|
Zombywuf posted:http://xkcd.com/810/ Where do the java ones fall then? coffeetable posted:speaking of, anyone seen any sample materials for the Principles and Practice of Software Engineering exam? kinda interested to know what it's like
|
#
?
Aug 16, 2013 20:55
|
|
|
|
| # ? Jun 11, 2024 05:59 |
|
|
Zombywuf posted:http://xkcd.com/810/ uh ones i know of off the top of my head are the sun certified java developer and microsoft MCPD i'm sure there are lots of others, especially for vertical market and single-vendor poo poo. there are probably as many dev certs as CJ certs
|
|
#
?
Aug 16, 2013 21:02
|
|
|
Notorious b.s.d. posted:uh ones i know of off the top of my head are the sun certified java developer This might actually count, don't know enough about it. Even then it wouldn't surprise me to find out that you can just memorise Sun's marketing materials and not actually know how to string two lines of code together. quote:and microsoft MCPD "can use visual studio" They're actually categorised by Visual Studio version.
|
|
#
?
Aug 16, 2013 21:07
|
|
|
Zombywuf posted:"can use visual studio" that's because visual studio versions have a year in the name the actual exams grill you on .NET features and APIs and poo poo
|
|
#
?
Aug 16, 2013 21:10
|
|
|
I'm guessing it would be nice for end-users if there were some basic requirements before being a professional devs, but at the same time, I know I wouldn't have the right to be a developer (or could have ever entered the industry) if it were like that.
|
|
#
?
Aug 16, 2013 21:17
|
|
|
MononcQc posted:I'm guessing it would be nice for end-users if there were some basic requirements before being a professional devs, but at the same time, I know I wouldn't have the right to be a developer (or could have ever entered the industry) if it were like that. Why? You seem to be making assumptions about what would be required. Many (most?) professions that require certification do apprenticeships. It's only in software where you have the kind of abusive certifications that cost you hundreds just to buy the training materials so that your future employers can get a discount on their software costs.
|
|
#
?
Aug 16, 2013 21:38
|
|
|
Zombywuf posted:Why? You seem to be making assumptions about what would be required. Many (most?) professions that require certification do apprenticeships. It's only in software where you have the kind of abusive certifications that cost you hundreds just to buy the training materials so that your future employers can get a discount on their software costs. Because I basically have no computer science nor engineering training at all at a formal place (I ended up getting a minor part time), and all professions around here have education as a prerequisite for apprenticeship in general.
|
|
#
?
Aug 16, 2013 21:41
|
|
|
MononcQc posted:Because I basically have no computer science nor engineering training at all at a formal place (I ended up getting a minor part time), and all professions around here have education as a prerequisite for apprenticeship in general. Well yeah, I am talking about a magical dream land where everything has been bootstrapped already. There is no reason there could not exist a body that could certify a basic level of programming skill via an inexpensive procedure that could replace the cargo cult requirement most companies have for entry level jobs that need 3 to 5 years experience and a high class degree. Other than that it doesn't currently exist and changing things is hard.
|
|
#
?
Aug 16, 2013 23:14
|
|
|
MononcQc posted:Because I basically have no computer science nor engineering training at all at a formal place (I ended up getting a minor part time), and all professions around here have education as a prerequisite for apprenticeship in general. if loud intern is any indication formal computer science training isn't worth much. they've really never taught ~software engineering~ which could and should be an entirely separate discipline like normal engineering.
|
|
#
?
Aug 16, 2013 23:36
|
|
|
The problem ain't certification it's rate-limiting the aggregate number of people who are allowed to pass. In general, professionals whose job it is to assist members of the general public by having deep narrow knowledge should have a licensure procedure because by design their clients are not equipped to know if they are doing a good job or not. But that test should be administered by groups whose only interest is reducing the number of say malpractice suits, not artificially inflating wages. The AMA does that and that's the problem, not the existence of medical licensure you crazy motherfucker. The other category that needs licensure is the one where fuckups can harm or kill people, like civil or nuclear engineering or certain types of comp e or whatever. anal volcano doesn't need licensure to foist the latest nosql bullshit he got off hn on his unfortunate clients. wow look @ me carepostin in the pos, if only the timphone made it easier to type in lower case
|
|
#
?
Aug 17, 2013 00:40
|
|
|
it's medical school that rate-limits doctors entering the profession, not anything the ama is (currently) doing. i don't see the bar association doing much to reduce the oversupply of inexperienced lawyers either certification for programmers is dumb. letting a computer decide anything safety critical is even dumber. once a computer is involved the mechanism is too complicated for people to verify that it wont kill people. yes i think ECUs are a terrible idea
|
|
#
?
Aug 17, 2013 02:59
|
|
|
Nomnom Cookie posted:it's medical school that rate-limits doctors entering the profession, not anything the ama is (currently) doing. i don't see the bar association doing much to reduce the oversupply of inexperienced lawyers either the ama has a vested interest in not certifying new med schools and they set the standards very high for good reason to be honest, u don't want shitlord, MD to treat u after having struggled to get into Joe's Caribbean Medical School and Seafood Shack but they make it v hard for FMGs from decent countries to get in, for reasons that aren't as clear cut aba has been sued a few times too Malcolm XML fucked around with this message at 03:34 on Aug 17, 2013 |
|
#
?
Aug 17, 2013 03:06
|
|
|
programmer certification exam only needs one question How do you feel about Tiny Bug Child's posting? throw out any answers less than five pages, certify the rest
|
|
#
?
Aug 17, 2013 03:15
|
|
|
five pages full of poo poo and i'm certified
|
|
#
?
Aug 17, 2013 22:19
|
|
|
tef posted:before it was wealth, now it is moreso gender and race. where should we move the hurdle to? way to completely miss my point. i don't give a poo poo about people code monkeying whatever the hell they want. what will be annoying is expanding the pool of people who list "embedded developer" on their resume from its current position of "every loving idiot who bought an arduino and ran the software that came with it" to now also including the morons who did it in javascript congrats, you still arent an embedded developer, you missed the entire point of being an embedded developer, now finding real embedded developers is even harder! if your notion of embedded developer is abstracted far enough away from the hardware that toy languages like js may be a viable option then spoiler alert it isnt embedded development its just development
|
|
#
?
Aug 18, 2013 01:43
|
|
|
Condiv posted:five pages full of poo poo and i'm certified more work than a typical cert
|
|
#
?
Aug 18, 2013 01:44
|
|
|
Bloody posted:way to completely miss my point. i don't give a poo poo about people code monkeying whatever the hell they want. what will be annoying is expanding the pool of people who list "embedded developer" on their resume from its current position of "every loving idiot who bought an arduino and ran the software that came with it" to now also including the morons who did it in javascript Every idiot who runs the Arduino sample code is also looking to add C++ to their resume. I think a lot of people that get CS/CE degrees now know at least how to use an Atmel chip without the higher level languages. At what point in time would you believe you're not doing "embedded development?" It's required to take three semesters of Atmel poo poo at my school, so maybe I'm just sheltered in my belief.
|
|
#
?
Aug 18, 2013 05:14
|
|
|
UberVexer posted:Every idiot who runs the Arduino sample code is also looking to add C++ to their resume. I think a lot of people that get CS/CE degrees now know at least how to use an Atmel chip without the higher level languages. At what point in time would you believe you're not doing "embedded development?" when you're operating at a level higher than bare metal code/arent particularly invested in the hardware design of the system. it's a gray area for sure.
|
|
#
?
Aug 18, 2013 05:17
|
|
|
if u dont know the difference between hosted and freestanding C then u cant call urself an embedded dev imo
|
|
#
?
Aug 18, 2013 05:36
|
|
|
UberVexer posted:I think a lot of people that get CS/CE degrees now know at least how to use an Atmel chip without the higher level languages. what
|
|
#
?
Aug 18, 2013 06:24
|
|
|
yeah i've never even heard of this but i'm literally a highschool dropout so
|
|
#
?
Aug 18, 2013 09:47
|
|
|
"Writing an embedded program that blinks an LED is the easiest thing in the world" once you set up your linker script to emit a vector table in ROM at a particular location, write a crt to set up ur stack and clear bss, open the clock gate to your GPIO unit, and set up the drive mode for the pin in question, sure Embedded development is mostly poo poo like that, there's little in the way of algorithms or data structures most of the time.
|
|
#
?
Aug 18, 2013 18:23
|
|
|
lollin ps the one good post went "what if companies were held responsible for their code"
|
|
#
?
Aug 18, 2013 21:28
|
|
|
no one can afford to write software without bugs
|
|
#
?
Aug 18, 2013 21:35
|
|
|
Mr Dog posted:once you set up your linker script to emit a vector table in ROM at a particular location, write a crt to set up ur stack and clear bss, open the clock gate to your GPIO unit, and set up the drive mode for the pin in question, sure Is this some kind of deliberate irony?
|
|
#
?
Aug 18, 2013 21:41
|
|
|
every industry ever regulated posted:no one can afford to add safety features to their product
|
|
#
?
Aug 18, 2013 21:44
|
|
|
tef posted:no one can afford to write software without bugs nah, just bust out your hoare axioms and proof tableaux and get to work, if you do it fast enough you might get it done before your code gets obsoleted
|
|
#
?
Aug 18, 2013 21:55
|
|
|
are embedded compilers still poo poo? are there any sort of ides for anything?
|
|
#
?
Aug 18, 2013 22:03
|
|
|
Police Academy III posted:nah, just bust out your hoare axioms and proof tableaux and get to work, if you do it fast enough you might get it done before your code gets obsoleted software verification has come a long way since proof tableaux. the industries which are willing to splurge on safety (read: aerospace) have fallen madly in love with formal verification went to a great talk a few months back on building tools to automate the ESA's model checking
|
|
#
?
Aug 18, 2013 22:04
|
|
|
Wheany posted:are embedded compilers still poo poo? MPLab X is sort of ok but the compilers are still poo poo yes
|
|
#
?
Aug 18, 2013 22:09
|
|
|
coffeetable posted:software verification has come a long way since proof tableaux. the industries which are willing to splurge on safety (read: aerospace) have fallen madly in love with formal verification correct me if I'm wrong, but to be able to do useful formal verification, don't you have to use a language with different semantics than the typical imperative/mutable c/java/etc style? i heard that haskell is good for that sort of thing since it's so uptight about everything, and because recursion is easier to verify than loops.
|
|
#
?
Aug 18, 2013 22:14
|
|
|
imperative language semantics are pretty extensively theorized and the way to verify loops is to verify them as recursive functions (which they are)
|
|
#
?
Aug 18, 2013 23:04
|
|
|
Police Academy III posted:correct me if I'm wrong, but to be able to do useful formal verification, don't you have to use a language with different semantics than the typical imperative/mutable c/java/etc style? i heard that haskell is good for that sort of thing since it's so uptight about everything, and because recursion is easier to verify than loops. the way to think of haskell's type system is that it's doing a bunch of verification and inference for you You can encode proofs of quite a few things like the B-Tree invariants in the type system and be pretty confident you didn't gently caress up http://matthew.brecknell.net/post/btree-gadt/ Of course there's always Knuth's "I've only proved it correct, not run it"
|
|
#
?
Aug 18, 2013 23:19
|
|
|
software formalism is a vast wasteland btw
|
|
#
?
Aug 19, 2013 00:19
|
|
|
MononcQc posted:Using callback-based code, you do not control what (event, state) combination you will get, and while some are not gonna happen, many of them are possible through annoying combinations. You need to prepare code to handle that explicitly in the least pleasant way possible. There are ways to avoid these strawman N^2 callback/state combinations quite rigorously. Use the right design pattern. e: well maybe this is only really really true in advanced statically typed languages such as C++. shrughes fucked around with this message at 01:11 on Aug 19, 2013 |
|
#
?
Aug 19, 2013 01:06
|
|
|
Wheany posted:are embedded compilers still poo poo? If you're shipping less than like a million units of something and you have any say in the matter than use a Cortex microcontroller. Then use GCC to compile your 32 bit ARM code, I hear that compiler is pretty good. The cost is basically the same as MSP430 and PIC at this point, it's slightly more expensive, but the price difference is not worth the extra ballache unless you're shipping a shitton of units. Your IDE is vim you loving scrublord. Binary we're-only-technically-not-violating-the-GPL gdbservers for JTAG devices can be a pain to work with, though.
|
|
#
?
Aug 19, 2013 01:24
|
|
|
Police Academy III posted:correct me if I'm wrong, but to be able to do useful formal verification, don't you have to use a language with different semantics than the typical imperative/mutable c/java/etc style? verification is a long, long way from my speciality, but to me this sounds like another great reason to get rid of c/java/etc  (or at least enforce programming styles upon them that make verification easy) Mr Dog posted:Your IDE is vim you loving scrublord. Binary we're-only-technically-not-violating-the-GPL gdbservers for JTAG devices can be a pain to work with, though. aight so i know nothing whatsoever about embedded development, but my pa (whose current project is mesh networking for streetlamps) has been raving about the eclipse target communication framework. is it as cool as he's making out?
|
|
#
?
Aug 19, 2013 01:38
|
|
|
there are pletny of good embedded IDEs, but as has been said before they are mostly useful if you want to do step through debugging with your JTAG.
|
|
#
?
Aug 19, 2013 01:48
|
|
|
Mr Dog posted:If you're shipping less than like a million units of something and you have any say in the matter than use a Cortex microcontroller. Then use GCC to compile your 32 bit ARM code, I hear that compiler is pretty good. The cost is basically the same as MSP430 and PIC at this point, it's slightly more expensive, but the price difference is not worth the extra ballache unless you're shipping a shitton of units. yeah if you want to cj linker scripts you can use vim i guess, you also might want to program in asm i guess!?!? the rest of us are using ides based on any one of a couple different ides based on the vendor's whims, usually with v. nice jtag integration and investigation of system resources i mean come on, we just went over how a majority (though by no means all) of embedded systems work is peripheral setup janitoring, whyyyyyyyy would you not use every tool possible to remove the bullshit and let you focus on the actual important bits?!? things like ti's GRACE tool are a giant step forward in that regard, not something to let the 'weak' get into our jobs
|
|
#
?
Aug 19, 2013 01:53
|
|
|
|
| # ? Jun 11, 2024 05:59 |
|
|
Zaxxon posted:there are pletny of good embedded IDEs, but as has been said before they are mostly useful if you want to do step through debugging with your JTAG. or if you want something to pop up tab completion for whatever meaningless all-caps macros the vendor has decided to name everything
|
|
#
?
Aug 19, 2013 01:54
|
|