Tavern on the Meta

Anonymous

4:08 PM

@Magisch can't have that here! I was hoping to keep building in from out of bounds but didn't have enough resources to keep up with the eye of the storm when it moved away. ;)

user315433

> the most common example by far being "Thanks" deleted and "You're welcome!" left behind. meta.stackoverflow.com/questions/364957/…

user315433

Of course, if one is 1-flag and the other isn't.

user315433

But it's not like the remaining comment has less meaning now; it's impossible.

user315433

Math.SE is chock full of YW comments because one of the most prolific answerers left them after every single TY comment they got.

user315433

But I don't think mods would be excited if I started flagging those at any noticeable scale.

Anonymous

4:12 PM

@Magisch PHP has safe idiot-resistant cryptographic constructs (i.e. what 99% of developers actually need, rather than just pointy foot-shotgunning primitives) in its standard library now. Python, Ruby, JavaScript, and even .NET don't have that, as far as I know. We can thank Scott Arciszewski, who helped lead the clean-up of crypto answers on Stack, for that.

Anonymous

The Power Of Not Presuming Everything Will Always Be Awful :P

Anonymous

But let me tell you: Closure compiler is and will forever be awful.

Anonymous

It was the best in the world for a little while when there was zero competition and no bar to match.

Anonymous

and it has made ~0 progress since then. Sad!

MetaEd

@Derpy I've been derusting my JS to produce those animations. And yuck.

I hear some of the yuck has been fixed in the v6 standard.

Shog9

4:28 PM

JS is beautiful in the same way that BASIC was beautiful back in the '80s: immediate gratification

You can either use that, or fight it.

MetaEd

@Shog9 I learned BASIC in 1976, which probably spoiled me for learning decent languages. You can see it in the languages I became expert in subsequently: sh, csh, and Perl. So what you're saying is I'm staying true to form.

@Shog9 "It is practically impossible to teach good programming to students that have had a prior exposure to BASIC: as potential programmers they are mentally mutilated beyond hope of regeneration." —Dijkstra

Shog9

@MetaEd you kinda gotta read the context there to know what he was talking about though

MetaEd

Are you making a scope joke about a Dijkstra quotation? :-D

Shog9

@MetaEd naw. Dijkstra had a notion of what good programming should be that... I'm not sure it ever really matched up with what most programmers consider good programming.

Bart

The EWD documents are a great read

MetaEd

4:38 PM

@Shog9 Yeah, I've read lots of Dijkstra's remarks about good programming. His philosophy is hard to disagree with, but it seems largely ignored.

Shog9

I don't mean any slight to the guy's memory by that. Brilliant, and undeniably influential. But that particular quote gets tossed around a lot in contexts where it doesn't make a lot of sense.

I'm butchering the context by saying this, but my tl;dr would be... Programmers should know what their programs will do prior to running them.

That's what BASIC discouraged - by design.

And why I compared it to JavaScript.

Bart

There is one EWD where he visits a company (not sure if it was IBM) and he describes his realization that correctness is not part of the business model, but service contracts are

Shog9

The pitfall I see a lot of people fall into when they go down that path is... They believe systems documentation.

You can't trust systems programmers. Or hardware engineers. You gotta go in assuming that the hardware is going to lie to the system and the system is going to lie to you even when the hardware doesn't.

And, of course... You can't trust the users either.

Bart

Don't trust anyone, not even yourself

Shog9

So the users lie to you, and then your program acts on those lies and tells the system to do something, which tells the hardware to do something, but whether that actually happens...

At that point, it doesn't matter if your program is provably correct.

Some of the best planned software I've used has been an absolute nightmare, because it would completely lose its shit if a network connection dropped or a file wasn't where it was supposed to be or I clicked two pixels to the left of where I was supposed to.

And there's no diagnostics or fall-back or error-handling or undo... Because the program was correct.

So these folks are great when you put 'em to work building algorithms, because when they come up with something you can be pretty confident that it'll do what it's supposed to as long as all of the inputs are verified.

But... You gotta get a sloppy BASIC programmer to verify those inputs, including the ones that have no valid scenarios for ever being wrong.

And then design error-handling for when those inputs are catastrophically wrong.

And diagnostics for identifying how they got that way.

MetaEd

4:48 PM

@Shog9 Yes, that's not a bad gloss. And there are all sorts of reasons that's tough, starting from hardware. A great modern example of that is the recent security issues that came to light because of processor buffer design.

Shog9

And undo so that the poor operator who hit the wrong key doesn't get fired.

MetaEd

One of the most egregious hardware problems in the field today is heavy reliance on PCs without ECC memory.

Shog9

@MetaEd heh... Yeah, that'll be a great "don't be too proud" story for generations...

@MetaEd yeah. Still don't understand why that's not ubuquitous.

I mean... I do...

but, I figured by now it'd have just gotten rolled into everything and the cost eaten up

...oooh, Google Docs has overlapping circles for viewers now

user315433

While its script language is still based on ES5.

user315433

But hey, circles.

Shog9

4:54 PM

overlapping circles

MetaEd

@Shog9 Lack of ECC just makes me shudder. How many desktop PCs are there in process control, health care, and other critical applications and nobody though to put in ECC SIMMs?

user315433

Almost makes me about not having arrow functions when they would fit perfectly into array processing scripts in Sheets.

MetaEd

Cosmic radiation means it's not IF a bit changes, it's WHEN.

Wow! This means that about 1 byte in my PC gets corrupted every two days. — Stefan Monov Sep 26 '10 at 7:10

That whole answer is great on practical numbers, studies, etc.

Bart

@Shog9 are they freehand though?

Shog9

gotta be something for the .1 version

MetaEd

5:17 PM

tap tap this thing on?

Jenna Sloan

Did the chat go down for anyone else just a minute or so ago?

MetaEd

Yes.

All three chat servers seem to have bounced.

So maybe a network device in front of them had a cardiac event.

Jenna Sloan

I was like "OMG, the chat is down! What am I supposed to do with my life now?!"

MetaEd

@JennaSloan You could write a poem about it.

You could refresh the @StackStatus twitter feed until they admit it. Which didn't work for me.

user315433

7:43 PM

In a SEDE parametrized query, how do you use a string parameter that contains an apostrophe?

user315433

Okay, doubling ' works for simple cases but how about this:

user315433

set @sql = N' blah blah where lower(Text) like ''##pattern##'' and stuff'

user315433

Ah, quadruple ' of course.

Jon Ericson

This is why alternative quotation marks were invented:

46

A: How do I escape a single quote in SQL Server?

If escaping your single quote with another single quote isn't working for you (like it didn't for one of my recent REPLACE() queries), you can use SET QUOTED_IDENTIFIER OFF before your query, then SET QUOTED_IDENTIFIER ON after your query. For example SET QUOTED_IDENTIFIER OFF; UPDATE TABLE SE...

My test on SEDE.

Now with parameterizations.

user315433

8:06 PM

Yes, that one was a cross-site query so it had a parameter inside a query which was formed as a string inside the top query.

user315433

8:20 PM

How many declined comment flags does it take to get flag-banned?

user315433

Wait, so declining comments flags doesn't actually stop them from generating more crappy comment flags? Well crap. — animuson ♦ Dec 12 '16 at 3:09

Catija

A lot? I don't actually know that you can get banned from comment flags.

user315433

Current plan: flag YW comments until enough mods become unhappy to either suspend me or ask someone to add a regex for automatic handling of those.

rene

@SFTP add the type to the parameter and you no longer have to worry about escaping: data.stackexchange.com/stackoverflow/query/829126?pat=%27

Catija

Or you could just ask...

rene

8:27 PM

@SFTP I have written an answer here that explain the differences between parameters that take a type and those that don't.

user315433

+1

user315433

#FoundersForChange https://www.foundersforchange.org/

Tweeted by StackOverflow on March 22, 2018 at 8:15 PM

user315433

What's that graph behind his left shoulder?

user315433

Looks like it could be traffic to SO.

user315433

Or is that a frame from xkcd.com/time ?

rene

8:36 PM

I was more thinking the winterbash 2017 drawing. Can't find the chicken though.

user315433

"TWA flies direct"... yeah. It was the first U.S. airline I flew with. sigh

user315433

M.A.R. ಠ_ಠ

@SFTP is this the burnt people from TWA plane crashes?

I can't liken it to anything else

user315433

I thought the shape was similar in style, at least.

M.A.R. ಠ_ಠ

8:45 PM

Oh

The world makes more sense when I turn on VPN

Anonymous

9:10 PM

@MetaEd I anticipated that learning about rowhammer would lead the industry to broadly adopt ECC quickly.

Anonymous

Nope!

Anonymous

Maybe after its too late and we're all being rooted from our web browsers.

Anonymous

(I got on the confidential list before that vulnerability was disclosed, and was hyping up my coworkers for what a big deal it would be. The eventual response did not match my rhetoric.)

MetaEd

@JeremyBanks Well for most people it just means Bill Gates installs a new version of Windows that they didn't particularly want.

Sonic Wizard

@Shog9 Why did you change the order of close reasons in the close dialog? I have a dead-on muscle memory for RC flagging off-topic questions, and that's probably why others recently closed an off-topic question as the new site-specific reason you changed. Can we please get the old order back?

Shog9

9:20 PM

@SonicWizard to break muscle memory

Sonic Wizard

@Shog9 Why would you want to do that? Do you have data showing that under the old system, a lot of people had incorrect muscle memory?

Catija

They're unlikely to notice the change in text if the order is the same.

Sonic Wizard

Is that so?

Shog9

@SonicWizard you read the dialog more carefully now, didn't you...?

Yes, I announced the change and messaged a bunch of people about it, but I didn't exactly broadcast anything.

Sonic Wizard

@Shog9 Do you have predicted data showing that the number of incorrect closures done now would be far fewer than the number of future incorrect closures where people think of the old system?

Catija

9:23 PM

@SonicWizard How else do you know which close reason you want if the spot you were used to clicking is drastically different. Plus, if they keep using that slot, you can tell who's not really paying attention when reviewing.

Sonic Wizard

I think I saw a few people complaining about broken user scripts that picked the fourth option in the list

Catija

... darn.

Shog9

That... Is a stupendously bad way to write a userscript

Sonic Wizard

What's the correct way?

Shog9

Depends on what you're going for

If you want one specific reason and only that reason, then go by ID - they aren't re-used.

If you want to fail if a reason isn't available, match on the text of that reason if the option isn't available

If you're just listing reasons, then pull them at run time from either the dialog or the API

Sonic Wizard

9:26 PM

Thankfully if I picked the wrong reason, it wouldn't show up publicly - I'm just RC flagging, not voting to close. Plus, there's still room for error for one voter to screw up.

@Catija What do you think of my new name?

MetaEd

@SonicWizard RC flagging

Sonic Wizard

@MetaEd Recommend closure flag. It's the type that doesn't notify moderators, just sends posts to the CV queue and is marked helpful as soon as someone casts a close vote.

Catija

@SonicWizard It makes me think of the hamburger chain.

Sonic Wizard

Which?

MetaEd

@SonicWizard I'm sure you really meant radio-controlled flagging.

Catija

9:28 PM

sonicdrivein.com

Sonic Wizard

Ah, right. I was referring to the Sega character

MetaEd

Of course, if you're using the app over wireless, that is radio-controlled flagging.

Sonic Wizard

The hedgehog behind the mask @Catija

Catija

@SonicWizard Yeah... I know... but... it doesn't mean anything to me. I have no connection to that franchise, only the burger franchise.

MetaEd

I really only use one as a screwdriver.

Catija

9:31 PM

And that connection is... I've been to one in the last decade.

Sonic Wizard

Which explains my tropes: Shadow the Hedgehog, Amy Rose, and Chaos Emeralds.

Shog9

Mmmm... Chaos Emerald Cream Slush...

MetaEd

Can I just say I am really starting to hate depreciation rules.

assetAge = 12 + 12(y-y0) + (12-m0) + (d0<=15 ?1 :0)

Catija

Clearly, Sonic (the drive in) is missing out on some great cross promotion with Sonic (the Hedgehog).

Félix Gagnon-Grenier

9:56 PM

omg ~~ponies~~ tags

user202362

10:21 PM

@SFTP for change? like this?

user202362

10:45 PM

founders begging for changes from investors </trollololo>

Sonic Wizard

11:01 PM

-1

A: I'm changing the "not generally applicable" close reason here on Meta SE

I agree with this idea. Since we've started modifying the text of these close reasons, I'd like to propose something else. I've noticed a similar problem with a different off-topic close reason, though to a bit of a smaller degree: people tend to close fixed bug reports as off-topic with the fol...

@JourneymanGeek What do you think of this?

Never mind, deleting it and posting it as a separate FR

@Catija Actually, I deleted it with the intent of posting it as a separate FR, but realized it's already requested, so mine would be a dupe.

(cc @Shog9 )

Never mind, not a dupe upon close inspection.

@SmokeDetector tpu-

sd k

Shog9

11:19 PM

Hey, @SonicWizard... When I decline one flag with "no flagging twice" after you flag twice for the same reason... Flagging twice more is kinda insulting.

Sonic Wizard

@Shog9 Twice more? Yes, I reflagged as NAA again, but then I retracted it and replaced with a custom flag. So, only once more.

Anyway, sorry for the extra flag. I didn't notice the question explicitly said it wanted a test answer, and just noticed the answer.

Shog9

Yeah, context is important.

Sonic Wizard

Out of curiosity, what made you think of checking my flag history? One of those flags was disputed in review, and never made it to moderators.

user202362

I am still limping

Shog9

11:41 PM

@SonicWizard I see the full history of flags on the post when I handle a new one

nicael

11:56 PM

Looks funny

Transcript for