|
Lumpy posted:So one of our favorite derails has been settled. Turns out doing_it_like_this is easier on the eye and faster to read thanDoingItLikeThis. Well, they are both better than \doingitlikethis as you get in LaTeX. (It's good that people are coming up with actual evidence for using one thing over another, maybe people doing that kind of thing will reduce the number of stupid arguments we get about ~coding styles~.)
|
# ? Aug 30, 2013 15:17 |
|
|
# ? May 28, 2024 09:17 |
|
You are naive if you think objective measurements will change people's opinions.
|
# ? Aug 30, 2013 15:24 |
|
bucketmouse posted:Oh the horror stories I could tell about one of my upper division cs professors. Have you ever calculated a md5 by hand, on paper? I have. If it makes the course more difficult, it's a more valuable program, right?
|
# ? Aug 30, 2013 15:30 |
|
I'm glad that study exists, because now it means I have a reason to pick a style and stick with it in solo projects.
|
# ? Aug 30, 2013 15:31 |
|
Even tech companies that have been around forever don't do passwords right: Granted, 16 is a lot better than 8, but why the hell is there a cap anyway? It's just going to be hashed (hopefully) so it's not like you're going to be burning storage space with longer passwords.
|
# ? Aug 30, 2013 15:32 |
|
Bognar posted:Even tech companies that have been around forever don't do passwords right: A limit on password length DOES make sense but only from a computational standpoint. There is no logical reason to forbid a 256 character password other than if you're using a lovely cipher. With that said, lovely ciphers are those that are not bcrypt or scrypt for example.
|
# ? Aug 30, 2013 16:27 |
|
I wouldn't recommend bcrypt without pointing out that it truncates excessively-long passphrases which is an awfully subtle trap compared to PBKDF2 which seems to just work.
Crosscontaminant fucked around with this message at 16:58 on Aug 30, 2013 |
# ? Aug 30, 2013 16:56 |
|
Here's to magically breaking APIs and no interest in figuring out what the gently caress is wrong with it. Seriously. We have a site where we need to create product orders through a 3rd party api. Said api randomly breaks, and once it breaks, it will always fail until they do something about it. When we let them know it breaks, it's a series of "I've restarted it, is it fixed/no it isnt fixed" emails. Sometimes it takes up to 20 restarts before its magically fixed. Sometimes it magically breaks again after we send just a couple orders over. And this has been going on for half a year. The last couple months it happens at least 2 times/week.
|
# ? Aug 30, 2013 21:08 |
|
Freakus posted:Here's to magically breaking APIs and no interest in figuring out what the gently caress is wrong with it. I would consider finding a new 3rd party.
|
# ? Aug 30, 2013 22:07 |
|
Wow, someone doesn't write tests. 20 restarts is a whole lot of money lost. Are you the only user or something?
|
# ? Aug 31, 2013 07:57 |
|
Are you paying them money? If you are, then what the hell are you doing giving them money without having an actionable SLA?
|
# ? Aug 31, 2013 08:17 |
|
I can almost guarantee that if they are paying money, there's a more reliable service at the same price (likely even better value with the SLA) too. Ditch that poo poo unless you're absolutely forced to keep it, downtime in a purchasing/product order environment can cost a lot more than the obvious "x number of sales were delayed" metric.
|
# ? Aug 31, 2013 08:28 |
|
Lumpy posted:So one of our favorite derails has been settled. Turns out doing_it_like_this is easier on the eye and faster to read thanDoingItLikeThis. I've always preferred snake_case(it's Pythonic and therefore superior) but it's annoying when working with camelCased libraries, combining those with snake_case just looks off to me. Although maybe it's a good thing that it's a quick visual way to distinguish between library code and my code. Would be interesting to see a similar study done on colorschemes and fonts just to see if there's a significant productivity/performance difference assuming a non-insane choice.
|
# ? Aug 31, 2013 10:50 |
|
Bognar posted:Even tech companies that have been around forever don't do passwords right: They use a block cypher that takes 128 bits, and a previous implementation ran multiple blocks but their new one only uses one?
|
# ? Aug 31, 2013 14:48 |
|
I wrote something that was supposed to be like 20 lines of python and run from the command line. Who needs tests for that? It's slowly morphed in to a few thousands lines of code, running on two servers communicating over a REST API, and still no tests. I hate myself.
|
# ? Aug 31, 2013 15:07 |
|
Just remember, ABC. Always Be Ctesting.
|
# ? Aug 31, 2013 15:28 |
|
I don't know what the original arrangement is, we've been with them for about 10 years now. We're the only ones using this particular api, and we don'thave nearly as much trouble with the other ones of theirs we use. It's their products we're selling, we just get a cut of it.
|
# ? Aug 31, 2013 16:56 |
|
yaoi prophet posted:Just remember, ABC. Always Be Ctesting. After Building, Check!
|
# ? Aug 31, 2013 16:57 |
|
Sticky Profits posted:From the syllabus: Alice enters class on time, looks at the quiz, and goes to the bathroom. In the bathroom, Alice texts Bob, outside, with the quiz questions. Bob looks at the text(s), finds the answers, and texts them back. Alice and Bob (re-)enter the classroom and ace the quiz. Easy. Lumpy posted:So one of our favorite derails has been settled. Turns out doing_it_like_this is easier on the eye and faster to read thanDoingItLikeThis. Having read through this, I appreciate the work the authors put into establishing statistical significance, but I still have a very hard time taking a study of 15 people seriously.
|
# ? Aug 31, 2013 23:23 |
|
code:
|
# ? Sep 2, 2013 03:21 |
|
I am a MUMPS programmer -- Ask me anything
|
# ? Sep 2, 2013 08:52 |
|
On the subject of MUMPS, here's another specialized business language.
|
# ? Sep 2, 2013 08:55 |
|
bucketmouse posted:A bit of a derail, but does anyone know a good standalone XML doc generator for C++? I'm staring at something like 200 completely undocumented lua-c bindings and beginning to feel nauseous. We use doxygen at work for generating documentation from c++. http://www.stack.nl/~dimitri/doxygen/
|
# ? Sep 2, 2013 11:02 |
|
redleader posted:On the subject of MUMPS, here's another specialized business language. That sure is special, all right How is your first thought when you see that not to make mnemonics for all those numbers?
|
# ? Sep 2, 2013 17:38 |
|
This is the kind of poo poo you come across frequently in the hobbyist (Arduino) community. Everything is an unreadable unmaintainable poorly formatted loving mess.code:
|
# ? Sep 2, 2013 17:43 |
|
status posted:Seriously, who the gently caress puts comments like this? Any change to the code (to fix the horrendous style or to make changes) makes it even worse than it already is... Looks like someone was brought up on assembly.
|
# ? Sep 2, 2013 17:49 |
|
EntranceJew posted:Looks like someone was brought up on assembly. Even in assembly I don't often see large chunks of code with ; and no actual comment after it though.
|
# ? Sep 2, 2013 17:52 |
|
redleader posted:On the subject of MUMPS, here's another specialized business language. To be fair, it sounds like it was a glorified low-level VM language not meant for human consumption, only the tooling was so bad they made the decision to program it directly.
|
# ? Sep 2, 2013 20:04 |
|
EntranceJew posted:Looks like someone was brought up on assembly.
|
# ? Sep 2, 2013 23:44 |
|
status posted:This is the kind of poo poo you come across frequently in the hobbyist (Arduino) community. Everything is an unreadable unmaintainable poorly formatted loving mess. I'm not defending this particular style, but it looks like the empty comments signify the code described by the nonempty comment at the top. Like #region in C#.
|
# ? Sep 2, 2013 23:49 |
|
pokeyman posted:I'm not defending this particular style, but it looks like the empty comments signify the code described by the nonempty comment at the top. Like #region in C#. Huh, I had no idea why they did that when I first saw it, but now that you mention it, I've done the same thing when writing assembler. I think it makes sense there; it just looks bizarre in C.
|
# ? Sep 3, 2013 00:52 |
|
pokeyman posted:I'm not defending this particular style, but it looks like the empty comments signify the code described by the nonempty comment at the top. Like #region in C#. Yeah, that's definitely the reason. Could have gone about it a bit better though.
|
# ? Sep 3, 2013 03:34 |
|
status posted:Seriously, who the gently caress puts comments like this? Any change to the code (to fix the horrendous style or to make changes) makes it even worse than it already is... That's easy to fix, just put the comments before the code.
|
# ? Sep 3, 2013 15:01 |
|
This tweet:https://twitter.com/0x17h/status/374818579339698176 posted:github search is the gift that keeps on giving https://github.com/search?q=id_rsa+in:path&type=Code&ref=searchresults (thanks @casiotone!) Edit: Someone pointed out the description of this one: https://github.com/philchristensen/txsftp posted:☣ A highly secure SFTP implementation, supporting virtualized users, a chroot'ed environment, automatic GPG encryption and VPN detection. Jewel fucked around with this message at 15:12 on Sep 3, 2013 |
# ? Sep 3, 2013 15:06 |
|
Thermopyle posted:I wrote something that was supposed to be like 20 lines of python and run from the command line. Who needs tests for that? Never show a business person a working prototype because that poo poo will become production before you can sit down.
|
# ? Sep 3, 2013 20:52 |
|
Monkeyseesaw posted:Never show a business person a working prototype because that poo poo will become production before you can sit down. This is the story of one of our nearly unmaintainable applications. We stood up a staging site for them to poke around at and make suggestions/critiques. A day later they sent us an e-mail saying "Looks great, we've already got 5 clients using it!"
|
# ? Sep 3, 2013 21:25 |
|
Bognar posted:This is the story of one of our nearly unmaintainable applications. We stood up a staging site for them to poke around at and make suggestions/critiques. A day later they sent us an e-mail say "Looks great, we've already got 5 clients using it!" We had a client sneak a development build into production by putting it in an iframe on their live site. Without telling us, of course.
|
# ? Sep 3, 2013 21:30 |
|
Monkeyseesaw posted:Never show a business person a working prototype because that poo poo will become production before you can sit down.
|
# ? Sep 3, 2013 21:36 |
|
DaTroof posted:We had a client sneak a development build into production by putting it in an iframe on their live site. Without telling us, of course. Welp, I just added a clause to my standard contract form.
|
# ? Sep 3, 2013 22:01 |
|
|
# ? May 28, 2024 09:17 |
|
redleader posted:On the subject of MUMPS, here's another specialized business language. That lead me indirectly to this. https://github.com/jloughry/BANCStar/blob/master/C16LNAPP.pdf?raw=true e: if nothing else, ctrl-f "Interesting features" on that page e2: whoops wrong link bucketmouse fucked around with this message at 19:47 on Sep 4, 2013 |
# ? Sep 4, 2013 08:57 |