(this is long.)

2/13/2012

[09:51] <ursamajor> it helps that we have others around here who have been marinating in this stuff for years if not decades to help supplement
[09:52] <ursamajor> but it takes awhile to *get* to that deep understanding, and there's a whole bunch of intermediate understandings to go through in between.

[09:52] <ursamajor> plus it's often easier to communicate to a layman's understanding while you're not too far past layman yourself :)
[09:53] <shadowspar> believe me, when you feel like you have a good grip, on, say, perl, there's nothing like going to a conference with all the Lords Of The Perl Planet to make you feel like a dumbass again.
[09:53] <denise> yeah, i never know what i know until i try to teach it
[09:53] <shadowspar> I speak from personal experience.
[09:53] <denise> (or to brainstorm something with it)

[09:53] <denise> like, i keep startling myself because i make what i think are stupid suggestions to mark and he says "hey, that'd work!"
[09:54] <ursamajor> because you remember most easily and identify most easily then. I'm not saying that we don't have good teaching and information provided by people with more experience - but that teaching and learning in all levels are important :)
[09:54] <denise> HELLO IMPOSTOR SYNDROME


[09:55] <ursamajor> rah: hee *G* i so identify with this. To a probably unhealthy point, because honestly I have no idea what my actual professional abilities and strengths are
[09:55] <randomling> lol yeah I don't feel like a real programmer. but that's because I'm not one. apparently I know next to nothing about what's going on behind the scenes :P and I've only been doing this for 6 months and I *still* can't figure out how to test my stupid patch
[09:56] <ursamajor> I mean, i'm like, "Um. I like reading and finding out about things a lot and organizing them so that when it comes up in conversation I have relevant info and feel like less of a dumbass!"
[09:56] <stormerider> denise: just remember, the most non-technical member of SG-1 (Jack) was often the one suggesting life-saving solutions to problems :) not that I'm saying you're non-technical (far from it!), but my point is sometimes it just takes taking a step back and trying a different logical approach rather than getting bogged down in the details.


[10:02] <denise> randomling: with testing, you want to generally test at least one positive case (does it do what you want it to do) and as many negative cases as you can think of (does it not do what you don't want it to do)
[10:02] <denise> but you don't have to get them all!

[10:03] <mark> testing gives you confidence in your code. but that confidence is asymptotic. it approaches 100% but never reaches it.
[10:03] <denise> generally you just sit and go, okay, can i think of how people are going to try to use this? and then do what you think people will do to it
[10:03] <denise> exactly.
[10:03] <denise> you can never *guarantee* something is bugfree. at most you can say with confidence that you've tested lots of use cases.
[10:03] <denise> when you upload a patch, just include information on what use cases you've tested
[10:04] * randomling nods
[10:04] <denise> and if the committer thinks of anything major that should be tested that you haven't, they'll test it then
[10:04] <randomling> okay that feels a lot better!
[10:04] <mark> (the above logic is why writing 'some' tests is better than 'none', and why you shouldn't worry about 'WRITE ALL THE TESTS')
[10:04] <randomling> I was feeling all "but it's ALL MY RESPONSIBLITY" and panicking
[10:04] <denise> and if you *both* miss something and a bug gets onto the site: OH WELL
[10:04] <denise> IT HAPPENS

[10:04] <kaberett> also no no code should ever be entirely one person's responsibility
[10:04] <kaberett> if it is, the org's fucked up
[10:04] <mark> yes
[10:04] <denise> yes
[10:05] <ursamajor> yes to the yes.
[10:05] <denise> bugs are the respsonsiblity of the coder *and* the committer
[10:05] <denise> *and* anybody who read the code afterwards adn said "hrm, I wonder if that's a bug" and didn't say anything
[10:05] <kaberett> there's a *reason* that quite a lot of places use agile programming, i.e. every bit of code gets two eyes on it talking it over before it's committed
[10:05] <kaberett> yep!
[10:05] <denise> and anyone who reviewed the code


[10:06] <denise> anyway, you haven't arrived until you've made at least one mistake that brings the site to its knees ;)

[10:06] <mark> denise: dammit, I need to stop arriving.

[10:07] <denise> i think i got the pony last time i was writing code

[10:07] <mark> denise: I neglected to mention that the other day I oopsed a database and had to rebuild it. while debugging the slow page load thing, I installed something, that uninstalled MySQL :P
[10:07] <denise> ("the pony" = "the commit-and-ditch pony", aka the "you broke the site award", given equally and goodnaturedly to whomever wrote the code that broke shit and whomever reviewed and committed it)

[10:08] <idonotlikepeas> At work, I actually wrote code like seven years ago that caused a system to lock up last year due to weird database locking issues. That kind of thing just happens sometimes.
[10:08] <denise> I EARN THE PONY A LOT
[10:08] <kaberett> heh, at my ex's work it's a monkey :-)
[10:08] <denise> mark: you ... uninstalled MySQL?
[10:08] <denise> that's a good one :)
[10:08] <allen> i've earned the pony a few times, too.
[10:08] <idonotlikepeas> Yeah, that's pretty impressive.

[10:09] <denise> my most embarrassing sysadmin disaster was just your typical "rm -rf *.txt" vs "rm -rf * .txt" disaster
[10:09] <kaberett> randomling: also, if at some point you fancy holding my hand through learning to do a tiny fix I would love you forever <3

[10:10] <mark> to be fair, it was because we user Percona MySQL which is manually installed. so when I installed the webserver packages on db04, it includes mysql-client. which detected a conflict with percona-mysql-server and uninstalled it. I was not fast enough on the Ctrl+C

[10:11] <mark> but still, that meant the database was down for a while :P

[10:10] <denise> hell, writee a bug walkthrough for dw-dev-training
[10:11] <mark> but still, that meant the database was down for a while :P
[10:11] <randomling> hm, I might do that!
[10:11] <denise> i love it when people do those
[10:11] <denise> mark: you didn't mention that to me at the time. *G*
[10:11] <randomling> not sure I can remember everything I did
[10:11] <allen> ursamajor: i think now it will, and directories like '.config'.
[10:11] <randomling> also, a lot of it will be, "so then Sophie told me how to...."
[10:11] <ursamajor> that's okay, just start writing one for the bug you're doing now, then :)
[10:11] <denise> anyway, everybody makes mistakes.
[10:12] <sporkyrat> randomling: I think that's how a lot of tutorials go!
[10:12] <ursamajor> and yeah, randomling, that's part of the process :)
[10:12] <allen> (the directories are from the -r option - recursive)
[10:12] <denise> like, to the point where, if nothing breaks after a code push we get NERVOUS
[10:12] <kaberett> because clearly something major is going to go wrong very, very soon :D
[10:12] <idonotlikepeas> Yeah, everyone does that.
[10:12] <mark> allen: I have grown the habit of using '.??*' when I want dot files
[10:12] <ursamajor> "i did x! got stuck! asked for help. got help from soph! did x+! got stuck! soph helped again! got to x++! soph was asleep so geekosaur helped!"
[10:12] <ursamajor> etc.
[10:12] <denise> "fuck. nothing's wrong. this means we're going to have some lurking shambling bug that's going to turn out to explode in two weeks."
[10:12] <idonotlikepeas> Bug-free projects don't exist. It just means you haven't found it yet.
[10:12] <kaberett> randomling: also, d'you know what
[10:12] <kaberett> writing "and then I got stuck and asked Sophie for help"
[10:13] <ursamajor> but if that's how the coding process goes, then it's a good idea to portray it that way to newbs :)
[10:13] <kaberett> is gonna be SO HELPFUL to people like me
[10:13] <kaberett> who would othewrise go "CRI CRI CRI I SAID I WOULD DO THIS MUST FACE THE GAZEBO ALONE NO ASKING FOR HELP EVAR"
[10:13] <ursamajor> it helps illustrate to newbs that they don't have to KNOW ALL THE CODE right away
[10:13] <kaberett> which
[10:13] <kaberett> obvs ridic?
[10:13] <ursamajor> or even remember all the code they knew last week!
[10:13] <idonotlikepeas> Oh, a little programming joke. Some principles: 1. Every program can be optimized to reduce it by one line. 2. Every program contains at least one bug. Conclusion: Every program can be reduced to one line that doesn't work.
[10:13] * ursamajor giggles at peas.

[10:14] <denise> oh god if i had to do it all alone without asking for help i'd never get ANYTHING done
[10:14] <kaberett> randomling: well then :-p
[10:14] <denise> i mean, ask mark
[10:14] <denise> how often do i yell "mark, i'm stuck"
[10:14] <sporkyrat> ursamajor: YES.
[10:14] <sporkyrat> Oh my gosh, can we like, make little graphic medals for geekosaur and sophie?
[10:14] <denise> like the time he helped me troubleshoot my broken email for like three hours, only for me to discover the reason it wasn't working was because i'd let the domain registration lapse
[10:15] <denise> (he was very good. he hasn't teased me about it at all.)

[10:18] <denise> also, going back to the whole "people make mistakes" thing
[10:18] <denise> i can guarantee you
[10:19] <denise> no matter what
[10:19] <denise> no matter how bad your code is
[10:19] <denise> no matter how many bugs it contains
[10:19] <denise> it will *never* be worse than some of the things bradfitz got up to in the early days ;)
[10:19] <randomling> hee
[10:19] * randomling hits up the wiki to remember how to grep ffs
[10:19] <ursamajor> oh god grep
[10:20] <denise> "hey, so, i accidentally fucked up while trying to upgrade this one db cluster. we'll play back the backups later this week, okay?"
[10:20] <denise> six months later
[10:20] <ursamajor> i use it just infrequently enough that i can never remember the syntax
[10:20] <denise> "oh yeah, we were going to play the backups weren't we"
[10:20] <idonotlikepeas> oh god Brad
[10:20] * sporkyrat blesses brad's heart, the dear soul.
[10:22] <denise> and let's not even *think* about the UTF-8 conversion

[10:23] <denise> oh god i'm looking at 2002-era lj news posts and they're making me twitch, heh
[10:23] <kaberett> oh?

[10:24] <denise> kaberett: in 2002 i was sort of slowly taking over lj support by dint of simply showing up and doing shit
[10:24] <denise> so like
[10:24] <kaberett> lol, right
[10:24] <denise> brad would code something, push it live without testing
[10:24] <denise> support would get eight billion bug reports
[10:24] <denise> i would nag brad about it
[10:24] <denise> brad would ignore it
[10:25] <kaberett> ...
[10:25] <denise> i'd go to jesse or sherm or somebody and go "hey, nag brad about this"
[10:25] <denise> brad would ignore it
[10:25] <foxfirefey> Oh dear
[10:25] <denise> three months later somebody would contact brad directly about the prblem
[10:25] <denise> brad would be all "why didn't anybody tell me about this?"