The Whiteboard

Hexadecimal is a human representation, not a computer one. That's one of your invalid assumptions. It's a convenient way to express bytes in four-bit heximal places, that's all. The reason we don't use two digits per heximal place is because that would be really, really confusing. — Robert Harvey 36 secs ago

3:47 PM

What do you guys think of "find me a license" questions?

@RobertHarvey "find me a CV"

Q: Choosing a license for software

1

I am developing a javascript library that I want to release as open-source under something similar to the MIT license. The issue that I am having is that under this license others can sell my library standalone without modification. This is something I would like to protect. Is there a licens...

licensing mit-license

So quiet in here today. Did everyone get their fill of monkey snot yesterday?

user41796

4:08 PM

@RobertHarvey good fodder for VTC. It's just an alternative form of the community being a research librarian.

4:43 PM

@RobertHarvey I've got code to meddle about

woohoo coding

►

4:56 PM

@NickWilde: Demonstrating absurdity by being absurd: "What software license should I use for programming on a boat out at sea with my pet cat Schrödinger?" — Robert Harvey 45 secs ago

Ok, I wrote a few hundred lines, I'm good now until I get another review to move forward... One of the most annoying and tricky parts of this job is moving forward on things the perfect amount such that when you get review, if it's good - it doesn't take long and let's you continue moving forward, and if it's bad - it doesn't take long to correct. Move too much code and get a bad review, it'll take ages to correct what people don't like, too little and you're a reviewing too oft, wasting time

Getting the code right is the easiest part of this job...

@RobertHarvey The "Uncertain Seas" license, obviously. Voting to re-open in light of such an obvious objective answer.

Boat programming is always a good retort for anyone that thinks that any question in a given topic area should be allowed.

@JimmyHoffa We don't do code reviews here, unless the code is considered critical in some way, like real-time systems code or security-sensitive code. Not sure if that's a good or bad thing.

We don't code-review utilities, for example.

I just implemented a code-review process for our code base because we work with a couple of outside resources. So far it's proving to be more time spent and not much gain.

@RobertHarvey utilities sure, no need there, but I'm writing out code for a production feature... I've worked numerous places that "don't do code review" - I always find a way of getting review from someone anyway. I hate pushing code onto my team without someone on my team agreeing it's reasonable (unless it's a few lines - this is multiple classes along with changes to service initialization in global asax, relatively systemic changes..)

5:05 PM

@JimmyHoffa Yeah, we do standard "sanity check" type stuff. Usually it's just an IM or email saying "Is this a really stupid way to do this?"

"It's cool that I removed all the assignment statements, right?"

@psr "I replaced all of our for loops with monads and 7-level nested if-then statements. I can release now, right?"

5:24 PM

@Dr.Ibb ...the first part of that would be totally acceptable, preferable even. :) For loops? Pfah, such a terrible idea those were.

@JimmyHoffa I tend to use .map() a lot as a replacement for for loops, but I'm more familiar with it than the rest of my team so it's not idiomatic to the way things are written in most of our code.

Or list comprehensions since we work in Python and CoffeeScript a lot.

5:46 PM

@Dr.Ibb I rewrote that incomprehensible parser combinator thing into a nice clean case statement. I named each state after a constellation so they'll be easy to remember.

oh dear

I hope you documented them

YES. Self documenting.

constelations? at least have a list somewhere where the link is explained

user41796

@psr Too bad Hades isn't a constellation

Lists of constellations

The following lists of constellations are available: * 88 modern constellations - a list of the current constellations. * Former constellations - a list of former constellations. * Chinese constellations * List of Nakshatras - sectors along the moon's ecliptic * Asterism (astronomy)

Done!

Q: What are the benefits of using a 'decorator factory' that decorates objects?

6:16 PM

0

In a project I decided to implement the Decorator pattern. I have a class Thing with methodA(), and a class AbstractDecorator that inherits from Thing and that all decorators inherit from: ConcreteDecorator1, ConcreteDecorator2, etc. All of them of course override methodA() to add some functiona...

design-patterns object-oriented decorator factory

user55340

sigh

Q: How is the Decorator Pattern actually used in practice?

"I did this but actually I don't know why. Does this have benefits as opposed to having the client instantiate decorators directly?"

user55340

-4

I understand completely how to implement the Decorator pattern, and I also understand what it's intent is. The Decorator is used in one of two cases: As an alternative to subclassing - when there are multiple characteristics that an object can have, one could use inheritance in order to create ...

design-patterns object-oriented decorator

Decorator factories just industrialize the commercial decoration concept, it makes things pretty, en mass and utilizing all modern factory-work efficiency techniques. — Jimmy Hoffa 6 secs ago

Asking what might have been a good reason for doing what you did is just odd.

user55340

6:19 PM

May 16 at 21:39, by MichaelT

@gnat about the Decorators guy... he's young. He's trying to understand all of modern software development in one pass, without working from the fundamentals... trying to skip ahead to the "build big software" and "neat games". He reads something, doesn't understand it, repeats it and asks for it to be explained in concepts he understands. Thus the nature of the dups.

@JimmyHoffa A.K.A. "Put a bird on it".

@psr pretty strange. Who needs decorators so much as to create an abstract decorator factory that you might have a creational pattern for creating numerous exotic decorators. That just smells something awful to me...

google.com/…

Q: On large communities decaying over time, being nice or mean, and Stack Overflow

So I suspect most/all of you have seen this by now, but I suppose this is why I like you all so much more than SO:

157

Wandering about the Internet, I stumbled across why online communities decay over time. This is a rather good article and should be read in conjunction with a group is its own worst enemy. I strongly advise people to read both of these articles — they will give appropriate background for some of...

discussion

user55340

Why would you need a factory rather than just pass the object to be decorated into the constructor of the decorator? It sounds like you are over engineering for the sake of over engineering and then wondering why it doesn't have the benefits you seem to perceive that are there. Write the code first, then when you recognize you need a given Pattern to solve a specific problem, then you use it. Don't stick a Decorator or Factory (or heaven forbid a DecoratorFactory) in just because its a Pattern. — MichaelT 7 secs ago

user55340

6:23 PM

@durron597 I think I missed that one... ;-)

@durron597 Don't know why. We're very aggressive about closing off-topic questions here. It is a bit more of a small-town feel, though.

@MichaelT AbstractDecoratorFactoryFactory

user55340

@RobertHarvey there's more interaction here. I tend to always recognize half the names in the answers... I know how they write. I'll see them in answers again in other questions. Even that little bit is interaction.

user55340

I really believe that SE would benefit from a 4th place (main, meta, chat, ???) where people can interact with each other more. A social networking type place.

Why does chat not suffice? Do you want the structure that the Q&A format provides, without the restrictions?

user55340

Chat is too 'real time'... its closer to irc than google+

6:26 PM

@RobertHarvey I think programmers does a good job of keeping itself small, if you don't behave you get run out very quickly

user55340

@durron597 (see A group is its own worst enemy linked there...)

the only people that stick around are those that the community wants to have, and most of the questions that get burninated mostly send people to SO if they aren't already banned there

user55340

While SO/SE is the anti-forum, the forum exists because it fills certain needs. Its a good place for interaction (a poor place for questions). But sometimes, interaction is precisely what is needed.

then make your own forum

There's this thing called promotion.

user55340

6:29 PM

@ratchetfreak some do - sopython.com

@MichaelT I know YOU didn't miss that one, but I wanted to talk about it because it's interesting, so of course I come here

user55340

And while those can work, it lacks the integration with SE (oath, reputation, a handle you can invest in)

Promotion.

Or popularity, whatever you want to call it.

problem with a normal forum is that it is hard to create "privileges" that users can look forward to unlock

user55340

@MichaelT I agree that patterns need to be used only when there's a real need. Thing is I'm a hobbyist and my projects are currently too small to actually need too many of these things (1700 LOC this project). So I'm using them to practice these techniques, even though I don't actually need them in my current projects. — Prog 1 min ago

user55340

6:35 PM

So you are wondering why, when you force yourself to use a pattern that solves a particular problem that it isn't useful when you are not trying to solve that problem? — MichaelT 16 secs ago

user55340

@MichaelT I was wondering what are the 'text book' benefits of using a static factory to instantiate decorators, since I saw other guys doing this. — Prog 4 mins ago

user55340

You are partaking in cargo cult programming. Patterns are not meant to be building blocks. They are solutions to particular problems. Their use is when you recognize you have a given problem then you can say "a ha! this can be solved by a FlyWeight" and then you use it. Starting from the Patterns and trying to write software around them before the problem is needed leads to inflexible and contrived designs that resemble things from Architecture Astronauts. — MichaelT 20 secs ago

alrighty folks, I'm going to be interviewing for a software dev position next week (internal) focusing on test/TDD types of things. what sorts of questions should I ask or high level things to review?

Most of you know my background pretty well at this point, for better or worse ;)

user55340

"Can I use VBUnit for this code?"

6:50 PM

I will be more than happy to get away from anything with "VB" in it if I can...

user55340

Ok... more serisouly... ask about the process of getting the requirements. What existing test automation they are doing...

from what it sounds like, the team is relatively new - so I don't think there is much

user55340

The continuous integration environment. What is the process in the source code control for checking in a bug fix?

@MichaelT oooh good one

From looking at the hiring manager calendar he has had a lot of people start in the past few months, and my convo indicates it's an entirely new team

user55340

There are some where its 'just check it into the mainline' others are 'branch a hot fix, validate, merge to mainline'

user55340

6:52 PM

Ask about the process of gathering requirements. TDD really starts with requirements, but RDD sounds so lame because in theory, everything is RDD... one would hope.

@psr pretty sure they have a fulltime cake decorator on site to design whatever cakes you want for lunch, for free (ok maybe not but that'd be a cool perk of working at a software company)

@MichaelT That's also really, really good (thanks, I'm creating a google doc with my questions)

they are pretty heavily into Agile

user55340

What flavor of Agile? Scrum? Extreme? Kanban? Scrumban?

that I don't know

user55340

That's a question to ask them.

user55340

How do they prioritize tasks?

6:55 PM

Oh, ha. Thought you were asking me a clarification :)

user55340

You might look at you track ( jetbrains.com/youtrack ) though, saying that I probably owe @Ampt another nickel because thats from Jetbrains.

user55340

Actually, given the rick-coding I did of him yesterday, I probably owe him a strong drink...

Asking what flavor of Agile is often useless though. I've recently talked to someone who said his last job did extreme - then later said the core of extreme is a testability matrix. The label and the practice seem in general to have diverged almost completely, so if you care you need to ask details of what they actually do.

user55340

@psr it depends on what you are using the information for. Accept that you're going to get 'faulty' information. But there's also the "I am asking questions about your process and how you do things and have an interest in doing things right according to your existing workflows"

Blog - Stack Exchange: Podcast #59 – He’s One Of Those Science-ists

Yup. I did say "if you care".

Stack Exchange

7:05 PM

Abby T. Miller on May 29, 2014

Welcome to the 59th running of the Stack Exchange podcast, brought to you by Nutella! Your hosts Joel Spolsky, David Fullerton, and Jay Hanlon are joined this week by special guests Josh Heyer (aka Shog9) and Robert Cartaino (aka Robert Cartaino) of the Stack Exchange Community Growth team.

We’ve got a busy podcast scheduled, so let’s get down to business, starting with New Features with Uncle David.

We are revamping the user profile, and you can check it out and give us feedback (and please do!). …

A: Should delete votes be limited like close votes?

7:26 PM

@MichaelT Yeah I like this idea too, I'd be asking more as a general question hoping to get more detailed info for followup

user55340

7:41 PM

@GlenH7 @gnat apparently, don't have to down vote all the answers in a question to reduce the popularity adjustment for delete votes on questions:

user55340

31

We had discussed this before, but deferred until there was a problem - it seems we now have a problem :) The new rules: 10k rep users get 5 deletion votes per day on questions they don't own - deletion rules on questions one does own are still in effect popular questions require more deletion ...

user55340

(notice latest edit: meta.stackexchange.com/posts/51071/revisions - and its been that way for awhile, just not documented as being so)

Thomas Owens

DAMMIT, @ODED!

Bicoid

Is it better to major in computer science with virtually no projects of your own, or major in math with a minor in computer science, and have a couple projects and do research with faculty, in terms of getting the same programming job?

Vedant Chandra

Considering the job situation now, you're probably better off with the option giving more experience

I'm not sure about that getting you the same job though

8:10 PM

@Bicoid Anecdotally, I'm a programmer without a CS degree who got hired because I did my own side projects and learned a bunch of stuff on my own time, coupled with experience in IT and business.

There are a lot of jobs I don't even consider because they require a CS degree, but about 40% of the postings in my area say something like "CS Degree (or equivalent experience)" OR they don't even mention having a CS degree and focus solely on experience and skills.

Oded

@ThomasOwens you rang?

Q: If Scheme is untyped, how can it have numbers and lists?

Programmers or SO?

0

Scheme is said to be just an extension of the Untyped Lambda Calculus (correct me if I am wrong). If that is the case, how can it have Lists and Numbers? Those, to me, look like 2 base types. So I'd say Racket is actually an extension of the Simply Typed Lambda Calculus. No? What is actually th...

types lambda scheme racket

@RobertHarvey I'd say SO given that it's talking about a specific implementation.

Well, it's not really a "How do I do something with this programming tool" question.

True.

Bicoid

8:15 PM

@Dr.Ibb Oh, I didn't realize you actually have to have a CS degree to be considered for some jobs...that is interesting. thanks

@RobertHarvey P.SE

@Bicoid Some places won't even consider you unless they see the credential on your resume. I don't apply to those places because I don't think I'd want to work somewhere so rigid.

n/m, SO

The title is a conceptual question, the actual Q is a concrete question about racket's type system implementation

Thomas Owens

@Oded You closed a question way too fast and beat me to it. :(

no nevermind

his comment betrayed that he wants a conceptual question answered

@RobertHarvey that question is poorly written. Needs edit before migration maybe...

8:18 PM

@JimmyHoffa Definitely. It's unclear what he's asking, in my opinion.

Or he's looking for knowledge validation.

A: Sorting two arrays simultaneously

Does "untyped" in this context really mean dynamically typed?

Vedant Chandra

0

You should not have two parallel arrays. Instead, you should have a single array of WestWingCharacter objects, where each object would have a field name and a field number. Sorting this array by number of by name would then be a piece of cake: Collections.sort(characters, new Comparator<WestWin...

How much more obvious do I need to make it that it isn't a homework question

Yet the downvotes

@RobertHarvey editted

@RobertHarvey No. Untyped means typeless as he refers to like untyped lambda calculus

Then the answer is that scheme is a superset of ULC.

untyped lambda calculus has function application and composition, not more. The elements have no concept of type

8:22 PM

Works for me.

@RobertHarvey I think he's asking though - is that more akin to Simply Typed, or maybe even Lambda-mu

@RobertHarvey also, untyped lambda calculus doesn't really make a subset of a typeful language - the fact that a typeful language implements constraints that disallow functionality available to untyped lambda calculus means there are things untyped lambda calculus has which a typed one does not - so it's not a superset/subset relationship.

I presume you're writing a comprehensive answer. :)

@RobertHarvey pfah. I don't know Racket's type system.

I've not worked with it

It's Scheme's type system.

Unless you're talking about Typed Racket, which is a whole nother beast.

@RobertHarvey which differs from LISPs but I never took the time to figure out in what way. Never mattered to me, the lambda calculi type systems always seemed more relevant

8:27 PM

Except that nobody writes programs in Le Lambda Calculi.

lambda calculi speaks to an ideal, it's more useful to know how to operate within them than any given LISP because you can write any given LISP as a given lambda calculi

@RobertHarvey sure they do, I do

Using some language built on top of it.

Don't go all pedantic on me.

@RobertHarvey I think that's impossible around here.

Funny how Dokkat is a high-rep user here, but decided to ask his question on Stack Overflow instead.

Well, reasonably high-rep anyway. With an account in good standing and 32 questions asked.

8:45 PM

@RobertHarvey JavaScript makes a great implementation of untyped lambda calculus. Scheme and LISP are implementations of lambda calculi with a type system I genuinely don't know. I never looked close at their type systems because I don't care. I'm glad to just write lambda calculus in any LISP I play with as typelessly as the language will let me.

The question is about lists and numbers. You have those in Lisp because the language specification says you have them.

@RobertHarvey does it? Are you certain? I'm not. I literally don't know about the languages specced type system because I treat them like, for lack of a better term, atoms from Erlang. Those are the closest thing to lambda calculi terms in a programming language I can think of...

Well, the premise of the question is "how can I have lists and numbers if TLC is untyped?"

Lame question, I know, but there it is.

If it weren't Dokkat, I'd swear we were being Goma-trolled.

@RobertHarvey look at his questions, he has a habit of questions that are trollish and even some really bad ones.

that said, this question is decent

@RobertHarvey the question you phrase there, makes sense to me as a question (s/TLC/ULC/f)

One of these days you'll explain/these/path/things to me.

8:51 PM

and the answer is what the guy is deducing: LISP/Scheme are not actually untyped, so perhaps their type system is more similar to simply typed lambda calculus than untyped

@RobertHarvey you don't know what s/something/another/f is?

>:(

2

@RobertHarvey open vim, type s/something to find/something to replace it with/f and it will find "something to find" and replace the first one with "something to replace it with"

Yeah, I don't use vim. or vi.

g is global (I believe? I was never actually a vi user of any note)

@RobertHarvey me either, but I've at least picked that up :P

@JimmyHoffa You pick up lots of things. That's not necessarily a ringing endorsement. Some things you pick up you have to go to the doctor for. :)

9:02 PM

ULC is untyped because it doesn't concern itself with types. Scheme is typed, even if it is loosely-typed or dynamically-typed, because in the real world we have to store and manipulate real data, rather than mathematical symbols. That Scheme is based on ULC just means that it likes some of its features. — Robert Harvey 43 secs ago

Hey, who said that??

^-- I think this is actually the crux of the answer right there.

ULC is all about just working with symbols, LISPs may be based on some form of lambda calculi, but a LISP which only provided and worked with abstract symbols could never create real software.

that said, somebody really should actually detail the type system of "Scheme" in an answer on that question, and then talk about how it differs from ULC/STLC/etc. Like I said, I'm comfortable with the ULC "type" system, but I don't know Scheme's so I can't make an answer

@RobertHarvey +1 for your answer, makes sense and I think answers the question that was being asked. You can never be sure with dokkat though...

user55340

@RobertHarvey You should ask Yannis about him in the TL. Digging around MSE is a bit harder after the split. stackexchange.com/users/1021583/viclib?tab=activity shows that he's also moved some of his more... lets call them 'theory' questions to TCS. Asking a P.SE mod about his deleted questions could give some insight there too.

A: How are the conceptual pairs Abstract/Concrete, Generic/Specific, and Complex/Simple related to one another in software architecture?

9:22 PM

@RobertHarvey lambda calculus to functional programming is like a turing machine to assembly -> It defines the overall precursor concept, gives a frame to understand many different implementations from.

in the same way, learning an assembly, people learn the concept of a turing machine just by unintentional side effect, as it is learning a proper functional language you'll learn lambda calculus by unintentional side effect.

what's that damn word that means things have no relation - not perpendicular, it's ol___icular or something like that?

0

Generic/Specific and Abstract/Concrete are the same axes. They're synonymous, just different terms - when you make something generic, you're making it non-concrete, rather abstract. When you make something specific, you're making it non-abstract, rather concrete. Complex/Simple is an unrelated a...

^-- No idea why I felt like answering this question, it's kind of a bad question, but I'm rather pleased with my answer

I made up pseudo math. I'm interested to see if anybody looks at my approach to logically deducing whether a relationship exists and DVs/comments saying that's a bunch of hogwash - because it totally is. I have no idea if that approach to deducing something is right or not, but it sure is convincing :D

Q: How does I/O work for large graph databases?

9:51 PM

...most of the content in that persons question makes absolutely no sense. Just the questions at the bottom make sense, everything else is just some pointless jabber

user55340

> I should preface this by saying that I'm mostly a front end web developer, trained as a musician, but over the past few years I've been getting more and more into computer science.

user55340

1

I should preface this by saying that I'm mostly a front end web developer, trained as a musician, but over the past few years I've been getting more and more into computer science. So one idea I have as a fun toy project to learn about data structures and C programming was to design and implemen...

database-design database-development io graph-databases

@MichaelT ... ? asking for a CV?

I would love to work on a database system. That would be awesome.

I don't have a tenth of the experience or knowledge necessary to speak with any level of detail whatsoever about how a modern DB actually works internally however.

I just know enough about how they actually work to know those waters run deep

user55340

@JimmyHoffa He's not really sure what he's asking. There are some big ideas... but if you read that question you'll see the 'most of it makes no sense...' or 'based on mistaken assumptions'

user55340

again, the questions make sense if taken by themselves. The understanding of it is... confusing.

9:58 PM

@MichaelT yeah well, as you said, dude's in way over his head

user55340

And thus your issue with the question you answered.

aye

I want to post an answer to that DB Q: "B+ trees." submit

oh it's the same poster

user55340

Exactly.

user55340

He's in over his head with some big ideas that are based on mistaken assumptions.

@JimmyHoffa orthogonal

user55340

10:01 PM

(Whenever I read a 'wat' question, I tend to go poke at what other questions there are to see if I can get some context)

@psr thankyou! I can never remember that word. Oligarchy was coming to mind which is wrong to say the least.

This bit of code I've spent the afternoon writing

is literally the worst piece of crap i have written in 2 years

because i was unable to convince my boss that doing it right is better than doing it fast

</RANT>

i feel like i'm pulling my own teeth

@durron597 I oft wish I had that tradeoff in front of me, but I code very quickly... usually right and wrong vary in timeliness from me very little, but yet because others code more slowly I'm held to the tradeoff and told to do it wrong in the same amount of time

@durron597 Don't learn Haskell then. Ever. You'll realize everyone is doing everything wrong...

@JimmyHoffa well, the "right" thing is to refactor the thing i'm modifying before making the changes

doing it without the refactor first is the problem, it's basically half a day vs. 1-2 days to refactor, plus another half a day to add the functionality i'm doing now

and i will probably Ctrl-A+Del when i am allowed to refactor

all this

@JimmyHoffa I've had stakeholders tell me I have to do the option which takes more time because they know better than I do that it will really take less time.

So they get both slow and crappy.

user55340

10:17 PM

@psr heh... Employer^^ that was standard practice. The business always 'knew' better than the devs how to implement something.

@psr yeah, that's fun. On the bright side, they know better than you so the end result is actually quite good, you just don't know it! :)

user55340

The biggest issue was that we were effectively given the technical design from the business rather than the requirements... "put this field in the database"...

@MichaelT ugh.

user55340

Employer^^ has job openings (heh)... and given the overtime and bonuses they toss around, my W2 for 2013 at Employer^^ was 2x that of Employer^ ... that said, there's a reason I'm not considering that at all.

@MichaelT and the difference between W2s still didn't cover the amount you spent on the mental health services Employer^^ caused you to require...

user55340

10:20 PM

@JimmyHoffa I suspect I am still catching up on sleep from there.

@MichaelT Yes, that was generally how that would happen.

@MichaelT It might be worth it to say "Hey I know your system really well and I know how to fix your workflow to right the ship; give me a job with enough authority and salary and title to fix it and I'll do it"

user55340

@durron597 I know the place... it wouldn't happen until there was a change of ownership.

@JimmyHoffa Even if they persistently hated the end result, they would still know with absolute confidence that it's better than it would have been if they let the developers design it.

@MichaelT what's the worst thing that could happen, if they laugh at you, nothing lost, if they follow through, you'll know inside 2 days whether you actually get the authority you asked for, if they don't give it then just leave

user55340

10:22 PM

@durron597 Seriously, the worst that would happen is they say "yes"...

@MichaelT even if you're given the authority you'd need to right the ship?

@durron597 They are unlikely to follow through and they pay no price for MichaelT having a job-hopper's resume, but MichaelT does.

user55340

@durron597 I have no desire to be CIO... which would require sacking the existing CIO, most of the directors, and probably everyone above too. The structural problems are significant.

@psr Being at a job less than 6 weeks actually doesn't harm your resume (unless you do it over, and over, and over, and over)

The harm is the 1 year stay. Work somewhere 6 weeks, they say "why did you only work 6 weeks?" "it wasn't a good fit", and you have something interesting to talk about

2, 3 years, or more, it was time to move on

gnat

@MichaelT yes, great thing to know, especially when I'm short on votes :)

10:24 PM

@durron597 6 weeks isn't enough to get far in a big structural change. It's enough to find out if the culture is broken, but he knows the answer to that.

but 1 year, after 1 year you've just finally settled into their system, any training and overhead has been paid, now it's time to get stuff done... except you leave.

@psr of course, they got what they wanted -> The developers did a great job, and they couldn't fuck up so I'm sure it's all correct.

@MichaelT So they'd laugh at you. Maybe not worth the trouble. But if they actually did do all those things, maybe it would be worth it... except you actually don't want to be CIO

user55340

Managers don't have the authority to fix it. Directors don't have the authority to fix it. They can shelter their own somewhat... but you've got battling directors pointing the blame as to why things don't work at each other. That means the CIO is the proper spot to target, but that's got its own problems (private company and a good deal of cronyism)

but that's what i was saying; inside 2 days you'd know if the owner planned to follow through on sacking the CIO

user55340

10:26 PM

But I know now that it would never happen.

@MichaelT apply to be owner.

@JimmyHoffa Now there's an idea.

user55340

It would take a bit to buy him out of a family business.

@MichaelT all I'm saying is that if the owner is losing money hand over fist, he might be willing to do more than you think.

@MichaelT buy? No, they'd be buying a new owner. You would own the business and collect a salary.

user55340

10:27 PM

@durron597 he's not though... and it would take a LONG time to lose 7.5B, even hand over fist.

@MichaelT you'd be surprised how billionaires move if they are losing even 1% per year. but if they're not actually losing money, then that's a different story

user55340

@durron597 compared to the competitors, Employer^^ is the company that is growing (and has been growing). Part of the difficulty is that IT is seen as an expense (to be minimized) rather than a source of savings (efficiency) across the company.

Q: Software Development Costs Pyramid

@MichaelT aha

I'm about to email this to my boss. What do you guys think of my tone:

> I made the changes.

FYI we have been talking about unit tests today, I created a bug when I was making the change, and the unit tests caught it. With the unit test I found and fixed the bug in under 5 minutes. Without them… it would have taken a lot longer.

It’s easy to underappreciate unit tests because the problems they prevent don’t stay in the forefront of one’s mind for very long, and the ones they miss take so much longer to find and fix. But I thought you should know.

(context, we had a big discussion earlier about the value of unit tests, and he was saying he doesn't want me to spend time on them and i should put more time into integration testing)

because i spent a lot of time this week debugging something that wasn't covered in the unit test

maybe i should just tell him tomorrow instead of putting it in writing

mmm. i'll mull it. have a good night all

user55340

10:46 PM

Show the output of the test suite before the bug was fixed, the delta for the fix, and the output after the bug was fixed.

user55340

Point out that this is to ensure the functionality continues to work and changes in one part of the code don't inadvertitenly break the code in another.

user55340

6

A friend was telling me the other day that there is a pyramid for the costs of fixing a problem in the software development life cycle. Where could I find this? He was referring to the cost of fixing a problem. For example, To fix a problem at the requirements stage costs 1. To fix a ...

software-engineering software-quality

user55340

The thing is, that doesn't quite capture it correctly. Its more of 'the longer it is between bug creation and identification, the longer it takes to fix - and that longer goes up exponentially'. Fixing a coding bug in CI unit tests is the 'same' as fixing a design bug in design review. Fixing a design bug in integration testing is very expensive.