeny posted:The idea that + should be commutative is not remotely unreasonable It's clearly not always, though. Floating point isn't guaranteed to be commutative.
|
# ? Oct 26, 2012 04:53 |
|
fritz posted:It's clearly not always, though. Floating point isn't guaranteed to be commutative. Floating point underflow and related issues are to be expected though.
|
# ? Oct 26, 2012 05:07 |
|
Floating point + clearly isn't associative, but I can't think of anything off the top of my head where it's not commutative.
|
# ? Oct 26, 2012 05:08 |
|
fritz posted:That was only the switch in the UK & subsidiaries, and I think they were late because the Gregorian calendar was papist. (and didn't places Greece switch really late, like early 20th century?) http://en.wikipedia.org/wiki/February_30#Swedish_calendar the swedish changeover was the best
|
# ? Oct 26, 2012 05:11 |
|
Jabor posted:Floating point + clearly isn't associative, but I can't think of anything off the top of my head where it's not commutative. String concatenation in languages that use it as such, for one obvious (but vaguely cheaty ) answer
|
# ? Oct 26, 2012 05:31 |
|
|
# ? May 21, 2024 07:32 |
|
Ahaha, I found some good addon for Chrome that I wanted to use, but it had a lot of features that broke some stuff, where I only wanted one of the features, so I downloaded the .crx and extracted it with winrar, giving me access to all the code so I could just change it and put it back into extensions. This line of code is in one of the files, I have no idea what it does:JavaScript code:
JavaScript code:
Edit: Oh, right, debugging exists, could change it to 1==1 whenever they wanted to debug. Carry on!! Jewel fucked around with this message at 05:39 on Oct 26, 2012 |
# ? Oct 26, 2012 05:34 |