« first day (1809 days earlier)      last day (2176 days later) » 
00:00 - 18:0018:00 - 00:00

user114359
6:04 PM
@Duga the same person thought that question would be more appropriate at both Super User and at Programmers.
 
@Snowman the same user avoided some pretty classic train collision action then
little older story from a Toys buyer for Amazon gawker.com/inside-amazons-bizarre-corporate-culture-1570412337
All the stories seem to be pretty consistent in details about how it's all handled, but few of them come from the engineering side of the house; who knows
 
6:20 PM
@ThomasOwens: I think it doesn't really matter because for any company, no matter what kind of evil stuff they're up to, the CEO is going to defend the companies actions and say it's a great place to work
the tbray article writer is completely clueless as to how different his situation is from 99% of other human beings in the US
 
the gawker email sounds the most candid and creepy, but it's just as much anecdotal as the defenders. Really the NY Times article is an accrual of over a hundred interviews it said, so while it clearly started with a bias- and is written by a competing media outlet, it can still claim the most substantive data. Still, who knows
 
Hi peeps
-_-
Ummmm
 
6:39 PM
hi
 
Ooh, someone said hi back :)
 
user55340
At half past the hour, we hold five minutes of silence in honor of the beer that was spilled last night, or that it's Monday, or for the project that was killed, or for the build that failed, or if we just want to kid you.
 
user55340
;-)
 
@ThomasOwens just because a CEO says "no one will work long hours" doesn't mean that the requirements of their teams will not be that way :(
 
Alright :)
 
6:42 PM
we had dozens of failed builds today
our inboxes got spammed hard by the CI bots
 
user55340
@Ixrec retaliation time?
 
@MichaelT nah, if we go after them they'll counter with the worst thing a CI system can ever do
be silent the next time it fails
 
nice
 
user55340
As I suggested- "time for retaliation?"
 
6:45 PM
sadly most of our build failures can't be pinned on a particular individual, or even our team
but if we put one of those in New York...
 
user55340
You need to nuke it then...
 
I'm having difficulties with data vs widget lifetime in a Qt app. I'm not sure if I can pack it in a good question for the main site
I try to divide data and widgets, so I have a POD struct that I use for I/O and a controller class that wraps around the data and emits signals when something changes. Now I have two options for the widgets: create them with a reference to the controller class, or add a method that accepts a pointer to a controller later.
 
user114359
@Christoph Why not create a widget that has a functor or lambda so the widget can be decoupled from the controller and model?
 
what would I gain with that, and why?
 
user114359
7:00 PM
Separation of concerns, cleaner design, object reuse, makes sure all code for modifying models is located in the same area for easier debugging
 
all the logic is in the controller, apart from those parts that are specific to the widget
Creating the widget with a reference to the controller (in the constructor args) makes sure that the widget only gets a valid object unless I screw up the controller. This, however, can't be done with Qt creator because it only allows me to create widgets with a single parent pointer argument
 
user114359
I guess I don't understand the question. All the logic is in the controller, but the widget needs a pointer to the controller somehow?
 
so I tried adding a setController(controller*) method, and end up checking for situations that might arise when no controller has been seet
 
user114359
The widget must be doing something with the controller
 
the widgets connect to the controller
so what you basically say is that when my widgets need to know about the controller, there's still too much logic in them?
 
user114359
7:11 PM
widgets shouldn't know anything about a specific controller. They should send events (signals/slots) or if that is not possible, use a high-level interface.
 
user55340
data.stackexchange.com/programmers/query/350204/… - discount the stci and look at what months have the most closed questions.
 
Ummm... Quick question: Did any of you peeps talk to some members from Open Source about site overlap and stuff?
 
It wasn't me...
 
user55340
Nope. My only venture there was wondering about the unlicense vs unlicensed code and tags.
 
@Zizouz212 Nope.
 
7:15 PM
Oh. I remember that
Well, ok :) I can't chase down anyone from my site or this one...
 
I want to, but I want to talk to SE staff first. Both Law and Open Source are...lackluster...in the visible statistics.
I don't want to bring it up if one site will just close down and we'll be back to where we are now.
 
@ThomasOwens I'm here if you want to talk to me :)
 
@Zizouz212 You're not SE staff, though.
 
user55340
@ThomasOwens don't worry, when os closes, they'll do an es-engr style "move everything" here.
 
I probably misunderstood...
@MichaelT My site shall never close!
 
7:17 PM
@MichaelT See, I don't want that to happen. If anything, Open Source and Open Data can be merged into some kind of Open Stuff, Yo!
That would make more sense to me.
 
@ThomasOwens It would make things a site like that really broad... I think like 80% of the questions at open data is for data-requests...
 
user55340
@Zizouz212 mostly poking fun of @GlenH7 recent tribulations with that move and mess.
 
@Zizouz212 Maybe true.
 
There's a new Open Science site that's in closed beta
 
@Snowman I'll try to do it that way but it might be an adventure for more specialized widgets
 
7:20 PM
does "Open Science/Data" mean anything special beyond "no paywall"?
I know Open Source has the whole free as in speech thing
 
Basically, what I've seen is that Open Data is more for data, but now it's all become like where can I find this i.e. Data requests
 
user41796
@MichaelT We survived. Fortunately, embedded was still fairly small by the time they killed it off. So there was a workable number of questions to review.
 
Open Source is more the software and hardware side of things, and deals with more aspects: in particular project management, monetization and stuff like that
 
What is Open Science?
 
user114359
@Christoph It has been a few years since I used Qt but I know signals/slots are baked into the framework at a fundamental level - are you implying this will not work in your case?
 
7:21 PM
@ThomasOwens The open version of science...
 
user41796
Advanced notice would have been nice, but whatevs.
 
@Zizouz212 What is the open version of science?
 
user41796
@Zizouz212 Sounds like "we're a research librarian for you" type site, which doesn't do so great in the SE format.
 
@ThomasOwens I'm on the site, but it doesn't look distinctive at all... 90% of things can probably fit on Open Data/Source
@GlenH7 Honestly, look at the top tag: opendata.stackexchange.com/questions
 
@Zizouz212 Open Data : Open Source :: Software Rec : {Stack Overflow, Super User, Programmers, etc.}?
 
user41796
7:23 PM
> Do my work for me, please.
 
1st place: [data-request] 934 questions.
2nd place: [usa] 181 questions.
You would imagine that Open Data and Open Source would have more things in common, but geez, in my view, Open Data is just, "oh find this for me, please!"
 
@Snowman signals and slots work very well, but let's take the following scenario: I can, as in so many applications, have an create or open a "project", and close it. So the data my widgets are supposed to display might be there or not. When I lay out my widgets in Qt creator, the "project" lifetime is completely decoupled from widget lifetime
 
user41796
Only 1500 questions after nearly 2.5 years in Beta? That's not promising
 
@Christoph are you only talking about creating widgets in QT's designer?
 
Also, I should make you guys aware that Open Source is trying to change its name, in case that may be a factor in something. (reason for much lower activity now?)
 
7:26 PM
@Zizouz212 I've noticed that. Does it have support from SE?
 
user55340
What are you trying to change it to?
 
And that's my second question.
 
So it has full support from SE
Originally we were trying to change it to Free & Open, but the issue was ambiguity, so it's back to the drawing board
 
@enderland not really - I'm talking about managing data vs widget lifetime.
 
user114359
@Christoph If a widget cannot do anything because there is no model to back it, why is that widget not disabled? I.e. grayed out?
 
7:28 PM
@Zizouz212 Have you considered Free Culture or Free Content?
 
Alright, so the rationale for the change was to find a neutral name (Open Source veers with the Open Source Initiative) and we were trying to find something that eliminated any political bias.
@ThomasOwens I suggested Open Info, that was shooter down. Apparently, there was concern over Free Culture - associations with Creative Commons
 
@Snowman I'm not sure - if that's the usual thing to do I can try that of course.
maybe I'm a bit confused about the need to disconnect widgets. When I delete my sending objects, the widget doesn't need to be disconnected explicitly because the object management takes care of that
 
@Zizouz212 Free Content!
 
user114359
@Christoph Things I would look at are disabling widgets that cannot do anything useful, or at the very least adding null checks. No controller or model or whatever? Don't do anything.
 
You can go ahead and suggest that :D - meta.opensource.stackexchange.com/questions/542/…
 
7:34 PM
@Snowman these null checks piled up quickly. that's why I started thinking about this
 
user55340
Having free, unless closely committed with open or libre will have difficulty with the free beer seekers.
 
@Christoph ick, null is a monoid, if they're piling up you need to come up with a way to use coalesce to let the nulls bubble out and only check at the last moment (absolute top of the stack)
alternatively never null check and do a try-catch near the top of the stack with the expectation that you'll get lots of dereferences and deal with them accordingly
 
aaaand so my alternative approach I thought about was to create the UI dynamically, passing in a reference to the data to connect to in the constructor of a widget.
 
@JimmyHoffa my first thought was "in Javascript I just use || and && chains to deal with this" and then I remembered that booleans are monoids too
 
so when my data goes out of scope I can just delete the widgets, and create new ones when I have new data
 
7:38 PM
@Christoph I use upserts a lot: Update if it finds, create if not.
Then you can delete all you want and rely on the upsert to handle the scenario where it is or isn't there
 
user114359
@Christoph You could also use the null object pattern to help with null checks.
 
I have never heard about upserts
and apparently I used null objects before without calling them that way...well ok that's doable I'd say, just return something dummy-like if nothing more useful is available
 
@Christoph it simplifies things a lot by giving you one entry for processing of data changes when you don't care whether the thing already exists or not. Keeps you from having to branch the code that's just trying to pass the data around. Of course in transactional scenarios you may wish to be sure you're not updating an already existing entity if you specifically don't want to, but you can discern if that's your scenario or not case by case
 
user41796
@Christoph It's a combined insert / update. Update if it exists, insert if not. I believe later SQL Server versions offer it as a direct command.
 
@GlenH7 merge statement makes it an atomic high performance set operation
but that's SQL, I'm just referring to the generic concept of data; if I have things in memory, widgets or backing objects as he's talking about, I'll create upserts happily
 
user41796
7:42 PM
@JimmyHoffa We're still hand-rolling our own. :-( One day we'll get to that level of fanciness
 
ConcurrentDictionary recognizes it's a valuable sensible operation especially necessary for being atomic in the event that you want multi-threading safety by giving AddOrUpdate right on the class
 
I wish we had upserts
 
@JimmyHoffa I would upsert the widget, or would the widget upsert the data? I'm not sure about your point of view there
 
it's important for multi-threading because you have an inevitable race condition between does it exist and ok now create it unless you bundle those into an individually atomic operation.
@Christoph just referring to the idea of getting rid of the null checks, you said delete the <data? widget?> and then when the <something> comes back you'd recreate it. I was saying use an uspert so whether you're "recreating" or just "updating" doesn't matter. Simplifies that return vs. first time bit of code
 
understood that now, thanks
 
user114359
7:56 PM
I smell a flame war with @MichaelT.
 
user114359
In my opinion, you are both wrong. Recommended reading: Why was my question closed as primarily opinion-based?Snowman 43 secs ago
 
I think everyone agrees that question is getting what it deserves right now
 
user55340
Needs four more close votes to prevent answers (that would get deleted)
 
user114359
@MichaelT I've given her all she's got, Captain. Come back in four hours!
 
user55340
Been out since noon.
 
user41796
8:02 PM
Needs one more
 
The tragedy of that question is that he's only scratching the surface of Java's/C#'s verbosity.
 
user55340
"Use scala instead"
 
8:14 PM
added VTD
 
1
Q: Revisiting what is on topic, with respect to new sites: Open Source, Law, and Software Recommendations

Thomas OwensIn the past 2 months, two new sites in the Stack Exchange network opened up and are now in Public Beta: Open Source and Law. These are two topics that our Help Center's /help/on-topic page explicitly calls out (emphasis mine): If you have a question about... software requirements so...

 
user55340
@Telastyn those votes I still have.
 
2
Q: Revisiting what is on topic, with respect to new sites: Open Source, Law, and Software Recommendations

Thomas OwensIn the past 2 months, two new sites in the Stack Exchange network opened up and are now in Public Beta: Open Source and Law. A third, Software Recommendations, has been around for about 1.5 years. These are three topics that our Help Center's /help/on-topic page explicitly calls out that are well...

 
user114359
@ThomasOwens can I give your answer to that question a super +10 upvote?
 
@Snowman No. The other mods/users from the other sites need to weigh in first
 
8:25 PM
is that intended to be a change we make now-ish or some time in the future when these new sites have clearly "made it" and won't be disappearing overnight?
 
Feel free to ping me if you want my input in anything about that post above
 
@Zizouz212 We want your input.
 
(btw, I'm a moderator at Open Source)
 
Please comment or leave another answer.
@Ixrec I asked SE staff. It's OK if we talk about these changes - the sites with moderators are not likely to go away.
 
user55340
Btw, remember to up vote the community promo advert.
 
user114359
8:29 PM
@ThomasOwens I know, my point is I agree 1,000,000%
 
Because one million percent is a perfectly valid concept.
 
hyperbole is a perfectly valid concept
 
@Ixrec exaggerating is always a bad idea
 
OK. I'm out for now.
gn Programmers.
 
@ThomasOwens don't miss!
 
user55340
8:32 PM
@enderland always is an exaggeration. Sometimes is closer to the truth.
 
user114359
Never say "never"
 
user114359
Only Siths speak in absolutes
 
user55340
 
9:03 PM
My main concern is how the migration would work. Because we can't have migration votes to beta sites, if we go this route, everyone who asks a previously on-topic licensing question here will immediately have to be flagged for mod intervention, and there's likely going to be a big delay before they end up on the "correct" site, during which time they might cross-post. Not to mention the sudden need for a "massive cleanup" as durron pointed out. For that reason I can't honestly support making these off-topic, even though I would support recommending OS.SE as the "more on-topic" site. — Ixrec 1 min ago
 
How can you have an online training that goes down due to internal java errors for 4+ days... seriously?
 
@MichaelT I've got to become a better writer. Your answer is basically the same as mine yet you are getting votes and I'm not ;)
 
@durron597 SE is really good at helping you get better at writing...
 
user41796
@durron597 His writing skills are almost as good as his google-fu, so that's no small hurdle you're setting for yourself there.
 
@GlenH7 Yup.
 
user55340
9:10 PM
@durron597 join a writing site for a few years.
 
user55340
everything2.com Is where I spent some time.
 
user55340
everything2.com/?node=m_turner - take a fair bit of it with a heavy seasoning of mid 20 angst.
 
@MichaelT Lol my brother was obsessed with puzzle pirates for awhile
I never played it though.
 
user55340
It was a good game when I was there. Many years ago. The open economy is very interesting in a game.
 
user55340
9:24 PM
Haven't been on in otherwise recently. Last I looked, the mac installer was broken.
 
In other news, Amazon loved my code review from yesterday
hyachachaachachaa
 
@durron597 :D
That was fast turnaround
 
@enderland Maybe they're desperate re:high turnover
 
9:40 PM
Well you also said you aced the code questions
 
@enderland Well that too, if I had bombed them they wouldn't have called me no matter how desperate, lol
 
Well you never know... though if you listed high salary requirements, probably not in that case ;)
 
@enderland 105k or whatever someone put (from glassdoor) was not close to what I put.
 
Which makes sense, a cross country move to a HCOL area means you should want a lot more
 
that number is much closer to what I make now than what I put.
And I definitely want a pay bump well above COL changes, so yeah.
Really I just want to make more than my brother who's 10 years older than me and is also a programmer :D
 
9:43 PM
lol
 
My other brother is a 40 year old doctor specialist, I didn't ask for more than what he makes ;)
 
Depending on what type of doctor, you might
oh
guess not ;)
 
pulmonary and critical care. Plus he cosigned my apartment lease a few years back so I have some idea what he makes ;)
i put a range, that number is pretty close to the bottom of it.
 
yeah I figured looking at COLA between locations
 
and I said "close to 105k" but you don't know if that means "I make 85k" or "I make 125k" currently (and I won't say, obv)
 
9:46 PM
I think you've said more accurate than that before anyways ;)
 
probably.
Well, at least I would (removed) a more accurate number.
but, yeah, googling "senior java developer salary houston tx" is accurate enough, that's why I don't care about sharing.
 
Yeah
 
Oh, neat, I think I capped on SO today.
20k here I come
then immediately stop posting answers ;)
 
Haha
 
user41796
@durron597 You get more delete votes all the way up to 35k.
 
9:53 PM
@GlenH7 Don't you want me to put effort into getting 10k on Progs?
 
user41796
@durron597 There's that too. But then you could always bounty me up past 3k on SO... :-)
 
programmers.stackexchange.com/questions/293917/… wow that guy would have got some abuse if the question wasn't quickly closed
 
@GlenH7 As long as you bounty me up on Progs, hey.
 
user55340
@whatsisname I'm rather quick on the delete vote for poor questions that would take excessive heroism to rescue. Get then off the front page and newest questions to allow other questions more visibility.
 
yeah
it was comically bad
 
user55340
10:21 PM
Override_function? Ok - so that's what those php coders are smoking.
 
10:42 PM
great, the device appears in Device Manager but doesn't work with our application
fun
 
11:09 PM
has anyone seen an article recently about the whole amazon thing, but where the writer has actually been there for more than 18 months, and isn't the CEO?
 
Yes, we talked about it earlier this morning.
Sorry, that was the 18 month manager
There was another one, but I didn't look closely.
 
11:33 PM
This may be a better fit for programmers.SE, but falls a bit off-topic here. There's no problem to be solved, it's an opinion question. — Adam Smith 22 secs ago
 
00:00 - 18:0018:00 - 00:00

« first day (1809 days earlier)      last day (2176 days later) »