The Hangar

mee

04:51

hello, anyone in?

Bret Copeland

05:11

I don't know what's going on.

I didn't notice this meta post over the weekend, but I've responded now. cc @JayCarr @Farhan

0

A: Who is this user, who has been removed?

This, ladies and gentlemen, is why you don't create sock puppet users to upvote yourself, and then try to disguise those sock puppets by having them upvote other users. It might not get noticed right away, but eventually you're probably going to get caught, suspended, and have the sock puppet use...

SentryRaven

06:10

@mee Yes, there is people in here. Not sure if all are humans, but hey... can't be picky!

mee

sweet, any RAAF pilots in?

SentryRaven

Not that I know. Most here are from the private or commercial sector. Not many people read this chat anyway, it's like a dozen maybe?

Jay Carr

12:46

@BretCopeland - Is Aviation.SE actually graduating? I thought the site design thing was more or less a joke (right?)

casey

13:46

@JayCarr it was a joke from the podcast.

they were just looking at a list of sites with >10 q/day and no design and throwing some names out there

and someone in chat made that "design"

we don't know graduation status because they made sure not to read that list when picking a site

though I'd say we are a decent candidate for the quasi-graduated status they mention (e.g. higher rep requirements, removing "beta" from the name, and moderator elections, but no design)

Jay Carr

Ooooh, gray area for us! Would it change the rep required for privledges too?

(0ddly the only thing I really want on that list is the site design but...eh)

Oh, jeez

Sorry, it's early for me

I can see it now "higher rep requirements"

Well, that'd set me back a bit. But I looked at what I'd have at that point and it wasn't too bad so...

There you go I guess.

casey

14:48

@JayCarr as long as you are above 3k rep you still would have access to all the review queues, close votes, etc. The biggest change is moving the mod tools from 2k to 10k

Jay Carr

16:28

@casey Yeah, I'd be missing the mod tools for a while at least, but I don't use them all that much to begin with, so I'd be fine.

Bret Copeland

@JayCarr we are starting to get into the territory of consistently 10 questions per day, but that will need to be sustained for several months to show it's not a fluke. Just a guess, but if we kept these numbers for the next 6-9 months, we'd be a candidate for graduation.

Jay Carr

For full graduation or half graduation or...?

I guess there is no third option :)

voretaq7

@BretCopeland . . . so basically if airplanes keep smacking in to each other or mysteriously vanishing without a trace we'll be graduated? :)

I'm actually kinda surprised we haven't had any "How the hell could an F-16 and a Cessna 150 hit each other?!" questions

Farhan

16:45

@voretaq7 So what do you think that how the hell could an F-16 and a Cessna 150 hit each other?

FreeMan

@voretaq7 Well, when they're both attempting to occupy the same 3 cubic meters of atmosphere and the same point in time...

Jay Carr

17:02

Well, when you think about how much air there is out there it really doesn't make much sense when any plane hits any plane ever. But it happens on occassion... I'm curious to hear about how the 150 got where it did. I'm told it was in class C airspace and it seems like the F-16 would have been in contact with the tower so....

Yeah, just waiting for the NTSB report I guess.

voretaq7

17:20

@Farhan When a Cessna and a Fighter Jet love eachother VERY MUCH. . .

Textron AirLand Scorpion

The Textron AirLand Scorpion is a proposed American light attack and Intelligence, surveillance and reconnaissance (ISR) jet aircraft. It is being developed by Textron AirLand, a joint venture between Textron and AirLand Enterprises. A prototype was secretly constructed by Cessna at their Wichita, Kansas facility between April 2012 and September 2013 and first flown on 12 December 2013. == Development == === Background === In operations in Iraq and Afghanistan, the U.S. Air Force primarily used A-10 Thunderbolt II, F-16 Fighting Falcon, and F-15E Strike Eagle aircraft for patrols and cl...

^^ They made a baby! :)

@JayCarr Yeah - my guess is they were on one of the military training routes & maybe not talking to ATC (yet), and I'm assuming the Cessna was VFR and not talking to anyone. But nothing's confirmed yet & I could be completely wrong.

All the reports I've seen say "11 miles north of Charleston" so maybe outside the Class C airspace, but that's all mainstream media and who knows where they're measuring "Charleston" at :)

Jay Carr

17:44

Mmmm...yeah...

Though, it was the FAA that reported the "11 miles north of Charleston"

CNN seems to be providing a few more details from the FAA statement or something: cnn.com/2015/07/07/us/south-carolina-aircraft-incident

The military says the pilot was in contact with ATC when the accident happened, so...there's that.

Sorry, I tend to have a real fascination with accidents for some reason. Always curious why they happened. I doubt there will be much further detail until the NTSBs preliminary findings report though...

Farhan

Pilots who don't have an IQ of at least 130, should call themselves jerks.

Dan Hulme

18:06

That's a really tricky question. I know some people who definitely should not try to fly aircraft. Maybe the instructor is really saying, "I don't know how to teach you." Maybe they're saying, "You can become a good pilot, but if you're always this slow picking things up it will take you way too long and you'll become disheartened." Maybe they're saying, "You have the wrong attitude to it, and I don't know how to change that, but maybe in five years you will be a different person."

Bret Copeland

IQ is definitely the wrong way to think of it though.

ratchet freak

hasn't IQ been debunked?

Dan Hulme

I dunno, IQ is pretty strongly correlated with other measures of ability and potential, so it's probably the easiest thing to judge even if it's not directly involved in learning to fly

e.g. IQ and sense of balance are positively correlated

Bret Copeland

I think of it in terms of aptitude, which is hard to put a number on. Similar to programming, most, if not all of the good programmers I've known, said that the basics of programming came easy to them. If you struggle a lot early on, then you're probably never going to be very successful at it. But I've known smart people who wouldn't make good programmers.

Programming came easy to me, so did flying. Again, I think it speaks towards aptitude more than anything else. Sure, you could say it's the same thing as intelligence, but it's more like selective intelligence than general IQ.

Dan Hulme

Yeah, I think the answerer's original thing of "you must have this IQ to be a good pilot" is too inflexible, but if he said "if you don't have this IQ, you will probably find it harder to be a good pilot," I could get behind that

Also, he doesn't actually say "at least", just 130-135. Does that mean that if your IQ is > 135, you shouldn't be a pilot? :-)

Bret Copeland

18:14

eh, I wouldn't even mention IQ. I would just say that if you find it a major struggle early on, that it's probably always going to be a struggle.

Dan Hulme

@BretCopeland that's both true and more helpful

casey

yea, if a student struggles in a slow VFR bird things are only going to get harder when you add clouds and a faster, more complex airplane

Farhan

@casey This isn't a general rule, is it?

With experience, don't you get more comfortable to once-harder tasks?

casey

@Farhan to a point. eventually ability comes into play

Farhan

76

A: Why are the cockpit controls of airplanes so complicated?

I would argue that the controls of an aircraft are not complicated, but rather that they are simply foreign to you. In the vast majority of cases, the various controls in the aircraft do one thing: Turn something on, or turn it off. They are quite simple actually, but what makes it appear co...

The analogy presented in this answer (...^) is really applicable.

casey

18:28

and airplanes in particular are machines that you need to be thinking a few seconds into the future about and not reacting to the present. That is probably one of the harder things to do I'd guess

@Farhan take all that complexity away and boil it down to stick and rudder skills. Aside from starting up the plane and shutting it off there is surprisingly little interaction with most of the cockpit buttons.

Dan Hulme

@Farhan Right, but if it takes you twice as long as normal to get comfortable with the basics, it will probably take twice as long as normal to get comfortable with the advanced stuff

casey

and the stuggle is that you need to be flying the airplane first, while doing all of that other stuff second.

I taught instrument students in multi engine airplanes and getting them up to speed, talking to ATC and briefing approaches was... fun

Dan Hulme

I had so much trouble talking to ATC last time

Farhan

@casey I see your point here.

casey

you really see it once you instruct. You can look at a student, what they are doing and what the airplane is doing and know that in 5 seconds they are going to do X because they are missing Y.

and that is expected when you are new, and most people improve and do fine. Some never get past that stage.

We had a student that we flew with once, then put in an FTD for a few sessions and then decided he was untrainable. That is the only instance I can recall of telling someone flying wasn't for them.

Dan Hulme

18:36

FTD?

casey

compounding that issue though was the place I instructed at expected a fast pace and results. He might have been ok in a traditional setting, but I doubt even that.

@DanHulme flight training device

Dan Hulme

what's that?

a kind of simulator?

I mean, an aircraft is a flight training device :-)

casey

FTD is generally a simulator with a partial or full cockpit that might have visuals and an instructor station.

no motion

generally

frasca.com/index.php/fixed-wing/portfolio

ours were from that company ^^

modeling a piper seminole

but not as fancy as the one shown on that page

Dan Hulme

OK, thanks

casey

they are mostly just training aids, but the they can (and ours was) certified to use a certain amount of hours toward the instrument and commercial ratings

Farhan

18:45

@casey Do you know how much do they cost? :)

casey

@Farhan no clue, but probably not cheap.

I know at least the garmin units in them are real

Farhan

I think the only thing cheap in aviation is a logbook.

casey

I dont recall if the cockpit stuff was actual piper stuff or just generic replica stuff

Farhan

Since we are on the topic of simulators, has anyone have experience at uFlySimulator?

casey

nope

Dan Hulme

18:51

@Farhan the radio delivers live ATC from Heathrow? that doesn't sound legal

casey

electromagnetic radiation just wants to be free

3

ratchet freak

19:32

as long as you don't transmit they can neveer tell if you are eavesdropping

Jay Carr

19:44

@casey @BretCopeland My only complaint about the answer is that it doesn't account for the fact that sometimes a certain teaching style doesn't fit a certain learning style. I admit, it's not a super common issue, but perhaps in the case of that question the student pilot and the instructor just didn't get along for whatever reason, and the instructing was poor.

I've had it happen to myself on a couple occassions. Where one instructor tried and failed and another, who took a different approach, I was able to learn the same skill from in a relatively quick amount of time.

Hence my comment on the question I guess...

Bret Copeland

yeah, certainly possible.

Jay Carr

Looking over the comments and answers...that questions sure has people being fairly passionate.

voretaq7

@DanHulme There are definitely people who can't fly - just like there are people who can't drive. Some people just can't keep up with processing events and reacting at any pace faster than a brisk walk. The idea that you can know after 1-2 flights if someone can be taught to safely operate an aircraft is just as bogus as expecting someone to drive smoothly after their first 2-3 hours behind the wheel though.

Jay Carr

@voretaq7 I enjoyed your manual transmission analogy, btw.

voretaq7

20:01

@JayCarr It's not directly comparable, but I think if you can handle a 5-speed stick you can probably learn to fly

Jay Carr

I keep grinding in third on my Matrix, should I give up the dream then? ;)

voretaq7

@JayCarr . . . or get the synchromesh gears checked, third shouldn't grind :)

Jay Carr

@voretaq7 -- Huh, hadn't thought of that. I thought maybe I just had a special talent for grinding into third. Syncro's...one of the mysteries of cars I haven't read into yet. Time to go cruise wikipedia...

Farhan

@JayCarr You do that to pass the guy who cut in front of you. :)

Jay Carr

Grind my car into theirs?

voretaq7

20:05

@JayCarr My new battle cry when people try to pass me in the rapidly-vanishing acceleration lane is "I will put you into the fucking trees!"

Jay Carr

Remind me to never drive on, around or near Long Island...

voretaq7

It only bothers me because there's nowhere for them to go: They drove ALL THE WAY TO THE END OF THE MERGE LANE and now it's gone. Just stop. Stop your car, turn the engine off, and hang your head in shame until rush hour is over.

Jay Carr

lol, that does remind me of one of my pet peeves. You know those big flashy arrows that show the lanes are merging for construction? And how most people merge early to be on the safe side. But a few people just feel this need to go all the way up until the merge point (kind of like you're saying). If you could put them in the trees as well, I'd appreciate it.

ratchet freak

they are using the zipper method

Merge (traffic)

In traffic engineering, the late merge or zipper method is a convention for merging traffic into a reduced number of lanes. Drivers in merging lanes are expected to use both lanes to advance to the lane reduction point and merge at that location, alternating turns. The late merge method contrasts with the early merge method. A related scheme is the dynamic late merge. The late merge method has not been found to increase throughput (throughput is the number of vehicles that pass through a point in a given period of time). However, it considerably reduces queue ("backup") length (because drivers...

Jay Carr

I just hate that they cut in line. I always find myself being a jerk and pulling in real close to the car in front so they can't use that space to get into my lane. Which, for me at least, is odd. In most merging situations I leave a lot of space for people to get in.

@ratchetfreak Or they could just "zip" earlier like the rest of us, at highway speed, so we don't have to stop for them at the point of merging... (In Missouri they often come to almost a complete stop, the merges are pretty steep fairly often.)

voretaq7

20:15

@JayCarr "And how most people merge early to be on the safe side" <-- No. No I do not know this thing at all. Here in New York you ride that lane until you hit that flashing arrow :)

ratchet freak

or don't overtake when they have picked their spot to merge into

voretaq7

@ratchetfreak The zipper method actually causes more congestion because drivers get to the end of the lane and come to a full stop, as opposed to merging in earlier when there's gaps and traffic is still moving at something resembling a normal speed.

ratchet freak

you shouldn't need to come to a full stop

voretaq7

Surprisingly though it's more advantageous to be the inconsiderate shitstain who makes everyone else stop: You race to the head of the lane being closed and force your way in.

ratchet freak

the guys at the other lane should leave room for you

Jay Carr

20:17

@ratchetfreak "shouldn't need" and "what happens" are different is what we're saying.

voretaq7

@ratchetfreak You do if you drive RIGHT UP TO THE END OF THE LANE in the middle of New York rush hour - if you continue moving at highway speed (or even 40mph Construction Zone speed) you will go through the back of the work truck.

Jay Carr

Though, to be fair @ratchetfreak, what you suggest might work in my state if merge sections were longer than, oh, 100ft.

voretaq7

You also can't zipper merge effectively when roads are operating at 120% of their designed capacity but that's what we get for using a road system designed in 1930.

ratchet freak

the zipper method doesn't mean drive to the end of the lane

it means leave room for the guy merging in

voretaq7

@ratchetfreak but that's what people do.

They race down the lane being closed until they reach the obstruction, stop, and then pull into the other lane, forcing every other vehicle on the road to stop for them.

(Nobody wants to implement my solution of leaving a state trooper with a shotgun at the back of each work crew: If you get to the cop he shoots your dumb ass.)

Jay Carr

20:20

For the record, I leave all hordes of space for people who are trying to merge properly...

@voretaq7 I like that plan, I like it a lot. I'd join the trooper just to get that job.

voretaq7

People should begin merging where that first sign that says "Lane closed 1/2 mile" is - there's all sorts of space for them back there and plenty of time to coordinate it so nobody has to stop

but people are inconsiderate.

ratchet freak

in belgium you have to leave space when the lane is closing

but not on on-ramps

voretaq7

@ratchetfreak in the United States the vehicle and traffic law is not enforced :)

Farhan

The worse thing about zipper lanes is when the guy in front of you starts driving in between two lanes, before reaching the end.

voretaq7

@Farhan Oh I do that. I'm a total asshole about it because otherwise the people behind me in the lane that ends will do that "race to the end, stop, and block traffic" thing.

You're also supposed to stop at the beginning of the zipper/merge lane (at least in NY) so you have enough room to accelerate to something resembling highway speed before you try to merge with the rest of traffic.

ratchet freak

20:23

in the merging lane you should be matching speed to the lane you are merging with

voretaq7

@ratchetfreak . . . around here we think that lane is for testing our 0-to-60 and 60-to-zero speeds to see if the car really performs the way they say it does.

ratchet freak

which it doesn't of course

voretaq7

(0-60, realize there's traffic backed up on the parkway and you can't just throw your car into the lane, slam on the brakes and stop right as the merge lane goes away. That's how it works right? <insert Picard double-facepalm here>)

Farhan

@voretaq7 I see people doing it (driving between two lanes) 1/2 mile before the zipper lane.

voretaq7

@Farhan . . . you don't recognize fellow pilots lining up for takeoff when you see them?

ratchet freak

20:26

@voretaq7 but you still don't have clearance to take off yet

voretaq7

(they drive down the middle of roads with double-yellow lines here. I want to shout at them "TAXIWAYS ONLY HAVE ONE STRIPE NOT TWO! YOU'RE IN YOUR CAR!")

Farhan

@voretaq7 That's different story.

I haven't seen as many stupid pilots as stupid drivers.

voretaq7

@ratchetfreak That's OK - with our traffic you're never going to get to rotate speed anyway :)

@Farhan I've never seen a pilot taxiing with a coffee in one hand and an electric razor in the other. (and it wold be marginally more acceptable for a pilot to be doing this because on the ground we can steer with our feet - idunno how they think it works in a car!)

The things I've seen on the roads man - it changes you. It makes you want to buy a surplus tank so you can just roll over shit and knock SUVs out of the way with the gun barrel :)

ratchet freak

or support more highway patrol enforcing the rules of the road

Farhan

The bottom line is because a driver's license only cost $25.

ratchet freak

20:29

and a car only a grand

Farhan

If it is like €400, then many more traffic laws will be followed.

voretaq7

@Farhan If there were mandatory driver training instead of just handing out licenses that might help too :)

but yes - part of the reason Germany has a lower fatality rate on the roads is because if you start tailgating or riding down the passing lane looking at the clouds they pull you over and ticket you

ratchet freak

how does getting a driver's license work in the US?

voretaq7

around here people just slam on the brakes when they see a cop (doesn't matter if they're going the speed limit - they have to slow down anyway)

Farhan

@voretaq7 I always wonder if cops turn on their flashing lights for fun?

Jay Carr

20:33

@ratchetfreak It's been changing a lot lately. It used to be that anyone with a license could teach you (usually your parents) and then you took a 30 question test, then an testor took you out for a driven exam. If all went well you could have a full license at 16.

voretaq7

Married With Children BEST MOMENTS - Drivers Licence TEST

►

Kinda like that.

Jay Carr

These days it's fairly different, but I don't have any kids near driving age so... I just know you have to take a class now and you don't get a full license until you've been driving for a year or so (at least in Missouri)

Which, all and all, is probably for the best.

Farhan

It's like this:
1) Turn 16 or 18 (depends on state)
2) Go to DMV
3) Pay a little money
4) Get license

voretaq7

in NY you can get a learner's permit at 16 - any licensed driver over 18 can accompany you (theoretically they're supposed to be teaching you). If you take a DMV-approved course you can get a full license at 17, otherwise it's 18.

Farhan

@voretaq7 I knew you'll link that :P

Bret Copeland

20:35

@JayCarr I'm not sure I realized you were from Missouri.

ratchet freak

@JayCarr it's nearly the same in belgium, 50 question test (you can only get 5 questions wrong) and then a practical doing the basic tricks: hill start, U-turn and parallel park

Jay Carr

@I am currently living there at least.

Here...yeah.

Bret Copeland

I have a step-sister who lives in Springfield. I'm from Kansas City.

Jay Carr

@ratchetfreak Nothing like Finland, eh?

ratchet freak

what's that like

voretaq7

20:36

@JayCarr Interestingly in NY if you're 18 you can get your permit on Monday, schedule your road test the next day, and if you don't run anyone over or go over a curb you'll have a license on Tuesday.

ratchet freak

if you fail the practical 2 times you have to take a course from a licensed instructor

Jay Carr

@BretCopeland Really? Wow, it's just a couple degrees of seperation at that point. Where does she work? (I'm wondering if we work at the same place, I work for a pretty large company.)

@voretaq7 It's because we Missourians are more civilized than you New Yorkers.

voretaq7

@ratchetfreak I think ours is a 50 question test and you need to score 70%. Same practical test (on real roads, not a closed course like some states).

@JayCarr I don't know what you're talking about. Now if you'll excuse me I need to find a subway elevator to take a leak in :)

Bret Copeland

@JayCarr I actually don't know. We're not particularly close. My step-mom works for an FBO in Kansas City.

ratchet freak

Belgium's practical is also on the open road

Jay Carr

20:38

@ratchetfreak Oh yeah, as @voretaq7 implies, it varies by state here. There's no national test (though your license is valid in all other states anyway...)

Bret Copeland

@voretaq7 you live way out on Long Island, there's no subways where you are. Find someplace else.

voretaq7

@ratchetfreak If you fail the road test twice in NY you have to re-take your permit test (and pay another application fee).

ratchet freak

though getting a 90% on the test is a bit harcher

Jay Carr

@BretCopeland Ah, well, if I ever use an FBO in Kansas City I'll ask if anyone knows a Bret Copeland (assuming that's your real name.)

voretaq7

@BretCopeland LIRR has elevators too. (and hey, it's cleaner than the men's room at the station!)

Bret Copeland

20:39

I don't think the Farmingdale station has an elevator.

voretaq7

@BretCopeland no, that's still one of the lines where we let people commit suicide by driving around the little poles :)

Jay Carr

@voretaq7 Have you found that different licensing bureaus run the test slightly differently in your state? When I got mine in Arizona I got to choose between a really tough bureau in downtown phoenix (parallel parking, hill starts, freeway driving, the lot) or I could go to Sun City, where all I had to do was drive around the block and park in a parking spot.

Bret Copeland

@voretaq7 it's called freedom.

voretaq7

@BretCopeland it's called "Quit delaying my train you selfish bastard, just use a toaster and a bathtub!"

(I used to be a LIRR commuter, you get jaded real fast)

@JayCarr they're all a little different because the road course is whatever's around the road test site

Bret Copeland

how often do you actually take LIRR? You still haven't come in to SE.

voretaq7

20:41

@BretCopeland I try to avoid the LIRR - last time I was on it someone wet themselves in the middle of the afternoon.

Jay Carr

LIRR? Long Island Rail Road or something?

voretaq7

(and I was in Manhattan last month but I was stuck in the datacenter for 9 hours both days. Stupid hardware failures)

Bret Copeland

yep

voretaq7

^ Official mascot.

Bret Copeland

When we upgraded our data center recently, we offered our devs and sysadmins the opportunity to claim one of the old servers. I claimed NY-WEB01.

It's supposed to be delivered soon. I can't imagine I'll ever turn it on.

voretaq7

20:43

@BretCopeland terget practice? :)

Bret Copeland

More just something to put in storage and be able to say I own a server which ran Stack Overflow for five years.

voretaq7

I have a stack of hard drives on my desk from the old CVS server we decommissioned. I'm waiting until our lead developer is back in the office from his vacation - I think I'll let him smash them :)

Bret Copeland

haha, I'm not sure what we would be angry at enough to want to smash.

Most of our smashing is done virtually.

voretaq7

@BretCopeland SQL cluster.

Nobody loves SQL.

Jay Carr

I can't wait until we can smash our old CVS server..... ><

Bret Copeland

20:46

It's actually Windows Cluster... which makes it even worse.

ratchet freak

@BretCopeland windows server doing SQL?

Bret Copeland

People still use cvs?

Jay Carr

Yup..... ><

Farhan

@BretCopeland We still use VSS. :D

voretaq7

@JayCarr It actually hadn't been a CVS server for about 5 years, we just never changed the label when we moved to git. we FINALLY migrated to the (now 2-year-old) replacement git server last month.

Jay Carr

20:47

We're trying to transition to git, but we've got all these legacy projects nobody can be bothered to update.

voretaq7

(to give you an idea how old this machine is: It's got a 32-bit CPU.)

Bret Copeland

@ratchetfreak SQL Server Availability Groups are built as a service on top of Windows Cluster.

voretaq7

@JayCarr cvs2git :)

Jay Carr

Could be worse, I could be working on our RPG projects on our as400 IBM servers running DB2 for i v5.1...

I missed out on that fun...

voretaq7

@BretCopeland You forgot "because Microsoft hates us, and wants us to be miserable."

Jay Carr

20:48

cvs2git is a thing? I might have to suggest it... None of my projects are on git though, so...

lol, on cvs

Pardon

voretaq7

@JayCarr mcs.anl.gov/~jacob/cvs2svn/cvs2git.html

Jay Carr

Cool, do they have anything for DB2 -> postgres?

Bret Copeland

Stack, in seven years, has gone svn->hg->git

voretaq7

if your repos are complex you really need to do cvs2svn and then svn2git, but probably 90% of repos can just do cvs2git

Jay Carr

Well, I'm just a lowly BSA, I'll throw it over to the server admins though. I'm not into telling them their jobs, but I don't mind sending a link on occassion.

Is that for the db2?

voretaq7

20:50

no, for the DBA.

Jay Carr

lol

You are confused, we have no such thing.

voretaq7

You can't kill DB2. You cannot kill that which is already dead.

Jay Carr

We do have some Application Server Admins who occassionally do some stuff with our servers though.

lol

Bret Copeland

Sounds like there are too many job titles where you work.

Jay Carr

I'm pretty sure they can be killed, because they've died a couple times...

Nah, we just have a lot of people.

voretaq7

20:52

DB2 should never come out from beneath the loving embrace of the OS/400 or Z/OS object management layer :)

@JayCarr . . . but no DBA.

Jay Carr

There's all of...15 job titles in my department of...well, lots.

Well, I'm not being completetly accuarte, we just hired a couple, like, a month ago.

But they aren't up to speed yet.

voretaq7

ah so they're still in the "What the actual FUCK?!" phase of their onboarding

Jay Carr

We are very very excited to have them though.

Yeah, I try to avoid them.

voretaq7

Soon they will graduate to the "THAT'S IT! I'M ENFORCING CONSTRAINTS!" phase.

Jay Carr

I wonder if they knew about the whole as400 running DB2 for i v.5.1 thing...

voretaq7

20:53

Try to be on vacation when that happens - the area between Devs and DBAs tends to turn into a nuclear war zone when that happens :)

@JayCarr 0_o

Jay Carr

Fortunately my little team doesn't involve any RPGers... We're all Java, we don't touch the DB2s all that much.

Yeah............

voretaq7

that's a 5?

Jay Carr

It is.

voretaq7

I haven't seen a 5 in so long... I'd forgotten what they looked like!

Jay Carr

We liked teh 90s, what can I say?

My favorite thing is that the DB hasn't been upgraded since the company started using Java :D

Anyway, these are problems that are (thank god) being rectified these days.

But when I first showed up 2 years ago....it was a scary time.

voretaq7

20:55

@JayCarr Start with "Bless me IBM, for I have sinned. It has been 25 years since my last confession. The AS/400 is still running in System 380 emulation mode. Also we're still calling it an AS/400."

Jay Carr

roflol....... oh goodness.

voretaq7

I started here.... about 7 years ago. They were still running Postgres 7.

Now we're running 9.4, with real live replication like a big-people database :)

@JayCarr $JOB[-1] had a client that was running an AS/400 in System 380 emulation mode. (Law Firm.)

Jay Carr

Yeah, lemme tell ya, you really don't want to hear about our current DB situation. RPG is all 5.1 'n stuff. Our Java DB approach is best described as *

Lordy

voretaq7

We used to call them the dingoes (because they were skittish about any minor change, and very suspicious of outsiders just like a pack of wild dogs - but once they accepted you into the pack they would trust you and share their lunchstuffs)

Jay Carr

lol

That sound vaguely familiar.

Bret Copeland

20:58

Our db situation is that we're "highly available" which means we get long outages due to the "high availability" system.

voretaq7

Just...uh...bang really loud on the senior partner's door before you go in. He liked to browse porn between meetings. In fact just let his secretary get his attention for you, it's safer for everyone - she had him on a short leash.

@BretCopeland That.... sounds about right :)

our DB failover is semi-automated: If the primary blows up you'll get a read-only secondary and I'll get a page telling me to determine if we should make an actual failover

(the answer to which is almost universally "NOPEnopeNOPEnopeNOPEnopeNOPEnope")

Bret Copeland

in theory, that's the way ours should work too.

Jay Carr

We're going with the cornecopia approach. We have postgres, mysql, Microsoft SQL, Oracle...and some other's I can't recall. Why have standards when we can have a bunch of DBs that have no way to communicate with each other, I mean really.

voretaq7

@BretCopeland in practice..... Windows Cluster.

Jay Carr

I'm thinking of adding in CouchDB, Mango and Neo4j on my next project, just for fun.

voretaq7

21:01

@JayCarr you have standards.... let's see the lowest common denominator is.... MySQL. So you support.... um.... hmmm...

about 7/8ths of SQL89 - there you go!

Bret Copeland

We did get to execute a command with FORCE_FAILOVER_ALLOW_DATA_LOSS recently. That was a good day.

Jay Carr

Well, we use Hibernate, so it' all abstracted away, right?

voretaq7

@BretCopeland Oh. Oh my. Oh my word....

Bret Copeland

It was a split-brain issue.

I'm surprised we still haven't blogged about it. It was a major outage.

voretaq7

. . . so the solution was an icepick lobotomy? :)

"Where are you going with that power drill?!"
"SQL Cluster."

Bret Copeland

21:04

well, luckily we were attempting to failover to the node which had been master when everything in the cluster went "fuck y'all, I'm out"

So the chances of data loss was essentially zero.

voretaq7

Postgres clustering is decidedly simpler - though writes are a terribly inconvenient bottleneck.

(I believe if we ever get multi-master replication that will be the milestone for Postgres 10)

Bret Copeland

How will multi-master work?

voretaq7

@BretCopeland . . . fairy dust and unicorns I assume. Multi-master is HARRRRRRRRD!

Jay Carr

What is multi mastering?

voretaq7

Right now postgres gives you master, read-only slaves (which can convert to read-write in a few seconds if you have to), and cascading slaves (basically slave-of-slave). The relationship for each of those can be either sync or async.

@JayCarr instead of master/slave you have peers (so you could theoretically execute write transactions on the slave, and they'd propagate to everyone else)

except now you've got state information being created on the slaves (like transaction IDs and sequence values), and right about here is where the developers start swearing and drinking heavily :)

Bret Copeland

21:11

it would have to be either a hashing/sharding strategy, eventually consistent, or hella blocking. I don't see another option.

voretaq7

Yeah, no matter what you do there has to be some blocking calls. Eventually Consistent won't fly because it breaks ACID

Jay Carr

...I just feel this sudden need to go buy our new DBAs a beer or something. I always forget how little I understand DBs until I listen to people who actually get them discussing them...

voretaq7

(Postgres: The database that brought you "Oh that's what you 'NoSQL' people want? OK, HERE, this is how a REAL database does it")

Jay Carr

NoSQL, another thing I'm not clear on. I keep hearing that in relationship to graph/node DBs, are they the same thing?

voretaq7

@JayCarr I…don't think the NoSQL people know :)

I've heard it used for a lot of different things

Jay Carr

21:16

I'm told Graph/node doesn't work well with SQL, but... I mean it's not a rejection of SQL out of spite, it's just that it doesn't work. Whereas the NoSQL movement, at times, feels a bit spiteful?

voretaq7

all they seem to be clear about is they don't like the table/row model - sometimes I wonder if I could repackage Postgres and just call them records & fields :)

Graph/Node works "OK" with SQL it's just modeling it in the table-and-row universe requires bridge tables and looks funny.

(and requires creative partial-indexing or the performance is shit)

Jay Carr

But wouldn't that defeat the whole purpose of node/graph? Ie., all those lovely relationships you can pull out of any one node?

I mean, if I wanted table structure, why not just, you know, use postgres and call it a day?

Bret Copeland

I wish postgres would allow named query parameters and not freak out about uppercase letters. If it could just do that, I would be incredibly happy with it.

voretaq7

21:40

@JayCarr You can pull that from SQL - select * from <bridge table> where source_nod = <id> :)

@BretCopeland Postgres allows named query parameters - not sure what you mean about freaking out over uppercase letters though

(table and column names should be lowercase, SQL keywords should be uppercase, and any other text should be in quotes! :-P)

((developers HATE working with me as their DBA.))

Bret Copeland

yeah, it freaks out about uppercase columns and tables unless you put them in quotes, which I find unacceptable. I'm a big fan of PascalCase.

voretaq7

@BretCopeland . . . You make Codd cry!

Bret Copeland

And since when does it support named query params?

I don't mean for procedures, I mean in a query.

like, select * from table where id = :id

or @id or ?id or whatever.

voretaq7

@BretCopeland idunno, you've been able to use them in bound statements for years in every language I've used - maybe they're hacking around the limitation

Bret Copeland

but where id = ? sucks when you start to get several params.

voretaq7

21:45

let's see if I can do a prepared statement with one on the command line

Bret Copeland

github.com/brianc/node-postgres/issues/268

from a couple years ago, but someone claiming psql doesn't support it.

voretaq7

hmm, yeah Postgres doesn't seem to support it natively in PREPRE, I guess Perl/Ruby/PHP just hack around it in the driver

Bret Copeland

yeah

voretaq7

well that's all the language I use :-P

Bret Copeland

I don't know anything about the protocol psql uses. I'm guessing it's just not built in.

It's built into TDS.

Although TDS is pretty hellish for other reasons.

voretaq7

21:49

@BretCopeland you'd have to store the parameter name/position mappings with the statement server-side

which isn't hellishly difficult to do but positional arguments are easier & apparently nobody's been motivated enough to do better yet :P

Bret Copeland

apparently

voretaq7

I may plant the seeds of discord the next time I have occasion to see a Postgres developer in person though

Bret Copeland

I bet it's something they've heard before.

voretaq7

@BretCopeland I blame the driver writers for hacking around it so the server people don't have to do better! Damn lousy developers! :P

Bret Copeland

But that's a lesser issue to me. god_i_hate_snake_case

voretaq7

21:51

@BretCopeland yeah I'm not a fan of snake case or pascal case (and camel case can die the death of 1000 sand fleas)

but if you need more than one _ you're probably naming it wrong

Bret Copeland

so... nocaseforyouoryoujusthatenames?

voretaq7

I we've got {first,last,middle}_name {start,end,create,....}_date, and maybe a few others

Bret Copeland

camel case is great for local variable names.

voretaq7

I don't think we have any x_y_z names

@BretCopeland aw just use a freakin' class, we're not writing BASIC anymore :)

Bret Copeland

one or two words is probably most common, but certainly not enough for all uses.

voretaq7

21:54

usually when I see more than 2 words I also see a normalization problem

Bret Copeland

lol, no local variables? Time to send you off to the loony bin.

voretaq7

wednesday_sales_totals <glare angrily>

@BretCopeland you can use functions to do your scoping too if you won't need that data later :)

but if I catch anyone here doing myfunctionTotalKittensEaten I'll probably garrote them with a mouse cord

Bret Copeland

yeah, those pesky Lisp people have been trying to make that argument for decades.

voretaq7

. . . actually I'll probably leave them tied up in the lead developer's office with a post-it note stuck to them explaining their sin :)

@BretCopeland Anything worth doing can be done in Common Lisp. Including driving your enemies insane.

Jay Carr

@voretaq7 I'm guessing you are not a fan of Cocoa then, they're method names are...long, to say the least.

voretaq7

21:58

@JayCarr Apple should stick to designing user interfaces and quit trying to take up all of my disk space with function names in header files :)

Jay Carr

For examples: github.com/Quotation/LongestCocoa

lolz

voretaq7

(Apple SDKs are either excellent or astoundingly terrible)

Jay Carr

I personally like longer names. But then again I've been dealing with a lot of legacy code that involves a lot of methods with variables like "a" and "lr" and "n"

Can you bring that shotgun up again? Because that's how those variable names make me feel.

Bret Copeland

I like names which express intent. Sometimes that requires a longer name, but I prefer that over ambiguity.

Jay Carr

+1

Bret Copeland

22:03

I seriously need to go flying. My currency lapsed about a month ago.

Jay Carr

Is it hard to get your currency back now?

I'm not familiar with how that works...

voretaq7

@BretCopeland delete_all_the_files_i_need_but_leave_the_bosses_porn_stash_alone() :)

Bret Copeland

eh, not really. If I had my own plane I could actually just go up by myself and regain it with three takeoffs and landings. Since I rent, I have to go up with an instructor and do that.

@voretaq7 I mean, if that is what the function does...

voretaq7

@BretCopeland I've not been flying nearly enough - I went up last weekend because I looked at my logbook and said "Oh shit, I'm out of currency as of tomorrow!" :-/

@BretCopeland well Microsoft calls it the "cleanup utility" or sommat :-)

Shalvenay

23:25

@BretCopeland and @voretaq7 -- re the whole DBA thing -- I'm one of those devs who actually rather likes having a dedicated set-theory/data-management package at his disposal (because that's what a RDBMS is)

but the biggest problem with constraints is that real data can get very nasty very fast -- you're often limited in what constraints you can put on something by wacko corner case situations

and it's only worse when you have a large pile of existing data imported from a previous system

Bret Copeland

I think most developers actually do want an rdbms, but the nosql world is very tempting because of how quick it is to get going and the promise of scale.

Shalvenay

yeah. my answer to the startup issues is twofold -- 1) SQLite is fantastic for apps where you don't need to be client-server at the database layer

and 2) when you do need a client-server RDBMS -- PostgreSQL is cheap, good at ACID, and supports SQL well

Shalvenay

23:42

@BretCopeland -- the other thing is that app developers and DB developers (there should be a distinction between a DB developer and a DB operator!) wind up separated by a very thick firewall in most orgs -- having to put tickets in to make dev DB schema changes is rather annoying

Bret Copeland

I guess so. At Stack Exchange, we just keep our migrations in the repository, and they run automatically when we deploy the app, so it's not particularly difficult to make the changes you need.

Shalvenay

yeah -- schema-in-the-repo is the best bet, but it's also hard to implement for an existing app that isn't doing that

Bret Copeland

I actually don't think it's that hard to implement. You just have to start doing it. Our migrator is dead simple. It just looks for .sql files in a directory, and checks to see which ones haven't been run, and runs the ones which haven't (each inside their own transaction). It adds a row to the Migrations table for each file with the file name and md5 so that it will know not to run it again.

There's no such thing as a backwards migration, only forward. If you want to revert something, you write some sql to do that.

Shalvenay

@BretCopeland -- well, I'm talking about having the schema in the repo, not just a string o' deltas -- although I wonder if would be possible to write a diff tool that took the SQL for two schemas and then spit out the apropos statements to get from A to B

Bret Copeland

23:57

Having a schema file is not particularly maintainable, in my opinion.

If you have an existing schema, sure maybe you copy the current state of it into the repo, and then just execute the migrations on top of that to get up to date. If the schema doesn't already exist, then just start with an empty db and build it up with migrations.

anyway, I'm out for a while...

Transcript for