|
pointers posted:np complete. Yeah, um, you don't have a clue, you're trying to pattern-match your way into seeming like you do by throwing in technical terms and it's not working because their use is completely nonsensical.
|
#
?
Jun 8, 2014 07:00
|
|
|
|
| # ? May 22, 2024 19:01 |
|
|
pointers, putting aside the fact that NP-hardness has nothing to do with what we were chatting about, please tell me more about how the human brain's special structure allows it to efficiently solve NP-complete problems.
|
|
#
?
Jun 8, 2014 12:32
|
|
|
it's too soon to know if that's true or false. i'd bet operating on heuristics helps a lot (because that's what you'd use while programming for np problems), but that's just like, my opinion, man. brain and turing have different strengths. brains do well at sensory processing/associative memory. turing architecture does much better with numerical processing. neuromorphic chips could make the tasks brains are strong at easier for computers, they still differ in that they do not store all memory in registers. that is probably not a limitation, but it's too soon to tell. it could be that it's only the amount of processing power that make brains better at these problems (which i think is your point?). that's also opinion, for now, because computers aren't going to reach that amount of processing power anytime soon.
|
|
#
?
Jun 8, 2014 16:21
|
|
|
Real talk tho We can legitimately say that our brains function nothing like a Von Neumann architecture machine. There's a fusion of data storage and processing that creates behavior quite different from what we're used to thinking about in computer science. I also think you could make an argument that this structure leads to having only the capability to carry out probabilistic and relational calculations in way that leaves us as a non-Turing complete machine due to being error prone and thus unable to reliably simulate a Turing machine. A non-Turing machine capable of creating Turing machines, though. Which really, you could argue would mean that since we are able to produce something with those properties that we, in fact, inherit those properties. Therefore, my brain IBM QED
|
|
#
?
Jun 8, 2014 18:11
|
|
|
pointers posted:it's too soon to know if that's true or false. i'd bet operating on heuristics helps a lot (because that's what you'd use while programming for np problems), but that's just like, my opinion, man. You should try having reasons for your opinions. It's a good strategy for being right. I would also recommend identifying a single NP-hard problem that brains are better at than computers.
|
|
#
?
Jun 8, 2014 18:41
|
|
|
pointers posted:it's too soon to know if that's true or false. i'd bet operating on heuristics helps a lot (because that's what you'd use while programming for np problems), but that's just like, my opinion, man. You're missing a lot of fundamental, important points. First of all, every computer we know of, theoretical or real (with the possible exception of the human brain), is equivalent to every other computer. If you assume infinite memory and time, you can emulate a Playstation 4 or 128-core server on an 25 cent PIC or a Turing machine that doesn't have registers at all. When we talk about computability, we're just talking about what machines or systems can come to the same result. And, unless you believe that there's something about biological processes that makes them impossible to simulate on a computer, a computer can do everything a human brain can do. And, Turing machines can use heuristics, too. NP-complete and NP-hard problems are often only that difficult for an exact answer - there are all kinds of algorithms for the traveling salesman problem, for instance, which run in fairly short polynomial time and give either a near-optimum route with high probability. Yes, different types of machines are more efficient at different things. But, one of the key questions in computability theory is "can we do this at all?" When it comes to emulating the human brain, or anything that it can do, the answer is almost certainly "yes." Dairy Power posted:Real talk tho Turing machines are radically different from von Neumann machines, too. Computability theory doesn't give a drat. They might be very different to actually program, but you can get them to give you equivalent results, and that's all that matters.
|
|
#
?
Jun 8, 2014 18:57
|
|
|
Anything that a Turing Machine can compute a human can compute with enough pen, paper and patience. There are differences in the efficiency with which a brain and a Von Neumann machine can run certain types of computations. Thinking purely about what can be computed and ignoring the energy and time cost of computing the thing is a bit silly. I don't know of any good studies comparing the speed and energy cost of various types of computation between humans and computers. As far as I can tell the human brain is many orders of magnitude more energy/time efficient at pattern recognition than computers, computers are better at simple math(and everything that we can build on top of that). That's about all we really have at the moment.
|
|
#
?
Jun 8, 2014 19:30
|
|
|
Space Gopher posted:Turing machines are radically different from von Neumann machines, too. Computability theory doesn't give a drat. They might be very different to actually program, but you can get them to give you equivalent results, and that's all that matters. Yup, I was never claiming otherwise. My first thought was that our brain architecture really isn't comparable to what we usually think about in computer science classes, so direct comparisons to the type of algorithms you'd write for a computer is fruitless. My second thought was that our brains, due to the likely probabilistic nature of our thought processes, and, thus, inherent proneness to error, can't reliably reproduce a Turing machine, and therefore could be argued to not be Turing Complete. Those were supposed to be two related, but separate, thoughts. Of course, I suppose a computer could have an electrical error of some sort, so it's not 100% and you could argue that a computer isn't Turing Complete for similar reasons. Though, that's more of a physical reason than an architectural reason, so is a little different on a philosophical level. Of course, we are obviously capable of producing (down to physical limitations) Turing machines so I don't even know. Anyway, I'm not so certain that this train of thought is going to help anyone find a job lol.
|
|
#
?
Jun 8, 2014 20:26
|
|
|
Sil posted:Thinking purely about what can be computed and ignoring the energy and time cost of computing the thing is a bit silly.
|
|
#
?
Jun 8, 2014 20:30
|
|
|
For anyone in the position, how would you look upon my two projects here: https://github.com/Noppadet/Example-Code-Dynamics-SQL-More ? The Dynamics code is x++ which is a C like implementation. TSQL speaks for itself. I think the Dynamics code is a little fat, but how's the readability? The TSQL was very complicated to write, but I think it shows that I can at least do it? Thanks.
|
|
#
?
Jun 8, 2014 20:31
|
|
|
pointers posted:Hey, yes, exactly this, thanks. Like, I thought the question was 'will computers take programmer's jobs anytime soon?'. But then you started talking about NP complete problems.
|
|
#
?
Jun 8, 2014 20:41
|
|
|
Knyteguy posted:For anyone in the position, how would you look upon my two projects here: A single function with over 300 loc kinda stretches it for readability. Break it up into methods that are named the content of your comments and get rid of the comments.
|
|
#
?
Jun 8, 2014 21:30
|
|
|
Knyteguy posted:For anyone in the position, how would you look upon my two projects here: Building up error/warning strings via endless concatenation is a smell to me. Instead, maintain a data structure that contains all of the things that failed validation and the reason. Then pass that data structure off to something that's responsible for formatting the contents. Imagine how much of a pain in the rear end adding a new business rule would be in the current code. And then at the end, bam! There's presentation mixed in with business logic. That's a red flag. Make one method per rule you're checking. Keep passing your data structure that maintains validation errors/warnings through these methods, where each method adds what it finds. Then once you're done running your validation checks, pass the data structure back to your UI and let the UI figure out how to interpret and display it. The SQL is impenetrable. How do you validate that that monstrosity returns the correct results? How do you maintain it and modify it down the line? I haven't given it an in-depth analysis, but I'd expect you could probably make that one stored procedure into about a dozen smaller queries that are more reasonable in size and scope. Let SQL be responsible for data retrieval and only data retrieval. Let your code be responsible for figuring out what data it needs and asking for it. New Yorp New Yorp fucked around with this message at 22:06 on Jun 8, 2014 |
|
#
?
Jun 8, 2014 21:56
|
|
|
Holy poo poo that's the second page-long debate about something ridiculous I've started here this week. I need to stop making flippant/joke replies in this thread apparently. Edit: VVVV Oh god, quick, let me add something on topic: Going to quit soon. Slightly terrified of trying to find an Android role with only eight months of experience, and I feel like I haven't really learned anything like as much as I should have done, but really had enough of the company that I'm at. Tunga fucked around with this message at 22:43 on Jun 8, 2014 |
|
#
?
Jun 8, 2014 22:38
|
|
|
Tunga posted:Holy poo poo that's the second page-long debate about something ridiculous I've started here this week. I need to stop making flippant/joke replies in this thread apparently. I have a lot of angry and ridiculous opinions about this post:
|
|
#
?
Jun 8, 2014 22:40
|
|
|
Tunga posted:Holy poo poo that's the second page-long debate about something ridiculous I've started here this week. I need to stop making flippant/joke replies in this thread apparently.
|
|
#
?
Jun 8, 2014 23:01
|
|
|
Ithaqua posted:Building up error/warning strings via endless concatenation is a smell to me. Instead, maintain a data structure that contains all of the things that failed validation and the reason. Then pass that data structure off to something that's responsible for formatting the contents. Imagine how much of a pain in the rear end adding a new business rule would be in the current code. And then at the end, bam! There's presentation mixed in with business logic. That's a red flag. Thanks Ithaqua / Fuckbooks. We're about to do another workstation with Dynamics so I'll try to create something more robust this time (the CheckRules() has been delivered so I won't get a chance to refactor). The stored procedure was for SSRS. I now have a good idea of how I could've done this in a more pragmatic manner. I have another quick question that may help others: edit nevermind there's no good answer to quitting a small company without feeling a little bad, I don't think. Knyteguy fucked around with this message at 00:01 on Jun 9, 2014 |
|
#
?
Jun 8, 2014 23:04
|
|
|
Tunga posted:Edit: VVVV Oh god, quick, let me add something on topic: Here's a challenge that can help you: make a toy app that uses fragments, async tasks, background services, and list views with custom adapters and cells. Have several activities in there so you can demonstrate knowledge of intents to launch other activities and how to get the result back. Do the same with external stuff, like maybe make the app offer to take a picture with the camera/gallery and offer the option to share data out of the app. Make sure the app doesn't drop frames because it's doing work on the UI thread, and make sure you're not doing UI updates on background threads. Be sure the background service properly stops when it has no more work to do. Make sure the app properly handles rotation, and your UI or background tasks don't poo poo themselves when you rotate. If you can do this and do it well than I'd say you're pretty well situated as a junior Android developer somewhere. Based on my experience trying to grow our team, and from what I hear from others, finding Android developers is hard these days. Even some big players like Yahoo and Twitter seem to be acqui-hiring other Android app teams for their own purposes because recruiting is so tough. If you can handle the basics like I mentioned above I bet you probably would have good luck finding an employer who'd take you on and get your experience ramped up. kitten smoothie fucked around with this message at 23:11 on Jun 8, 2014 |
|
#
?
Jun 8, 2014 23:07
|
|
|
Hi again thread. How's it going? Congrats to the recent group of folks able to get jobs. Keep on learning, including learning from the massive mistakes I've done in my brief career so far. So when I asked the thread last time why I'm having such a hard time getting a job, the main response was to look outside south-east North Carolina. Well I've been doing that, and so far I've had a few interviews with Dropbox, Evernote, Apple, and Gawker Media. Dropbox was incredibly snobby to me, and then they put a war criminal on their board; Evernote's expectations for a interview programming assignment were beyond my capabilities; Apple's been honest to me in saying they went with an internal employee for a position I was considered; Gawker uses Scala and just getting setup for the programming challenge took longer than compiling the linux kernel. Other than that, my interactions with recruiters have been nil besides this disheartening gem:  It's surprising I haven't been contacted by startups or lesser known companies. I know I've been applying to nearly every job I seem qualified to do. Each job application has a tailored resume and cover letter for that position. Yet nothing comes in. Even the automatic application acknowledgement email is a nice thing to get. I know I sound despondent, but six months ago I was suicidal and went through multiple tragic family situations. I couldn't even solve a basic programming challenge from a fellow goon in a reasonable time. I still have a morbid outlook on life, but I keep it to myself and manage to be relatively happy and smile on most days. Today I'm medicated for depression, have gotten professional help, and thankfully living with my parents who are trying to be understanding. However it befuddles them how little contact I get for potential employment, even for menial local jobs like grocery stores and gas stations to at least get me busy doing something. Who I am LinkedIn, GitHub (autogenerated resume), Stack Overflow Careers, PDF Resume, Sample Cover Letter What I know Cocoa, Cocoa Touch, Objective-C, Ruby on Rails, Perl, Python, PostgreSQL, Heroku, Git, GitHub, CSS, JavaScript, jQuery, Bootstrap, PHP, and WebDNA What I've done Battlefield 4 Webcon (demo site) - Intended to be the website for all Battlefield-playing goons to join and manage all goon-run servers. - Used Steam OpenID and verified SA profiles via scraping. - Did have protocol to communicate with BF4 servers implemented and some work done for background process. - I stopped working on this because BF4 really sucked. NDA Mac, iOS, and Full-Stack Web Development - My most of my work for my former employer is covered under a highly-restrictive Apple NDA that doesn't expire for a few years. - About the only thing I can say about this was it was a complete from scratch solution for automating iOS device deployment. WebDNA Timeclock - My first WebDNA project. It was intended for production use, but it didn't interface with Quicken for the accountants so it was dropped - Learned Bootstrap, jQuery, JavaScript, and basic API implementation. I could strip out the WebDNA code to demo the UI capabilities, but it's been a low priority since front-end web dev isn't an interest of mine. Battlefield 3 RCON PHP Scripts - Think of this as like BF4 Webcon, but because I didn't have control over the hosting, I had to rewrite Python scripts in PHP. - Frankly surprised the scripts worked as well as they did. I wish I could advertise the fact these scripts kicked over 10,000 non-goons from goon-run servers, but most people see that as a negative statistic. - For a short time, we decided to play another Battlefield game, Bad Company 2. I made a new branch and made the necessary changes so both versions of the scripts could be run simultaneously without affecting each other. AWStats Scripts - Written at my last internship, these Perl scripts gathered logs from multiple servers and ran them through AWStats, the web stats software to use before Google Analytics came along. - I reduced the scripts' average running time from 45 minutes to under 5 minutes by storing the log's last timestamp between runs and checking for that timestamp during the next run. If found, the scripts only passed a single log file to AWStats rather than all the logs for a single website. AWStats already handled duplication and redundancy checks within logs. - As websites were migrated to different servers, the scripts accounted for that and maintained stats for the site regardless which server it was actually hosted on. Steps I've taken - Paid ResumetoInterviews and LinkedIn Radar to professionally revise my resume, cover letter, and LinkedIn profile. - My stepfather reviewed my resume, cover letter, and social media profiles. He just retired from his HR career which included being the head of HR. He doesn't know tech but from a HR perspective, it was top notch. - Setup automated job email alerts from LinkedIn, GitHub Jobs, WeHireRemotely.com, Stack Overflow Careers, Simply Hired, Glassdoor, Indeed, and Craigslist. I get about 25 emails a day and open each remotely-interesting job (usually 10-15 a day) in separate tabs. I taylor my resume and cover letter for each job application then apply, and repeat. I've separated alerts by location whenever possible, so I get jobs pertaining to SF, LA, NYC, Chicago, Portland, Seattle, Denver, Austin, Minneapolis, Charlotte, and Raleigh-Durham. - I've done practice interviews with friends, family, and one goon. I've been complemented on how well I've answer the non-technical questions. I do struggle with technical questions, but neither friends nor family know programming so I can't practice with them. - Triple-checked my domain's DNS settings to ensure I'm getting every email sent to it. Gotten to know Google Apps pretty well thanks to it. - Attended various local career fairs and young professionals meetups. I haven't found a non-Microsoft developer meetup around here yet. Admittedly restrictive stipulations - Company must meet my ethics. My main concern is that I want to neither track nor kill people. This leaves out most Google positions, advertising and marketing companies, and defense contractors. I did apply for a privacy engineer position at Google however. - Company knows how to handle sexism and man children. GitHub did an incredibly poor job with Julie Ann Horvath's resignation. Pretty much any company mentioned by Valleywag is off the list. Valleywag is the new FuckedCompany. Doesn't help I own CheckYourPrivilege.org. - No video game jobs. I love video games, but gently caress how the industry treats its employees. Twitch seems okay and I've applied there too.
|
|
#
?
Jun 9, 2014 17:15
|
|
|
I took a quick look at your pdf resume and it seems alright except that:quote:Languages: Git, Objective-C, jQuery, Perl, Python, SQL, Ruby, Ruby on Rails, Cocoa, Cocoa Touch, these things aren't languages
|
|
#
?
Jun 9, 2014 17:28
|
|
|
wolffenstein posted:- I've done practice interviews with friends, family, and one goon. I've been complemented on how well I've answer the non-technical questions. I do struggle with technical questions, but neither friends nor family know programming so I can't practice with them. This is jumping out and screaming at me as a major issue. Can you go into more detail about what trips you up? You're shooting for technical jobs, so the technical questions are do-or-die. Non-technical stuff is more of a pass/fail "is this person capable of doing work on time and passing the 'minimum acceptable human being to be around' bar?" and a potential tiebreaker between technically qualified candidates. Don't be afraid to apply anywhere that looks like they might hire you, just for practice. Right now, if you're unemployed, you've got time to schedule interviews anywhere. Even if a company says "our core business is providing disruptive solutions to enhance lethality over the critical link between warrantless phone surveillance and drone strikes PS no women, negroes, or Irish need apply," you can still use them for interview practice and decline the offer if they get back to you. Plenty of big corporations do a similar thing from the opposite direction: they identify a candidate before they open the hiring process, but still bring in a few outside applicants they'll never hire just for the sake of demonstrating due diligence.
|
|
#
?
Jun 9, 2014 18:01
|
|
|
Your resume's using the word "utilizing." Also there's bad grammar, unidirectional double-quotes, and unartistic use of whitespace.
|
|
#
?
Jun 9, 2014 18:13
|
|
|
shrughes posted:unartistic use of whitespace.
|
|
#
?
Jun 9, 2014 21:10
|
|
|
If your resume doesn't contain a quine written in Whitespace, don't bother applying.
|
|
#
?
Jun 9, 2014 23:56
|
|
|
As a hiring manager when I look at your resume I'm going to wonder why you have a one year lapse of employment. Maybe you went backpacking through Asia, maybe you were hurt in an accident, but all I know is there's a year lapse when (most) everyone in IT is getting hired. That's a hard hurdle to overcome. I would swallow your ethics and start applying to the bigger companies (Amazon, Google, Facebook, Microsoft) to see if you can get back on your feet. In a year or two you can start looking for a job with someone that aligns with you better.
|
|
#
?
Jun 10, 2014 00:11
|
|
|
Hypothetical: let's say a misdemeanor conviction turns up on a background check--something like disorderly conduct or a minor drug charge. On a scale from "instant reject" to "if all else is equal it would be the tiebreaker", how much is it going to affect an applicant's chances? I remember seeing this discussed here before but I can't seem to find it.
|
|
#
?
Jun 10, 2014 04:53
|
|
|
SurgicalOntologist posted:Hypothetical: let's say a misdemeanor conviction turns up on a background check--something like disorderly conduct or a minor drug charge. On a scale from "instant reject" to "if all else is equal it would be the tiebreaker", how much is it going to affect an applicant's chances? I remember seeing this discussed here before but I can't seem to find it. Depends on the company and the misdemeanor. There's a big difference between disorderly conduct, which is a made-up crime, and a "minor drug charge".
|
|
#
?
Jun 10, 2014 04:59
|
|
|
Those were just the first two examples that popped into my head. Someone was trying to convince me that tech was some beacon of light in terms of opportunities for people with criminal records, and I remain skeptical. I believe a direct quote was "Nobody cares about a misdemeanor." There was probably an unspoken "...if you're white" at the end of that sentence but I didn't call him out.
|
|
#
?
Jun 10, 2014 05:07
|
|
|
Yeah, I think that's correct. That's my impression, I only have one anecdote to think of and I think that involved a crime more serious than a misdemeanor (and a job-haver). And come to think of it I knew a former drug dealer that got caught and he seems employed.
|
|
#
?
Jun 10, 2014 05:14
|
|
|
wolffenstein posted:Stuff I am not one to normally say this, but honestly, you need to figure out how to reduce that one year gap on your resume. Did you do consulting? Work on your own projects? Edit: Not to say ValleyWag, Pando or TechCrunch are not sometimes on the money. But their entire existence is as a tabloid, not actual news. Strong Sauce fucked around with this message at 05:54 on Jun 10, 2014 |
|
#
?
Jun 10, 2014 05:46
|
|
|
As a guy looking to enter the software development workforce in a few years AND somebody who wants to do some backpacking and travel, what can I do to minimize this potential gap in CV problem? And also I quite understand why a gap on the CV is so bad if its for traveling/backpacking/volunteering overseas? Isn't that a character building good use of time for somebody in early/mid twenties?
|
|
#
?
Jun 10, 2014 06:55
|
|
|
You don't need character to write software, just sheer bloody mindedness. Declared periods of non-employment are fine, but straight out blanks are bad.
|
|
#
?
Jun 10, 2014 08:07
|
|
|
Lampsacus posted:As a guy looking to enter the software development workforce in a few years AND somebody who wants to do some backpacking and travel, what can I do to minimize this potential gap in CV problem? And also I quite understand why a gap on the CV is so bad if its for traveling/backpacking/volunteering overseas? Isn't that a character building good use of time for somebody in early/mid twenties?
|
|
#
?
Jun 10, 2014 08:09
|
|
|
Here's a couple who travel around the world while writing software: http://northwaygames.com/category/travel/ Seems like a wild way to live. I wonder if one of them has a ton of family money to support their adventure.
|
|
#
?
Jun 10, 2014 10:21
|
|
|
So just got back from my first Junior Developer interview. I felt like it went well...ish. How regularly do companies make you sit a little half-hour exam thing? It had questions about SQL, .NET, C# and stuff in it, and then I had to do a twenty minute Diagrammatic Reasoning test.
|
|
#
?
Jun 10, 2014 14:29
|
|
|
Surprise T Rex posted:How regularly do companies make you sit a little half-hour exam thing? It had questions about SQL, .NET, C# and stuff in it, and then I had to do a twenty minute Diagrammatic Reasoning test. That is uncommon -- I've never heard of a company doing that. I've heard of recruiters doing it, just to make sure you're not lying and saying you're a C pound developer when you actually have no programming skills whatsoever.
|
|
#
?
Jun 10, 2014 14:33
|
|
|
I had a year gap on my resume (and was applying to places while unemployed - also didn't do any side projects or anything). It probably hurt me in terms of getting interviews, but every place I got an interview with asked about it and when I told them I went traveling for x months that was either the end of it or it would lead to a 10 min tangential conversation.
|
|
#
?
Jun 10, 2014 14:36
|
|
|
I would say that's pretty common for a junior level interview, all the ones I went to in the past had at least a short "test" section. Kinda weirded out that the interview I just had didn't have any test section. There were technical questions and they've seen my github account but they seemed to take my word that I can code.
|
|
#
?
Jun 10, 2014 14:38
|
|
|
Yeah, I got the job through a recruiter, but the place itself was the one who sat me down with a test and left me alone in a room for the better part of an hour while I did the paper. Scored high on the diagrammatic one though, which reflects well 'cause it was basically algorithms. Input -> Process(es) -> Output, with one of them missing.
|
|
#
?
Jun 10, 2014 14:38
|
|
|
|
| # ? May 22, 2024 19:01 |
|
|
Surprise T Rex posted:So just got back from my first Junior Developer interview. I felt like it went well...ish. There was a company that gave me what amounted to a bespoke IQ test, all for a job that involved a bunch of B2B web app work. Apparently they are one of the best places to work in London? Failed it and went on to somewhere that pays better and is far more interesting. I'd consider it a red flag.
|
|
#
?
Jun 10, 2014 14:39
|
|
