|
rotor posted:i dont really find myself needing to "encode" my "specifications" into code because I dont write libraries any more. back when i did, I didn't find that "byte, int, char*, char*, byte" was significantly clearer that "var, var, var, var, var", you either have them descriptively named or you have docs. jesus christ how dense are you? my entire point is that you can do more with type systems than "byte, int, char*, char*, byte", and you might appreciate this if you learned about types beyond C or, you know, you can just reject a useful tool because you never learned how to use it properly. most programmers do too!
|
# ? Mar 4, 2013 22:13 |
|
|
# ? Jun 2, 2024 15:26 |
|
JewKiller 3000 posted:jesus christ how dense are you? my entire point is that you can do more with type systems than "byte, int, char*, char*, byte", and you might appreciate this if you learned about types beyond C wow, interesting, this is fascinating new information to me. thanks. ive never programmed these "object orientated" programs before so this is all so different.
|
# ? Mar 4, 2013 22:20 |
|
i get it though, you do enough of one thing for long enough and you forget about how other things have different uses and requirements.
|
# ? Mar 4, 2013 22:21 |
|
trex eaterofcadrs posted:vba and ssis both could have been so much better, maybe we'd actually be in a golden age by now if you could use c# instead of vba in excel people would start writing poo poo like email clients and webbrowsers in there. vba has to exist to prevent excel from consuming everything. im pretty sure you could create a library in .net an expose it to com for use in vba if you wanted. assuming ssis is sql server integration services, cant you write c# for that? if its all vba im gonna be disappointed when our 2000->2012 migration is complete cause that's what I was most looking forward to.
|
# ? Mar 4, 2013 22:24 |
|
my argument here isn't from a correctness standpoint, it's from an roi standpoint. type safety for me is where returns stop being real high. and as for the "encoding my specifications", well, i have had the luxury in the past decade or so to not have to worry about that because I'm not writing for a general audience.
|
# ? Mar 4, 2013 22:25 |
|
and of course you assume i'm talking about oop. respond to my loving post not whatever motivations you suppose i have i'm not saying that absolutely every program must be written in a language with a static type system, i never said that. you originally posted that type safety wasn't useful because of the sort of errors it catches. i'm disagreeing with that, it can catch much more important bugs than "true" vs True, if you have a powerful enough type system (not c) and you learn how to use it well enough (not tbc)
|
# ? Mar 4, 2013 22:28 |
|
JewKiller 3000 posted:and of course you assume i'm talking about oop. respond to my loving post not whatever motivations you suppose i have and you should settle the gently caress down, beavis. quote:i'm not saying that absolutely every program must be written in a language with a static type system, i never said that. you originally posted that type safety wasn't useful because of the sort of errors it catches. i'm disagreeing with that, that's cool but that's not what I said quote:it can catch much more important bugs than "true" vs True, if you have a powerful enough type system (not c) and you learn how to use it well enough (not tbc) and I said that generally in the code I've written catching those bugs is not worth the time it takes to deal with strict typing. please respond to my loving post not whatever bullshit you think i posted.
|
# ? Mar 4, 2013 22:32 |
|
Shaggar posted:if you could use c# instead of vba in excel people would start writing poo poo like email clients and webbrowsers in there. vba has to exist to prevent excel from consuming everything.
|
# ? Mar 4, 2013 22:35 |
|
I've never seen a downside to static typing and only the very real possibility of mistakes that comes with dynamic typing so idk why you'd ever want the later. Let the compiler tell you about your bugs instead of waiting to (hopefully) discover them in your tests at runtime.
|
# ? Mar 4, 2013 22:36 |
|
to sum up: im not arguing that typing isn't useful, i'm saying that in the things i've written, it's - generally - not worth the time. maybe you have a different project, or more time, or higher quality requirements, or any of a hundred different other variables. sure, that's cool. I've worked on a few libraries where things would have been a mess if it was untyped. But they've been the exception rather than the rule.
|
# ? Mar 4, 2013 22:38 |
|
Shaggar posted:I've never seen a downside to static typing and only the very real possibility of mistakes that comes with dynamic typing so idk why you'd ever want the later. types for everything makes refactoring harder, so you do it less.
|
# ? Mar 4, 2013 22:38 |
|
i have never had a bug where the root cause turned out to by php type coercion, ever
|
# ? Mar 4, 2013 22:52 |
|
i think what we've learned today is that the answer is that sometimes one thing is good and sometimes the other is good so can we close this thread now or what FamDav fucked around with this message at 22:56 on Mar 4, 2013 |
# ? Mar 4, 2013 22:52 |
|
can we make it so that only tef and how!! and tbc and shaggar can post
|
# ? Mar 4, 2013 22:56 |
|
Suspicious Dish posted:can we make it so that only tef and how!! and tbc and shaggar can post why would you ever do that teffu
|
# ? Mar 4, 2013 22:56 |
|
the three stooges vs the one who knows what hes talking about now previewing in theaters near you
|
# ? Mar 4, 2013 22:56 |
|
I enjoy posting in here and will not give it up no matter what people choose
|
# ? Mar 4, 2013 22:59 |
|
rotor posted:types for everything makes refactoring harder, so you do it less. we're through the rabbit hole here folks
|
# ? Mar 4, 2013 23:02 |
|
rotor posted:types for everything makes refactoring harder, so you do it less. I would say it makes it easier cause you can guarantee you don't miss anything when u refactor.
|
# ? Mar 4, 2013 23:14 |
|
Shaggar posted:if you could use c# instead of vba in excel people would start writing poo poo like email clients and webbrowsers in there. vba has to exist to prevent excel from consuming everything.
|
# ? Mar 4, 2013 23:15 |
|
I think strict static typing is great because when you run into a place where you have to resort to punning, that's a good time to think hard about what you're doing. When I've done lovely poo poo in languages with dynamic typing, I've always wanted a way to have my "magic thinking machine" warn me if I'm not explicit about what I intended.
|
# ? Mar 4, 2013 23:21 |
|
Shaggar posted:I would say it makes it easier cause you can guarantee you don't miss anything when u refactor. i would say it makes it harder because if it's just a pass-through, which so much stuff is, theres a lot more superficial change
|
# ? Mar 4, 2013 23:34 |
|
rotor posted:types for everything makes refactoring harder, so you do it less. it's a lot easier to deal with breaking changes when your whole IDE turns bright red with underlines when you gently caress up, compared to potentially dealing with case-by-case runtime issues.
|
# ? Mar 4, 2013 23:36 |
|
Optional typing systems
|
# ? Mar 4, 2013 23:46 |
|
MononcQc posted:Optional typing systems actionscript
|
# ? Mar 4, 2013 23:46 |
|
all interfaces need to specify types either way, with dynamic typing you will just have to put the specifications into the documentation instead
|
# ? Mar 4, 2013 23:48 |
|
Cybernetic Vermin posted:all interfaces need to specify types either way, with dynamic typing you will just have to put the specifications into the documentation instead says who?
|
# ? Mar 4, 2013 23:49 |
|
Cybernetic Vermin posted:all interfaces need to specify types either way, with dynamic typing you will just have to put the specifications into the documentation instead for adapters or facades or whatever, no, they dont. they're just pass-throughs that dont care what the actual object is, they just relay it to the next thing.
|
# ? Mar 4, 2013 23:51 |
|
well, depends a bit on your views sure, but at least if you are expecting anyone else to call into the interface you will probably want to tell them something about how it is done beyond the name of the function
|
# ? Mar 4, 2013 23:51 |
|
Shaggar posted:if you could use c# instead of vba in excel people would start writing poo poo like email clients and webbrowsers in there. vba has to exist to prevent excel from consuming everything. im pretty sure you could create a library in .net an expose it to com for use in vba if you wanted. i daydream about jamming an R-like into a spreadsheet program just to see if its as awesome as i imagine ssis does use vba until 2008, then you can use C# as well, but the whole thing is like an almost-ran integration system that only exists today because of intertia i think it sucks worse because of what it could have been
|
# ? Mar 4, 2013 23:51 |
|
Cybernetic Vermin posted:well, depends a bit on your views sure, but at least if you are expecting anyone else to call into the interface you will probably want to tell them something about how it is done beyond the name of the function That presumes the use of an float when I meant an int and it gets truncated instead of rounded which is what I would have wanted was intentional and how is there this much discussion about types. Surely there has to be better p fodder. Lazy evaluation or something.
|
# ? Mar 4, 2013 23:53 |
|
rotor posted:typescript
|
# ? Mar 4, 2013 23:54 |
|
rotor posted:dart
|
# ? Mar 4, 2013 23:54 |
|
Bream posted:That presumes the use of an float when I meant an int and it gets truncated instead of rounded which is what I would have wanted was intentional and how is there this much discussion about types. Surely there has to be better p fodder. Lazy evaluation or something. in that case you probably wanted better numeric types than int and float, maybe an integer or a decimal fetishizing an implementation detail such as uint32 or ulong128 is silly, math textbooks don't talk about those and don't have the concept of overflow either
|
# ? Mar 4, 2013 23:57 |
|
More languages with optional typing systems:
Those are the ones I can remember off the top of my head. They vary from type hints to full type systems that can be disabled or enabled on demand.
|
# ? Mar 5, 2013 00:13 |
|
also common lisp
|
# ? Mar 5, 2013 00:14 |
|
if it can be done with the code you wrote, perl will figure out and just do that so you can get on with your day. otherwise it'll die and will give u some hints on how to fix it, if possible no type problems in perl, give it a shot sometime.
|
# ? Mar 5, 2013 00:20 |
|
MononcQc posted:More languages with optional typing systems: Technically dynamic languages have exactly one type ( haskell's data.dynamic allows you to do some neat things) Also one if the dumber things about ghc is that with two commonly used type system extensions you can thoroughly break it by writing unsafeCoerce
|
# ? Mar 5, 2013 00:25 |
|
trex eaterofcadrs posted:i daydream about jamming an R-like into a spreadsheet program just to see if its as awesome as i imagine we're using dts on 2000 and barely suits our needs now. ssis has the potential to fix a lot of problems we have w/ the limitations of dts and I have no problem spending time writing connectors or whatever to provide more functionality. right now we have this piece of poo poo esb (mule) that I want to throw away and ssis will let me throw away like 90% of what we use mule for.
|
# ? Mar 5, 2013 00:39 |
|
|
# ? Jun 2, 2024 15:26 |
|
Malcolm XML posted:Technically dynamic languages have exactly one type ( haskell's data.dynamic allows you to do some neat things) Some of these type checkers are fully equivalent to a static one (Racket and Modula-3 for example) where you can block compilation on errors. Whether checks are duplicated at run time or not is an entirely different issue.
|
# ? Mar 5, 2013 01:12 |