EPTPlus I'm always wary of just visually hiding things people shouldn't have access to. Always feels like there's more margin for mistakes or unintentionally security issues that way, even if it's not.
The Big Programming Thread - Page 905
| Forum Index > General Forum |
Thread Rules 1. This is not a "do my homework for me" thread. If you have specific questions, ask, but don't post an assignment or homework problem and expect an exact solution. 2. No recruiting for your cockamamie projects (you won't replace facebook with 3 dudes you found on the internet and $20) 3. If you can't articulate why a language is bad, don't start slinging shit about it. Just remember that nothing is worse than making CSS IE6 compatible. 4. Use [code] tags to format code blocks. | ||
|
WolfintheSheep
Canada14127 Posts
Plus I'm always wary of just visually hiding things people shouldn't have access to. Always feels like there's more margin for mistakes or unintentionally security issues that way, even if it's not. | ||
|
sc-darkness
856 Posts
| ||
|
emperorchampion
Canada9496 Posts
On September 09 2017 20:57 sc-darkness wrote: I had an eye test. As usual, they recommend 20-minute break from computers after so many hours (I forgot). Does anyone do that? :D I feel like no employer would say yes to that. Aren't modern monitors good enough? Do you have a lunch break lol? Also every 20 min you should focus your eyes on something far away for a minute or so | ||
|
Manit0u
Poland17197 Posts
| ||
|
Deleted User 3420
24492 Posts
On September 09 2017 20:57 sc-darkness wrote: I had an eye test. As usual, they recommend 20-minute break from computers after so many hours (I forgot). Does anyone do that? :D I feel like no employer would say yes to that. Aren't modern monitors good enough? Don't you have any sort of work you can do that doesn't take looking at the monitor? Also it's probably a good idea anyways. Get up, walk around, stretch, etc. | ||
|
Isualin
Turkey1903 Posts
On September 09 2017 20:57 sc-darkness wrote: I had an eye test. As usual, they recommend 20-minute break from computers after so many hours (I forgot). Does anyone do that? :D I feel like no employer would say yes to that. Aren't modern monitors good enough? There is a time management technique called Pomodoro which is basically working 25 minutes and a 5 minutes break after that. It has a lot of implementations as apps or websites. As travis said, it is a good idea to take a walk, drink water, talk with a colleague etc | ||
|
sc-darkness
856 Posts
Either way, any successful H1B visa holders here? No one said yes last time. | ||
|
sabas123
Netherlands3122 Posts
| ||
|
Manit0u
Poland17197 Posts
For instance, when you create user model you must pass "password" and "password_confirmation" fields, and they must match. When you update it you also need to pass "current_pussword" (none of them being "encrypted_password" which is actually digested and saved to the db). When you pass "password_confirmation" as "nil/null" or you don't pass it at all, it turns off all passwrord/confirmation matching validations. WTF?! This is one of the most ridiculous things I've ever seen. Especially that it's at the very core of the framework and there's plenty of authn/authr libraries that make use of this feature. I've seen a lot of bullshit in my life, but right now this takes the cake. | ||
|
PM_ME_NICE_PUPPERS
Pakistan51 Posts
On September 09 2017 20:57 sc-darkness wrote: I had an eye test. As usual, they recommend 20-minute break from computers after so many hours (I forgot). Does anyone do that? :D I feel like no employer would say yes to that. Aren't modern monitors good enough? At SAP, we were allowed to take a morning walk and an afternoon walk around the local park. Staring at screens, sitting down, isn't great for your health, doing so without regular breaks is devastating. | ||
|
WarSame
Canada1950 Posts
I've been wondering if a "rolling" blockchain is possible, where some computers store only the last X blocks, and the rest store the whole blockchain, but that seems like it might be vulnerable. Is it? Are there better ways? | ||
|
Hanh
146 Posts
| ||
|
enigmaticcam
United States280 Posts
The general guidelines with OOP and classes are to keep logic well organized and simple. Objects should generally do no more than one thing. But what if you're building something where optimization is paramount? For example, suppose you're building a chess engine. Perhaps you might have some logic that calculates all the possible moves for a given piece. If that piece is a rook, the calculation would be a bit different than say a knight. Proper OOP would require something like an abstracted piece class that defines a single interface for calling that logic, and then you would have different instantiated objects that inherit from that piece class that each perform their specific logic based on that piece type. But rather than having to instantiate new classes just to perform this logic for every new position when you're having to analyzing literally millions of positions as fast as possible, maybe it's better to just use the trusty old case/switch statement and route that logic to a function in a class that's already loaded? Obviously I can do what I want, but just curious what your opinions are. Thanks! | ||
|
netherh
United Kingdom333 Posts
On September 15 2017 02:56 enigmaticcam wrote: Question on OOP and optimization: The general guidelines with OOP and classes are to keep logic well organized and simple. Objects should generally do no more than one thing. But what if you're building something where optimization is paramount? For example, suppose you're building a chess engine. Perhaps you might have some logic that calculates all the possible moves for a given piece. If that piece is a rook, the calculation would be a bit different than say a knight. Proper OOP would require something like an abstracted piece class that defines a single interface for calling that logic, and then you would have different instantiated objects that inherit from that piece class that each perform their specific logic based on that piece type. But rather than having to instantiate new classes just to perform this logic for every new position when you're having to analyzing literally millions of positions as fast as possible, maybe it's better to just use the trusty old case/switch statement and route that logic to a function in a class that's already loaded? Obviously I can do what I want, but just curious what your opinions are. Thanks! [Not too much knowledge about chess engines here, so this is quite speculative...] But... assuming you need a function for a piece that takes a position and returns the set of possible outcome positions after moving that piece. You could use a "Ruleset" class overloaded for each type of piece, or a set of free functions and a switch statement, whichever is faster (they'd be effectively doing the exact same thing). By "new" and "already loaded" are you referring to dynamic allocation? I don't think you'd need to dynamically allocate inherited classes, as you'd only need one Ruleset instance for each type of piece, not for every piece instance. You probably don't need an inheritance hierarchy for the Piece class itself. You might not even want an instance of such a class. Just an enum, and some conversion functions (for names, notation, etc.) might do. In general, I suggest thinking about the specific data you need (coordinates / type of pieces), and what the algorithms do with it. Then think about container types and how to group the data efficiently. At that point you have your objects. i.e. Don't think: "Well a chess board has pieces on it of different types, so I need a Piece class and an inheritance chain with Pawn, Knight, etc.". Think something like: "my algorithm needs a set of coordinates (int, int), and piece types (enum), that I process in [...] way, and they need to be stored in a [map / vector / list]". | ||
|
enigmaticcam
United States280 Posts
On September 15 2017 04:22 netherh wrote:i.e. Don't think: "Well a chess board has pieces on it of different types, so I need a Piece class and an inheritance chain with Pawn, Knight, etc.". Think something like: "my algorithm needs a set of coordinates (int, int), and piece types (enum), that I process in [...] way, and they need to be stored in a [map / vector / list]". Ah yes, that's a better way to organize it. Sometimes I get those logical blinders on and it takes a fresh pair of eyes to help me see it from a different perspective. Thank you! | ||
|
CecilSunkure
United States2829 Posts
On September 15 2017 02:56 enigmaticcam wrote: Question on OOP and optimization: The general guidelines with OOP and classes are to keep logic well organized and simple. Objects should generally do no more than one thing. But what if you're building something where optimization is paramount? For example, suppose you're building a chess engine. Perhaps you might have some logic that calculates all the possible moves for a given piece. If that piece is a rook, the calculation would be a bit different than say a knight. Proper OOP would require something like an abstracted piece class that defines a single interface for calling that logic, and then you would have different instantiated objects that inherit from that piece class that each perform their specific logic based on that piece type. But rather than having to instantiate new classes just to perform this logic for every new position when you're having to analyzing literally millions of positions as fast as possible, maybe it's better to just use the trusty old case/switch statement and route that logic to a function in a class that's already loaded? Obviously I can do what I want, but just curious what your opinions are. Thanks! You're over thinking it. Typically people that use OOP apply it everywhere whether or not it helps. If we're solving some kind of chess optimization problem, OOP probably won't help. | ||
|
Manit0u
Poland17197 Posts
Some reading: https://www.saturnflyer.com/blog/oop-dci-and-ruby-what-your-system-is-vs-what-your-system-does http://www.artima.com/articles/dci_vision.html https://stackoverflow.com/questions/378629/dci-data-context-and-interaction-successor-to-mvc | ||
|
WarSame
Canada1950 Posts
On September 14 2017 16:50 Hanh wrote: You can instruct your client software to prune the data which roughly does what you want. Thank you, that led me in the right direction. It's encouraging to know that a technical problem didn't just ruin what I wanted to do  | ||
|
spinesheath
Germany8679 Posts
On September 15 2017 05:29 Manit0u wrote: You will get the most performance in OOP by adopting the DCI (Data Context Interaction) architecture. I'm not sure if you can do that in Java. C#, Scala, Python and Ruby support this in various ways, most commonly traits that you inject into objects during or after instantiation (run time, not compile time) - only when they need them. Some reading: https://www.saturnflyer.com/blog/oop-dci-and-ruby-what-your-system-is-vs-what-your-system-does http://www.artima.com/articles/dci_vision.html https://stackoverflow.com/questions/378629/dci-data-context-and-interaction-successor-to-mvc Interesting. I've been thinking about concepts like this to fix some really annoying architectural problems at work but I haven't stumbled across DCI yet. Good to know that it's an established pattern. Probably won't be able to apply it large-scale anytime soon, obviously... | ||
|
Manit0u
Poland17197 Posts
Validating paths is a bitch:
The regex works fine, if not for one tiny little quirk. Matching longer strings results in catastrophic backtracking. Like this one:
I know it's caused by the negative look behind, but I don't know how to rectify this  | ||
| ||
