The Whiteboard

Robert Harvey

12:10 AM

This answer makes no sense whatsoever. — Robert Harvey 10 secs ago

Ixrec

I like "programmer is nice guy, who wants to write nice code"

Robert Harvey

12:29 AM

Yeah, that's a lie, isn't it?

Ixrec

it means whatever you want it to mean

user114359

1:03 AM

Next program I write I will write meanly. I will name variables stuff like "imNotSayingWhatThisVariableDoes" and "phlblblblblblbb"

Telastyn

"Nice" is the adjective you use when you can't think of anything better.

"You know Tel, you're a nice guy, but... (I wouldn't be caught dead in public with you)"

user114359

"Yes, I know her, she is such a nice girl!" (talking about the physically unattractive woman)

user114359

Ah, so much of our language is communicated outside of the literal meaning of words

user114359

Humans suck.

Jimmy Hoffa

public string ScrewYouBuddy(int getBent, YourDumb shutup)
{
    var goDie = getBent / shutup.AndDie;
    return new String(Enumerable.Range(0, goDie).Select(jerk => (char)jerk).ToArray());
}

user114359

1:13 AM

@JimmyHoffa ...but can you write that in Haskell?

Jimmy Hoffa

> screwYouBuddy getBent shutup = pack [0..(getBent / andDie shutup)]

user114359

@JimmyHoffa Sounds like Yoda coding: lawn my off get!

psr

1:33 AM

@JimmyHoffa Thanks for formalizing this. Frothing insults are usually so lacking in precision.

@JimmyHoffa And why shouldn't insults be declarative and immutable?

user55340

2:55 AM

Guns,Germs and Steel - Jared Diamond

►

user55340

3:11 AM

@amon and he deleted the product question. Now to think of a way to rephrase it and properly constraint it so that its a code golf question.

felipe

Hello everyone, I have a question regarding Git and SVN.
Imagine that in my SVN repository, I need an older version of one file to be committed on top of the new one. In order to do that, I need to update that file to a given revision, copy it somewhere else, update the working copy again, substitute the file and commit. Is there an easier way to do that with Git?

psr

@JimmyHoffa - Is there anything I can do to help with getting that web API code sample? Or is it turning out to be a pain? (I'm not sure what files would be involved so I really don't know).

Jimmy Hoffa

3:38 AM

@psr Nah sorry - just haven't gotten to it. I'll do it now, got an e-mail? I'll send you a zip. You can toss it on github or something to share; I don't have git setup to do such

Jimmy Hoffa

3:54 AM

@psr github.com/JimmyHoffa/OWIN-REST-Example

there, I pushed it up. It's a hand full of half-done things just to serve as a concept for architectural boundaries for a service as well as showing a console host for OWIN and an IIS host for OWIN

Jimmy Hoffa

4:16 AM

@psr let me know if that works for ya or you have any other questions

Jimmy Hoffa

4:44 AM

@psr do note - about the controller - and this is dumb as hell, inheriting from ApiController isn't enough, your controller classes actually have to end with "Controller" near as I could tell. I didn't dig too deep when dealing with that, just recall if I didn't name the class SomethingOrOtherController it wouldn't actually show up. Also, when the owin routing is set, you'll have had to do something to load the assembly with your controllers into your AppDomain.

A reference to a type in that assembly as you define your IoC is likely enough, that's more or less what I did.

oh and you have to turn nuget package restore on by right clicking the solution and selecting it - so you pull down the MVC stuff referenced from nuget

Duga

6:58 AM

p.s. Stackoverflow is for specific focused questions if you want to ask about design or algorithms use the sister site programmers.stackexchange.com or another one where open questions and discussions are welcome. — simbo1905 17 secs ago

Duga

7:18 AM

I'm voting to close this question as off-topic because it belongs on Programmers StackExchange — Yuval Itzchakov 47 secs ago

Duga

8:34 AM

@BrunoFinger: Programmers lists software licensing in their list of topics. Do search first, of course. — Martijn Pieters ♦ 44 secs ago

Duga

10:01 AM

This question would definitely be a good, on-topic question for Programmers. I can already think of a few FSF pages I'd quote to give it a super-thorough answer. At the risk of sounding pedantic, "legal advice" is off-topic everywhere; we allow software licensing precisely because explaining the intent of a license can be done accurately and concisely without turning into full-blown legal advice. — Ixrec 14 secs ago

Thomas Owens

12:16 PM

I just noticed that the "How to be a Programmer: A Short, Comprehensive, and Personal Summary" is released under the GNU Free Documentation License. If I wanted to, I could take it, put it on GitHub, and start editing it, I guess.

Duga

12:39 PM

@Pat Programmers.SE perhaps; they tend to tolerate opinion-based questions more. — Bartek Banachewicz 11 secs ago

user114359

I just learned something about Java and how it parses source code. Check this out:

user114359

335

Q: Why is executing Java code in comments allowed?

The following code produces the output "Hello World!". (No really, try it) public static void main(String... args) { // The comment below is no typo. // \u000d System.out.println("Hello World!"); } The reason for this is that the Java compiler parses the unicode character \u000d as a ne...

user114359

I have been programming in Java since 1998 and never knew that.

Thomas Owens

@Snowman It's probably a good thing that people don't know about it. Because it's a bad idea.

I wonder if any of the static analysis tools (especially the security oriented ones) detect this and report it as a finding.

user114359

12:54 PM

@ThomasOwens It is actually a really good idea for the reasons stated in the answers, but should only be used for those reasons. Of course it is not as big a deal in 2015 now that Unicode is used pretty much everywhere.

Thomas Owens

@Snowman Putting unicode in comments is a good thing. The side effect of turning what appears to be a comment into executable code is...dangerous.

Especially when IDEs don't highlight the "comment" as executable code - it still looks like a comment to me in Eclipse Luna. I'd hope that static analysis tools detect this. Ones that work on bytecode would, for sure. But human analysis could miss this, and that's problematic in highly secure environments.

user114359

@ThomasOwens I could see an argument for a compile error on Unicode escapes that change the semantics of the program from within comments

Thomas Owens

That's my whole concern. Not that unicode can go in comments, but that tools may not be properly implemented to detect this. Eclipse isn't, and neither is the SO code viewer. If I did this, I'd be in trouble at work.

I'm sure it's all kinds of security violations to actually use this.

ratchet freak

like trigraphs in C

little known and not all tools handle them correctly

user114359

trigraphs can do some nasty stuff. I remember a few IOCCC programs taking advantage of them.

user114359

1:03 PM

trigraphs are fairly useless given modern keyboards

ratchet freak

@Snowman trigraphs where created for non ascii 6 bit encodings

user114359

I thought it had something to do with early keyboards restricting what could be typed. I don't know, I have never used them.

ratchet freak

remember that C is from the early days of computing before ascii was the standard

Robert Harvey

1:28 PM

Seems like a lint rule that forbids unicode escapes in comments would solve this problem.

Thomas Owens

@RobertHarvey It would indeed.

Robert Harvey

Or at least forbids unicode newline characters.

user114359

I think it would be safe to exclude anything that is nonprintable

Thomas Owens

But if people don't know that this is an issue and if tools don't have these configured from go time, that's when you run into issues.

Robert Harvey

Someone might need to have a discussion in a comment about the way a particular unicode character is handled. Though you could have that discussion without the escape character.

Thomas Owens

1:30 PM

Unicode characters also tend to have names.

So you don't need the code.

user114359

Some of the common ones do, but most do not. But the named characters should cover many use cases while simultaneously helping by not making people look them up.

Duga

I think that the question might be too broad or subjective for Stackoverflow. It might be better suited for Programmers StackExchange. I'll see if I can get it moved over there. — JasonWilczak 56 secs ago

@Snowman thanks for the link. I still feel this would be on topic for Programmers.SE, based on their What topics can I ask about here page since this appears to be about "software architecture and design" and seems to follow in line with a Great Subjective Question‌. But again, just the humble opinion/vote of one developer, which, could be completely wrong. I do think, however, it is not a fit for SO since there is no inherit problem mentioned. — JasonWilczak 10 secs ago

@JasonWilczak you are correct it is on-topic for Programmers, but is currently too broad and needs work before a moderator would migrate it. Questions should be fixed before migrating, or the asker tends to get confused and frustrated. — Snowman 35 secs ago

Thomas Owens

2:05 PM

I really hate TDD. MetaFight's workflow is close to mine. I write code. I write some simple test cases, drivers, or whatever is needed. I refactor the code and tests. I revisit the test cases as I'm finding issues either reading or running the code, adding or removing them.

Robert Harvey

And writing tests while you doing that just doubles the workload, right?

Ixrec

my first attempt at TDD basically caused every change to the code to be twice as much work as usual because every code change required a test change too, when the whole point of good coding practice is to make change stay easy

and that just encouraged us to stop running and maintaining the tests entirely

enderland

@Ixrec why did your code change require a test change?

Thomas Owens

@RobertHarvey It's the whole test-first thing that is hard.

Ixrec

these days I only write tests if I have some confidence that I can come up with tests that won't need changing anytime soon, which for most code simply can't be done on the first few revisions

Thomas Owens

2:08 PM

Hard to the point of not adding value.

Ixrec

@enderland that question is making the false assumption that non-brittle tests are always easy to write

there's an article about this that got linked at work the other week which I think sums up my feelings perfectly: david.heinemeierhansson.com/2014/…

to be clear, there are parts of the code where non-brittle tests are very easy to write, and when working on them I do the whole red-green-refactor TDD thing; it's just not all the code

Robert Harvey

I see Test-First as a learning tool. Training wheels, of sorts. Great when you're trying to learn how to ride the bike, but you take them off once you don't need them anymore.

Ixrec

you're almost quoting the article I linked, lol

André

> It's given birth to some truly horrendous monstrosities of architecture. A dense jungle of service objects, command patterns, and worse.

My experience with TDD

user55340

Test first is an ideal. It is good when you know exactly the direction you are going. This is often the case with bugs.

Ixrec

2:13 PM

that part I personally haven't seen, but I have always been suspicious of the "inject this dependency because you'll need to mock it for a test" idea when real code won't ever need to inject a different version of it, which I assume leads to those jungles

André

Whereas just testing it later, with good coverage, made much more sense

Or maybe I just got TDD wrong

user55340

The problem with not test first is that test later often becomes test never.

4

André

That's real world business

Ixrec

the problem with test first is that test first often becomes ignore the broken test

Robert Harvey

Oh, funny. The author of that article also used the Training Wheels metaphor.

Telastyn

2:15 PM

I don't quite follow that workflow. Refactoring without unit tests is annoying.

Robert Harvey

@MichaelT Meh. "Gotta ship" trumps almost everything, especially Test First. Test First will be the first thing that's jettisoned when you're up against a deadline.

user55340

Test first works really well when someone hands you a bug you can reproduce. Write the test, yep - there is the bug. Now fix it.

Ampt

@MichaelT Tests? Tests?! We don't need no stinking test.

Ixrec

most of our bugs involve UI interactions that can't be automated, sadly

when we have a bug in the backend code, I absolutely write a test before fixing it

user55340

However, tests only have value when there is reasonable coverage. Sure you fix the bug, but lacking coverage you break two other things.

Ampt

2:17 PM

@RobertHarvey Congratulations!

Ixrec

I thought some tests were still better than no tests

Robert Harvey

@Ampt Thanks!

Ixrec

requiring perfect coverage upfront sounds like it would force you to go back to waterfall-style development, where no changes are allowed after the initial plan

user55340

Test first endeavors to instill the "good coverage" discipline. And it takes time and discipline to do.

user55340

Not perfect - sufficient.

Ampt

2:19 PM

Then again you need to explain why your development effort takes 2x longer now to a stakeholder who wanted it done yesterday, and by the way, here's 50 pages of new requirements

Telastyn

the problem I have with MetaFlight's workflow is that writing tests a month later assumes you remember what the hell this crap was supposed to do a month later.

user55340

@Telastyn exactly.

Telastyn

writing tests right with the code is faster and easier.

Ampt

Some efforts aren't worth the cost of test-first. we aren't all building rocket ships or pacemakers

Ixrec

I usually don't know what it's actually supposed to do until a month later, when business/UX have seen the prototype and had a chance to tell us what's wrong with it

Telastyn

2:19 PM

and it is cheaper than paying a bunch of QA drones to find the same bugs.

Ampt

Prepares for crucifixion

user55340

agitar.com/downloads/TheWayOfTestivus.pdf

Ixrec

@Telastyn that's what everyone always says, but every time I've tried it, this just is not even remotely true

user55340

>
The best time to test is when the code is fresh

user55340

Meh. Bad copy and paste.

Ixrec

2:21 PM

and you only get the long term benefits if the requirements for that code never change so you only have to fix bugs

Telastyn

I don't test first, but unit tests are a required part of "done" for my teams.

user55340

> The best time to test is when the code is fresh

Your code is like clay. When it’s fresh, it’s soft and malleable. As it ages, it becomes hard and brittle.

If you write tests when the code is fresh and easy to change, testing will be easy, and both the code and the tests will be strong.

If you write the tests when the code is old and hard to change, testing will be hard and both the tests and the code will be brittle.

Ampt

* Subject to constraints, time, and budget. Not applicable in all 50 states. See your local representative for terms and details.

Telastyn

@Ixrec - I'm not sure what sorts of scenarios you're in, since I have always found it to be true.

I find usually 2-3 bugs a week with unit tests

Ampt

Test first is great in ideal situations. In a perfect world, all projects would be accommodated to have 100% coverage and test driven design.

Ixrec

2:22 PM

my guess would be that for us, requirement changes happen more often than refactors

Telastyn

often subtle edge cases that QA would take forever to find.

Ixrec

so any attempt to encode the correct behavior in a test is out of date before that test has a chance to catch anything

though I'm mostly referring to near-frontend code, again the small amount of backend code I work on is perfect for TDD and it always works there

Telastyn

that's like saying any attempt to encode the correct behavior in code is impossible.

user55340

Business logic needs different tests.

enderland

one advantage of test-first is that your test by definition has to be written basically as a requirement rather than a "I think the code does X, so I'll test it as such" - this is a problem with @Ixrec's environment.. :)

Telastyn

2:23 PM

ah, I don't work much on frontend code.

you can't unit test that stuff

user55340

You can... Just that you need different test harness.

Ampt

Yup. We do that all the time

Telastyn

there's no way for a computer to programmatically determine if a UI is usable or pleasant or functional.

user55340

Look and feel... That is what is hard to test.

Ampt

Selenium is pretty good for doing business use case tests

Ixrec

2:25 PM

@Telastyn yep, exactly this

Ampt

You can test for functional;

usable depends on your definition

Telastyn

not well.

Robert Harvey

@Telastyn So you write the tests after the method has been written, before refactoring.

user55340

But I can test to make sure that the Dom is correct after some JavaScript function is invoked.

Ampt

pleasant is in the eye of the beholder.

Telastyn

2:26 PM

@RobertHarvey - generally. The tests are what verify that the code works. Functional code trumps clean refactored code, so I make sure I meet the base requirements first.

Ixrec

I've heard places that do real web development with direct access to the DOM and whatnot use tools like Selenium to deal with these problems, but sadly our infrastructure teams haven't provided any equivalents yet (despite many of us asking)

Telastyn

these days I will make some simple refactorings before tests since I'm often confident that I can do them correctly.

Ixrec

these days we're cobbling together a tool that will programatically search for images on the screen and click on them, which is at least starting to allow us to do some genuine smoke/acceptance/regression testing on UIs

even if "unit tests" for such code remain largely impractical

durron597

@Ixrec Computer vision is hard.

Robert Harvey

And UI tests are brittle.

Ixrec

2:29 PM

@durron597 there's an art to selecting exactly which region of the UI to take the screenshot of to minimize brittleness

still, it's already proven to be infinitely more useful than unit tests in the front end, since the first failure of this tool caught an actual bug instead of yet another requirement change

André

@Ixrec what is the web stack you work on?

Ixrec

a proprietary one that only exists within our company

enderland

there be dragons!

André

run for your lives!

Ixrec

shockingly, it's not actually that bad

André

2:35 PM

so html5/javascript?

Ixrec

yes, during the brief time I worked on the stack itself I was fiddling with css, html and javascript

the frontend part of the stack anyway

André

Intern is pretty cool, have you tried that? developers.mobilesystem7.com/blog/post/automating-ui-tests

Ixrec

it helps that this is the company's new bleeding-edge framework to replace the previous one that also only existed within our company

because of the way our company is organized, any automated UI testing that does not involve image files would require the infrastructure teams to take a tool (like what you linked) and write stuff on top of it

it's on our very long list of things we want them to do

André

well, good luck then :)

Ixrec

hence image files

but even if that tooling wasn't the issue

the fact remains that with UI development, the first iteration is frequently thrown out and redesigned from scratch, so any effort spent writing tests for it before it has a chance to settle is often wasted

André

2:39 PM

I saw somewhere, on github, some tool to generate the screenshots and automate the comparison

enderland

comparing screenshots can be tricky though...

Duga

This is not really a programming question might be a better question for programmers.stackexchange.com ( conceptual questions) or codereview.stackexchange.com ( peer programmer code reviews). Welcome to stack overflow, recommended reading: stackoverflow.com/help/how-to-ask — Dan Beaulieu 39 secs ago

André

saw a presentation about it, was somewhat cool, it did pixel by pixel comparison, somewhat clunky to setup

Ixrec

it can be tricky, but since we have no direct DOM access it's the most reliable option

André

I can look it up for your if you want @Ixrec

Ixrec

2:40 PM

(and even if we did the infrastructure teams are allowed to change the DOM structure of their widgets at any time, which is why we don't have access in the first place)

enderland

@André the trick with pixel by pixel comparisons is that... if your UI renders differently for whatever reason (we had issues with graphics card stuff) or the UI framework does things differently you get all sorts of false failures

Ixrec

yeah, we absolutely do not want pixel by pixel comparisons

André

@enderland, yup

Ixrec

the tool we're using defaults to a "similarity index" of 0.6

enderland

good :D

Ixrec

2:41 PM

and for the important parts of the test, I'm giving it an image that should match, and an image that should NOT match (ie a screenshot of a bug we had)

so that proves it's not simply accepting everything or rejecting everything

@enderland for us it's not even weird bugs, the VX requirements actually do change every couple months in fairly significant ways

e.g. only a month ago they decided grid rows should be 19px high instead of 22px

(and from a UX/VX point of view I strongly agreed with that change)

Ampt

Bought a boat last night

5

My only thought so far today

enderland

toyz

Ampt

Ended up going with an '88 Searay in pretty good shape... now I gotta figure out title and crap like that

luckily I've got the title for the boat and trailer.

durron597

@Ampt What'd you pay, like $7k?

Ampt

2:56 PM

@durron597 Nope. <5

Jimmy Hoffa

@MichaelT and the problem with test first is that a bunch of dynamic language snobs don't realize you can't write tests for types that don't exist in static typed languages

durron597

@Ampt ni han

Ampt

Thanks.

took some negotiation to get him down, but it was worth it.

Jimmy Hoffa

I'm On A Boat (ft. T-Pain) - Album Version

►

Ampt

I must have found every scratch, ding, and tear on that thing

had a price for every one of them too haha

@JimmyHoffa You at kinkos straight flippin copies.

durron597

3:00 PM

Star Trek set to "Move Your Feet"

►

Stack Exchange

Blog - Stack Exchange: Welcome, Nicolas Chabanovsky, and Stack Overflow in Russian!

Tim Post on June 11, 2015

Some time ago, we hired a Russian-speaking Community Manager named Nicolas Chabanovsky. He works remotely from St. Petersburg (we’re pretty sure it’s the one in Russia). In the past Nicolas was a software developer, and participated in the development of DLNA-stack at Motorola, webOS at LG Electronics and many other notable projects. What we found the most interesting about Nick is he’s one of the founders of the most advanced clones of Stack Exchange we’ve ever seen, which was ХэшКод, otherwise known as HashCode. …

user55340

@Ampt a boat you say?

user55340

Lyle Lovett - If I Had A Boat (live)

►

user55340

>
The mystery masked man was smart
He got himself a Tonto
'Cause Tonto did the dirty work for free
But Tonto he was smarter
And one day said kemo sabe
Kiss my ass I bought a boat
I'm going out to sea

user55340

3:16 PM

@Ampt the real question though is how you will program while on a boat.

Ampt

@MichaelT with my mobile hotspot, obviously

I would totally be out on the lake right now if my GF wouldn't murder me for not taking her with.

user55340

116

Q: What is the boat programming meme about?

I remember some time ago there was some huge problem regarding some question that had something to do with a boat, and I think I missed that topic completely. As I see it being mentioned every now and then, I'd just like to know what it was about.

discussion memes

enderland

@MichaelT hey his cats name was ender

user55340

@enderland is it a sphynx?

enderland

> You'll need an Ansible connection for sure. Ask your cat, "Ender" for some help with that.

durron597

3:31 PM

@MichaelT wtf, why is the original question gone not even as a 10k link

@Ampt Either take her with or tell her to call you when she's free and then briefly go back to dock to pick her up

user55340

So the 10ks couldn't undelete it.

Ampt

@durron597 unfortunately, she's at work til 5, then it would take her 1.5-2 hours in traffic to reach where I would be on the lake.

durron597

@Ampt You can go boating for 10 hours, nbd

Ampt

hahaha right?

user55340

Do a romancing the stone and pick her up in the boat.

user55340

3:37 PM

user55340

See? No traffic.

Ampt

Something about vacation time and not enough of it

user114359

Please read my comment (and see my most recent edit) before VTCing this question:

user114359

-2

Q: What is the difference between function() and function(void) in C?

I have heard that it is a good practice to write functions that do not receive anything as a parameter like this: int func(void); But I hear that the right way to express that is like this: int func(); What is the difference between these two function declarations?

programming-practices c

user114359

It rightly deserves downvotes, but should probably be left open

user55340

3:43 PM

@Snowman / @Ixrec consider either a significant rework of the question or write a new q&a pair that answers the core question so it can be duplicated here.

Ixrec

what's the core question not being answered?

I think Snowman's edit just turned it into something acceptable

durron597

25

Q: Understanding the difference between f() and f(void) in C and C++ once and for all

Possible Duplicate: is f(void) deprecated in modern C and C++ Ok, so I have heard different opinions on this subject and just want to make sure I understand it correctly. For C++ Declarations void f(); and void f(void); mean precisely the same thing, the function f does not take any ...

c++ c function void

user114359

Yes, I made it more general because it is not specific to main(). I believe this captures the essence of the problem:

user114359

> What is the difference between these two function declarations?

user55340

Acceptable and on topic here rather than asking about implementation details.

user55340

3:46 PM

(Rather than one I really itch to migrate)

Ixrec

@durron597 oddly enough, the accepted answer there seems to have vastly extended the scope to "how does parameter passing work?" so that's not a great duplicate either

durron597

@Snowman link thief

user114359

@durron597 :-P

Ixrec

I spent a few minutes searching for SO questions that actually duplicated this particular question and never managed to find any, just lots of related but different questions

durron597

@Ixrec My question is the same...

The only way it's not the same is the part about "what is the best practice"... and we all know how awesome best practices questions are.

Ixrec

3:49 PM

okay, I couldn't find any that had satisfactory answers

like that question, all the ones that included an answer to this question had it buried under lots and lots of other irrelevant information

as SO loves to do when there's a standards document to be quoted

durron597

Once again this really only serves to enrage me about how crappy cross-stack duplicates are handled on the network

Ixrec

yeah, there's still no agreement on what even is a duplicate

also, why the heck am I getting so many upvotes for that half-assed answer

user114359

@Ixrec because it is accurate and correct.

durron597

whoa, work-environment suddenly down to 34 open questions. @ThomasOwens did you take a sledgehammer to the tag?

user114359

If the author accepts the answer this will be a good place to point other quests as duplicates if they show up. I wonder if we have questions that are dupes but deleted, I am sure I have seen it here on Programmers before.

Ixrec

3:53 PM

it'd certainly be better than closing with a custom "This is on SO" reason, or closing as dupe of a locked/closed/deleted/crappy question like some sites love to do

durron597

If I edit the career advice bits of this question away, could the resulting question be good or would it still be too broad:

1

Q: How to test a random feature in MS Excel/WORD/PPT

Interviewers often ask questions about testing a random feature in Excel, Word, Notepad etc.. I am someone who will be interviewing for a SDET position and has a SDE background not specifically SDET background. I am extremely interested in testing though so applied for a SDET position. So how sho...

interview

Ixrec

first thought: I can't understand what the question is trying to ask

second thought: the answers seem to have interpreted this as either an opinion poll or an open-ended list question

durron597

@Ixrec "How does one test a feature in a deployed application?"

Ixrec

that definitely sounds too broad

what kind of feature, what kind of application, how is it deployed, do you mean manual or automated tests, do we need to lecture OP on why automated tests are superior, etc

durron597

@Ixrec That's what I thought too, I was double checking because I don't want to edit it and still have it be closed. Because then it's closed and on the front page.

Ixrec

4:01 PM

yeah

more importantly, because of the answers that are already there

any edit you did to make it not a list question or not too broad would almost certainly invalidate most of them

user114359

@durron597 don't worry about the front page too much

durron597

STCI Status: interview 174. Star this if you think the remaining 29 work-environment questions are on topic, and VTC the bad ones if you don't.

2

user114359

Poor Grizzly is probably wondering what we did to his question

Thomas Owens

@durron597 Most were on-topic, but primarily opinion or too broad.

durron597

@ThomasOwens I'm confused, do you think most / all of the remaining 29 should be saved, or do you think most / all of them should be closed

Ixrec

4:10 PM

does changing an answer to community wiki status "throw away" any rep I've already earned on it, or just make all future rep go to Ms. Community instead?

enderland

@Ixrec the latter, I think

durron597

@Ixrec Pretty sure it's the former actually

enderland

I'm pretty sure when some of my answers on Workplace went to CW I didn't lose rep

Ixrec

the canonical MSE answer doesn't seem to address this

Robert Harvey

That's not what the Wikipedia article says. The Wikipedia article says that the specification itself is informal (which you've already stipulated to), not the Wikipedia article itself. There are numerous citations at the bottom of the page. That there exist Wikipedia bigots does not, in and of itself, make the information inherently unreliable. — Robert Harvey 1 min ago

Ixrec

4:12 PM

wow, I'm glad I didn't have time to look at that .ini question

@durron597 have you done the CW thing before or just enderland?

Robert Harvey

I think I misread him a bit but still. It's an INI file, FFS.

Not rocket science.

psr

@JimmyHoffa Thanks! I'll pull this after I get home (er, or soon - parents coming in to town today).

user114359

@RobertHarvey he is still looking for a reference to a standard, therefor off-topic.

durron597

@Ixrec Hmmm, maybe I'm incorrect

Thomas Owens

@durron597 They are more likely to not be closed, I think.

Ixrec

4:14 PM

ok, if my function(void) answer gets another substantive edit I'll change it to CW and see what happens

durron597

I think I may still have 20 rep for this answer (not 80)

Thomas Owens

The others were so obvious closures.

Ixrec

btw, I'm looking at the work-environment tag now and it's all over the place

I definitely would not claim all of them are good on-topic questions that just need better tags

durron597

@Ixrec I actually don't even think the work-environment tag itself is necessarily bad

Thomas Owens

You can vote, then.

I just wasn't confident enough to mod hammer them.

Ixrec

4:16 PM

I will, just saying so here since durron appeared to be polling us

Robert Harvey

Goddammit, Skype.

durron597

@Ixrec If you think some of them need to be closed, the right thing to do is not star my pin and VTC the bad ones.

Ixrec

that's what I'm doing

Jimmy Hoffa

@psr just don't tell them all my secretz! They mustn't see my magical OWIN WEBAPI REST CRUD CODE !!

durron597

This question is probably fine, on-topic, and the tag is appropriate

but maybe there's another tag that conveys the same thing but better.

Ixrec

4:20 PM

I think it's at least being misused on some of these questions

have to finish going through them first

it doesn't strike me as a particularly good tag just because it's so vague, I think some of the questions that are using it correctly probably could have a better tag (I think a few of these could use project-management for instance)

durron597

@ThomasOwens should we synonym ssis to sql-server? What do people who are more familiar with those products think? (@RobertHarvey) ?

Stack Overflow doesn't have them synonymed. Here, ssis only has 7 questions, of which 4 are closed.

overexchange

You pass a string, which is a class name and a new object that type gets returned. Factory pattern. Where do we generally require in real time?

enderland

Google: 90% of our engineers use the software you wrote (Homebrew), but you can’t invert a binary tree on a whiteboard so fuck off.

Tweeted by mxcl on June 10, 2015 at 5:07 PM

durron597

I asked in the C# room on SO, they said "don't synonymize".

Thomas Owens

Did they say why?

It may not be appropriate to synonymize on SO, but that doesn't mean it's not approrpaite here.

Jimmy Hoffa

4:30 PM

@durron597 why the hell would you make those synonyms?

They're totally different things

durron597

in C# on Stack Overflow Chat, 2 mins ago, by Squiggle

@durron597 similar suite of products, but SQL-Server is a RDBMS and SQL Server Integration Services is an ETL tool built on top of SQL Server.

Ixrec

that sounds pretty definitive to me

overexchange

People use facebook/twitter/gmail/whatsappchat, when do they get time for office work and self improvement?

Ixrec

unless we want to merge them into something like sql-server-products

durron597

@JimmyHoffa Because SS stands for SQL server, and I've never used SSIS

Jimmy Hoffa

4:31 PM

@durron597 SSIS is an ETL tool - it has lots of very specific things to it that one would never deal with in normal sql-server dealings

@durron597 no worries, just think of it like this: It's the difference between IIS and ASP.NET

SSIS relies on sql-server, but it's a distinct set of tools

I haven't touched SSIS in years but I've done tons with SQL server

durron597

@JimmyHoffa Yeah, I understood the first explanation. I didn't understand it 5 minutes ago but I did when Squiggle explained it

Jimmy Hoffa

ok

durron597

I was just saying why I asked the question in the first place

/me doesn't do a lot of work with databases. Certainly not with ETL tools

Hmmm....

The market data we have, we store it all in .csv

would it make sense to keep each market event as a record in a database?

Would that make our historical simulations more performant? They currently have to parse data from these flatfiles

Thomas Owens

You're dealing with events? Should be using a complex event processing engine. But yeah, for history, I'd make each event a database row.

durron597

But we're talking about hundreds of gigabytes of data

This isn't just database anymore, it's "data warehouse" level

Thomas Owens

4:36 PM

What about the advantages of querying that data rather than parsing all of it?

Ixrec

where I work that's an entire division of the company I've never visited...

I'm fairly sure we aren't using .csv files though

Thomas Owens

You can make smaller, more focused tests and simulations.

durron597

one thing we never need to do is search the data. it's stored in the .csv file in the order it needs to be read, and it needs to be in that order every time

it's already sorted so that different securities are in different files

Jimmy Hoffa

@durron597 if you don't really know much about RDBMS and appropriate design techniques for it, it would be great to learn - extremely valuable, and likely a good solution for your problem. However it would also be fraught with peril due to your inexperience with designing that stuff (I assume?). CSV would likely be the safer thing as it's a lot harder to correct a badly designed database vs. some hand full of CSV files

durron597

if it were in a database, and I wanted to do a SELECT, i would never need to do a WHERE clause.

Well, I need to learn persistence anyway because we restart the production server every day and it loses everything. it doesn't matter most of the time

Jimmy Hoffa

4:38 PM

I would encourage you perhaps to start small - take some simple relational subset of the data you guys deal with, and try to work out a simple nominal database for it and learn to work with that, something where you're only dealing with 5-10 tables. See how that goes after a while

durron597

i would call myself proficient with basic sql syntax but not fluent.

I took a databases class in college

Jimmy Hoffa

@durron597 SQL syntax is a totally different ballgame than DB design

Robert Harvey

@Snowman Already voted to close. It's almost become a reflex.

durron597

As far as our historical data is concerned, i'm mostly talking about like, performance advantages

Jimmy Hoffa

It'll be a lot easier for you to correct those 5-10 tables over a couple of months as you see how DB's need to grow and change vs. if you like 40 or 50 that have flaws in their structural concepts. Add another subset of data a couple months later, and so on

durron597

4:40 PM

are there practical reasons to move the data from CSV files into a RDBMS when it's already in well-sorted CSV files and we don't need the indexing advantages of a DB

Jimmy Hoffa

@durron597 it would definitely have performance advantages; modern RDBMS have been tuned to be highly performant for decades, they've got tweak after tweak

durron597

Yeah, I plan to get better at db design as I learn JPA for our server persistence

@JimmyHoffa would it really? if I never need to do a WHERE query, never need to do a JOIN, and only one type of ORDER BY when the CSV file is already ordered that way?

Ixrec

I would think being in a DB also makes it easier to add various analytics when you do need them

Jimmy Hoffa

@durron597 ACID allows you to distribute the work far more easily and safely, RDBMS do lots of in-memory work reducing IO loads significantly, they structure their data in high performant structures

Ixrec

where I work things like VWAP are a big deal, and there's no way we'd be able to calculate that efficiently with .csv's

durron597

4:42 PM

@JimmyHoffa I should mention I never need to do the equivalent of an UPDATE or an INSERT either

It's really like select * from datatable, then process all the returned rows.

Ixrec

does that processing include anything like grouping or summing or sorting?

durron597

@Ixrec nope.

Jimmy Hoffa

@durron597 Like I said, I suggest you move a small subset of your data into an RDBMS, get acquainted with it, after you've learned more about it all you'll have a better understanding of the benefits. I mean, there's a ton to know about an RDBMS, I think for you to understand how one could benefit you, you'd have to know more about them. Your inexperience with the tech colors your judgement on this topic...

durron597

the production application reads data in chronological order (as prices change, etc.) and then makes decisions based on things like Exponential Moving Average

Jimmy Hoffa

It would take me too long to explain the benefits to you just as it would take too long to convince a procedural programmer why OO is good- some things must be learned on your own... Perhaps it's not the right solution for you but I don't think you have the experience right now to make that decision in an informed way.

durron597

4:45 PM

but it wouldn't make sense to store the EMA in the database itself, because sometimes we run simulations with an EMA alpha of .07, other times .1, etc.

@JimmyHoffa That's why I'm asking. I know the tech well enough to know about performance advantages on things like indexing and sorting and cross-referencing and filtering but I don't need to do any of those things

Ixrec

the short explanation is probably that an RDBMS has far more flexibility (in performance and functionality) than any "ordinary" file format

durron597

I don't need to do any of the tasks that I typically associate a database being better than flatfile at

Ixrec

the thing I really care about personally is the flexibility of the queries and storage schemas; if you're sure you'll never, ever need to add a feature that your .csv format can't handle, then I guess it's okay to stick to .csv

durron597

@Ixrec it's about effective use of time.

Ampt

Well this should be interesting... the boat trailer wasn't ever titled to the guy I bought it from... I'm guessing because the back of the CoO ran out of room on the back... the DMV is gonna love this.

Jimmy Hoffa

4:48 PM

@durron597 like I said, I think your inexperience with it is keeping you from making an informed decision... RDBMS is a large complex thing that takes a great deal of time to understand well. You haven't worked with one enough it sounds like to have more than a shallow understanding.

Ixrec

true, if it's anything like my job you'll just end up making the switch when you get asked to implement a feature that requires a real database

durron597

I actually agree that an RDBMS would be better, but is it a better use of time than, say, supporting a new order type. implementing a new algorithm take profit logic

Jimmy Hoffa

@durron597 frankly I don't know your exact data behaviours well enough to know if it's the right solution for you, and you don't know RDBMS well enough to know... I simply suggest you expand your knowledge on RDBMS with a little practice on some of the data that it would be safe to meddle out of the CSVs. Then you may have a better idea of if it makes sense or not.

@durron597 which would be more maintainable in the long run? In the long term maintainable solutions pay with their savings for features like you suggest. Again, I don't know your particular technical situation; CSV may be more maintainable in the long run.

Robert Harvey

@RobertHarvey I wish more people had this mindset. * sigh * — null 32 mins ago

durron597

Complete non-sequitur, OMG: reddit.com/r/AskReddit/comments/39d13i/…

Robert Harvey

4:54 PM

Huh? Even though I took a veiled swipe at him. He must be having trouble with his Wikipedia bigot boss.

durron597

@RobertHarvey Yeah agreed

Robert Harvey

@durron597 The comments are more interesting than the story.

durron597