The Whiteboard

gnat

12:00 AM

17 hours ago, by Ixrec

I'd probably edit it to something like "Why use an opaque 'handle' in an API rather than a typesafe struct pointer?"

Ixrec

I was feeling too lazy to go find that again but now I suppose I must

done

durron597

12:42 AM

@gnat @MichaelT @GlenH7 meta.stackexchange.com/questions/252690/…

gnat

1:04 AM

I can't understand something with this recent question. First, OP asks at GameDev.SE. Over there, it gets closed with reference and link to SO. Instead, asker quickly (in 5 minutes!) registers at Programmers and reposts over here. Why didn't they simply click SO, looks like they didn't even created account there, as if they didn't even try...

oh. They found the way to SO now, registered and reposted, and got to -5 (not surprising). And that account picked older question that was migrated from Gamedev to SO few weeks ago. Wow, what a mess. But still, why did they initially ignored reference to SO in close banner and got to Programmers, I can't understand that

user20683

meh

user20683

buy a damned book like the rest of us

user20683

Game Dev is hard

user20683

and people don't understand this

user55340

@WorldEngineer But do you know how big the C++ books are? Can't you fit that into a text area for me?

user20683

1:12 AM

mostly because they want game design but are trying to be game programmers

user20683

having known a goodly number of game devs of either stripe, it's a fairly divergent skill set

gnat

Sorry in the message posted in game dev about it being off topic when it said stack overflow i didn't know that it ment to be posted there because i thought stack overflow was just the name of this site. — Funshine 13 mins ago

^^^ but there's not just message, there's a link. All that guy needed to to is to click it. Instead they proceeded over here and reposted here. What a mess

user55340

@gnat that message was a bit... disingenuous given that the OP had posted on Stack Overflow 20 days ago.

gnat

@MichaelT no it's fair. I also thought they lie but then checked details. Their older post was migrated from GameDev, and they registered at SO only today. Today's registration picked that old migrated post and added to their profile. Doesn't make things less strange but looks fair

user55340

@gnat hmm. Well, they still should have been aware of it. sigh

gnat

1:19 AM

@MichaelT yes that strange. I can't understand that. Most interesting, what pulled them to Programmers when they had so many hooks to land at SO instead

user55340

Bob Ross Remixed | Happy Little Clouds | PBS Digital Studios

►

durron597

in SO Close Vote Reviewers on Stack Overflow Chat, 40 secs ago, by durron597

cv-pls unclear How to read a file in c++? - Funshine 2015-08-28 01:04:27Z

user20683

@MichaelT Bob Ross painting finally gave up and started uploading videos to Youtube rather than DMCAing everything

gnat

@durron597 looks like you're trying to break the beautiful race to help...

interesting, that's a bit different of what I thought. A different kind of race, a race to help. And it's not driven by rep gains, just a desire to help. And there are many willing to help, because the question is simple, and it is that simple so answers will be identical, and answers notification saves time to others willing to help, so that they can get to next simple question that also may benefit of quick help, correct? That looks like a beautiful picture — gnat 10 hours ago

we saw two help racers today at Programmers, they were in such a hurry to explain poor asker that syntax error reported by compiler was because of wrong kind quotes. Bad bad WorldEngineer broke that help race

durron597

1:35 AM

@gnat My most recent help race. Missing is the version that had no code in the first grace period

gnat

@durron597 yeah I saw it, I remember you mentioning it over here. And there were strange downvotes, at question and both upvoted answers, as if someone was afraid that 4 answers in half an hour to a question that doesn't look close-worthy could push it to hot network list

durron597

1:51 AM

@gnat I downvoted the other upvoted answer because it was really bad advice.

Then, he edited it so it was no longer downvote worthy, but I didn't want it to be higher than mine, which was, IMO, the correct answer.

As soon as mine hit +1 I removed the downvote.

gnat

@durron597 tactics! I thought about something like that when I saw that downvote was removed after edit over there

durron597

It wasn't rep tactics, I just didn't want OP to go with a bad answer.

gnat

@durron597 I think delaying vote revert after edit qualifies as tactics. Reasonably fair mind you (I sometimes do something like that myself, and I don't feel ashamed) but still tactics

durron597

@gnat Perhaps, but I didn't think of it that way.

Ampt

2:52 AM

@MichaelT All this needs is a sick bass drop

this will be echoing in my head for the rest of the weekend

user55340

3:09 AM

I believe in happy little clouds.

Ampt

Mister Rogers Remixed | Garden of Your Mind | PBS Digital Studios

►

You know what this is?

My sick beat machine.

Ixrec

6:46 AM

looks like that handles vs pointers question from CR was a good fit here after all

Duga

7:31 AM

I'm not sure this is the right place to ask this kind of questions. Maybe Programmers would be better. — Mr Lister 22 secs ago

enderland

11:46 AM

@ThomasOwens is this your alter ego? workplace.stackexchange.com/q/52638/2322

"Owen Thomas"

Jimmy Hoffa

12:16 PM

Happy Coffee Day

MetaFight

why thank you.

and a happy coffee day to you as well.

season's roastings.

3

Duga

12:39 PM

This is a great question, and probably suitable for programmers.stackexchange.com . Many people see TDD as more of a design methodology than a testing methodology. — David Wallace 19 secs ago

Niklas in Stockholm

12:55 PM

Any good books for learning the theory of front-end web development for an experienced software developer? Or other material to learn by other means than learn-by-doing.

Jimmy Hoffa

@NiklasinStockholm there is no theory, with front-end it's only Do or Do not.

Niklas in Stockholm

@JimmyHoffa Thanks for the answer Jimmy.

user55340

What do you consider to constitute front end?

Jimmy Hoffa

1:36 PM

@MichaelT he means access, he's trying to understand the theory of Microsoft Access

enderland

DONT DO IT

user55340

@JimmyHoffa I'm sorry for your data loss. Here is a book on that theory.

user55340

amazon.com/On-Grief-Grieving-Finding-Meaning/dp/0743266293

user55340

"No, that data has to be here somewhere" - "Damn it access!" - "ok, maybe I have a backup here that will load". "Noooooooo!" - "ok, I lost it. I'll have to re-enter the data and put in a request for Oracle in the next project."

overexchange

I need help in writing the constructor for derived class for the given base class

1

Q: How to write constructor in derived inner class?

Below is the base class relevant code DblyLinkList, complete code for DblyLinkList class here, package JavaCollections.list; public class DblyLinkList<T> implements Iterable<T>{ protected DListNode<T> sentinel; protected int size; class DListNode<T> { private T item; pr...

It is an exercise of inheritance and encapsulation

user55340

1:56 PM

@overexchange have you tried asking in the Java chat room on stack overflow?

Jimmy Hoffa

@overexchange have you tried striking the word derived from your vocabulary and seeing how you design software henceforth?

enderland

Have any of you positive experiences with Agile and a very small team (6 total people)?

user55340

Tends to six cowboys with weekly meetings. Good or bad depends on skill of being a cowboy.

user55340

2:12 PM

The agile part is "we're fixing bugs and adding features at this rate." Which can be used by managers to try to guess end date.

enderland

I'm debating trying to push more agile like stuff for our current team, our pseudo-standup has been very useful so far

user55340

Re read the agile manifesto and remember that is the goal and ideal.

user55340

Toss in the software craftsmanship manifesto - something to not lose sight of.

user55340

Software craftsmanship

Software craftsmanship is an approach to software development that emphasizes the coding skills of the software developers themselves. It is a response by software developers to the perceived ills of the mainstream software industry, including the prioritization of financial concerns over developer accountability. Historically, programmers have been encouraged to see themselves as practitioners of the well-defined statistical analysis and mathematical rigor of a scientific approach with computational theory. This has changed to an engineering approach with connotations of precision, predictability...

user55340

Anything that is done in the name of agile that takes you away from those manifestos needs to be considered critically.

Jimmy Hoffa

2:17 PM

@enderland just remember agile != scrum. Agile is about close-customer-contact and rapid feedback loops. Code small pieces at a time to get fast feedback from your development activities, feed those small pieces out to test often to get fast feedback on their quality, feed those small pieces out to customers often to get fast feedback on their validity for the use cases. Scrum attempts to be a framework for all those activities, but frankly I would avoid it like the plague for a 6 person team.

Ampt

Happy Coffee Day Progs

user55340

I'm also going to point out agile came from a bunch of consultants as a justification for more billable hours as requirements change. "Sure, you want that feature? Well add it to the backlog."

Jimmy Hoffa

Fast feedback is undeniably a good thing for anyone in a production position to have. Period. Focus on that because it's a simple goal, it's a simple concept, it's hard to fuck up when you focus on it. Fanciful process frameworks are all just distractions unless you have a lot of people you need organized to some nominally consistent system. It's like fast food, you can't customize for that many people so it's not made to order. For 6 people, you can make to order, so ignore the mass products.

If you're producing steel girders, coffee, or software, the quicker you get notified that you made a mistake or aren't meeting the clients desires, the quicker you can correct yourself and the less wasted time their is. Fast feedback is a simple concept to understand and a worthy goal to pursue.

user55340

There are good ideas in agile, but dangerous ones too.

user55340

> Individuals and interactions over processes and tools
Working software over comprehensive documentation
Customer collaboration over contract negotiation
Responding to change over following a plan

enderland

2:23 PM

Part of my concern with making our team "more agile" is that it's harder to break down our projects into user stories and have meaningful acceptance criteria that the development team can actually validate

user55340

Process? Itel? Be gone with corporate governance.

Ampt

@MichaelT Really? I would think the "We've got this 24 month waterfall project" has more dollar signs associated

user55340

Documentation? Writing comments takes away time you should be writing code.

Ampt

The code is self-documenting. Duh.

user55340

Screw the plan. Software is fond when it is done.

user55340

2:25 PM

@Ampt waterfall has a fixed $$$ to it. Agile may be $$ or $$$ or $$$$ when you are done.

enderland

@MichaelT you think waterfall has a fixed $$$ ?

Maybe initially, a fixed cost

user55340

if waterfall is done without expanding requirements part way through, it is less likely to have cost overruns.

user55340

"We'll do it as a change request and take some of the time out of QA" isn't waterfall.

user55340

I'll also point to the customer collaboration line - read that as what a consultant has in mind.

Thomas Owens

@MichaelT I don't know if I agree with that, given how people implement waterfall.

user55340

2:30 PM

Waterfall is rarely implented to its ideals.

Thomas Owens

Even if you don't change your requirements, if you push your integration to very late in the project, you don't know if you've understood your requirements. If you didn't understand your requirements, you can easily overrun in rework.

Jimmy Hoffa

honestly who damn well cares. Nobody- NOBODY -ever follows a process rigidly. They just take one as a model and change 50% of it to whatever management wants to make them feel good about it, then they bitch, because that's what people do about process. It's just like the difference between principles and patterns, all the AbstractManagementFlyweightStrategyFactorys in the world won't solve problems and a process won't cause software to be made.

user55340

You need good milestones.

Robert Harvey

You need good millstones.

enderland

problem with waterfall is business never knows what they want

user55340

2:32 PM

That's a man month chapter.

enderland

and most people don't plan for that

user55340

@enderland that is always a problem.

Thomas Owens

@MichaelT You do. But if your first integration milestone is at the end of the project, that's still not good. You could hit the milestone and still run into problems.

user41796

@RobertHarvey And thick ropes to tie them with.

Jimmy Hoffa

If you want it done, and done well, focus on the task at hand and think about the things that are beneficial to the quality of the task once completed. Process? Fast feedback, collaboration, peers with respect for eachother. No framework will give you those things just like no MVCVMCVMCMVPVMCPV design will implement SRP for you.

user55340

2:33 PM

@ThomasOwens then you set yourself up to fail in the project plan. This is where most waterfall fails.

Jimmy Hoffa

@RobertHarvey I want to see this wheat done by the end of the week, or I walk!

user55340

However, that isn't a failing of waterfall. That is a failing of the plan.

Thomas Owens

@MichaelT If you're planning to waterfall, how can you have integration sooner?

user55340

Agile tries to solve that problem, but means you can't ever set a fixed budget on the project. This is wonderful for consultants.

Thomas Owens

As soon as you move to iterative and incremental models, you aren't waterfall anymore.

@MichaelT One sec.

Jimmy Hoffa

2:35 PM

@ThomasOwens you may call it a river, I call it a horizontal waterfall.

user55340

Multiple prototypes with working interchangeable modules for the other groups to code against.

enderland

so basically the biggest problem with Waterfall is you (often?) have a plan that has a X month long project where the only deliverable is after X months?

user55340

That is how it's often done. That is a failing of the plan.

Jimmy Hoffa

@ThomasOwens untrue, incremental model is waterfall, iterative is more recognizably agile. People constantly use "incremental" and "iterative" interchangably when they mean exactly the opposite thing: Incremental means you complete something before moving on, iterative means you repeatedly iterate on something, doing partial completions bit by bit, moving things forward before you've completed

Thomas Owens

@JimmyHoffa Incremental isn't waterfall. Incremental is sequential, but waterfall is a specific implementation of a sequential model.

Jimmy Hoffa

2:38 PM

@ThomasOwens yeah, so waterfall is a specific implementation of an incremental model. Or, Waterfall is incremental.

Thomas Owens

@JimmyHoffa No.

An incremental model is a sequential model. Waterfall is also a sequential model. However, waterfall is not incremental.

user55340

Waterfall model

The waterfall model is a sequential design process, used in software development processes, in which progress is seen as flowing steadily downwards (like a waterfall) through the phases of conception, initiation, analysis, design, construction, testing, production/implementation and maintenance. The waterfall development model originates in the manufacturing and construction industries: highly structured physical environments in which after-the-fact changes are prohibitively costly, if not impossible. Since no formal software development methodologies existed at the time, this hardware-oriented...

Jimmy Hoffa

... 1 = 1, but 2+2 = 4 so 2 != 1+1 ?

user55340

> Coding: the development, proving, and integration of software

user55340

Integration isn't one thing in there at the end.

gnat

2:39 PM

speaking of methodologies, I for one am concerned that we seem to be handling Structured Tag Cleanup Initiative Phase II in a bit too agile way. I would like it to be a bit more waterfalley. More process over interaction so to speak. Plan it, list me what tags to kill, put them in the line, prioritize and I'll go over them

Thomas Owens

@MichaelT It is for Royce.

user55340

that remains a failing of the often implented plan.

Thomas Owens

His waterfall is System Requirements - >Software Requirements -> Preliminary Design -> Analysis -> Design -> Coding -> Testing -> Operation. Testing is integration.

Also the SEI is providing guidance to DoD contracting offices for how to deal with firm-fixed-price agile programs, so it's not totally incompatible.

Ampt

2:54 PM

@JimmyHoffa Have you thought about setting better milestones?

user41796

@Ampt I see that your employer has trained you well in the dark arts

Ampt

@GlenH7 The bloodmagic really helps me control the massive amounts of drain from casting all these evil spells too!

2

Robert Harvey

The fundamental problem with any software development methodology that seeks to enumerate every detail of a software project (primarily for estimation reasons) is that programming is a fundamentally creative activity. Rigid software development methodologies stifle that.

user41796

@RobertHarvey ~~Row~~ Code faster you scullywag!

Robert Harvey

Thus we have verbose bondage and discipline languages, static structural architecture, and voluminous project plans. There's nothing fundamentally wrong with any of that, but don't expect to produce anything groundbreaking with it.

user41796

3:00 PM

But things break all the time...

Thomas Owens

@RobertHarvey There are basic steps that need to be accounted for, though. You need some requirements and an understanding of what you are making. You need to design and write code. You need to test - write unit tests, manual test procedures. You need to build and package your software into some deliverable form.

Robert Harvey

Yes, you need to do all of these things. But they can all be done within the context of a lightweight process. I would argue that the lighter processes are ultimately more successful, because you're not spending all your time on ceremony.

Jimmy Hoffa

@GlenH7 like bones! Smile damnit; the beatings will continue until morale improves!

@Ampt you poor sod.

Thomas Owens

@RobertHarvey Isn't that the point of lean? Regardless of sequential or iterative or incremental, you want to do things like eliminate waste in product and process, to deliver quickly, to empower the team to make decisions.

Robert Harvey

Absolutely.

Jimmy Hoffa

3:04 PM

@ThomasOwens agh don't use the word empower, it makes my sores start weeping again..

Thomas Owens

That's what I like about Lean. Regardless of your methodology, it provides tools to make it better and focus on value-adding things, either for your team or your organization or your customer.

Robert Harvey

@GlenH7 Things only break because people keep changing things. If you lock down the OS and the hardware, software lasts forever. cf. banks running on big iron. It always amuses me when a software developer is reluctant to use a software library because it hasn't had a commit made to it in the last six weeks. Maybe that's because it already works.

Thomas Owens

@JimmyHoffa That's what it's actually called in Lean, though. The Popendieck's used it. Their word. Not mine.

Jimmy Hoffa

@ThomasOwens buzzword magnates have resorted to calling the pope a dick? Is nothing sacred anymore?

enderland

@ThomasOwens this is like the best buzzword bingo phrase ever :P

Thomas Owens

3:06 PM

@JimmyHoffa Nope, nothing.

ಠ_ಠ

enderland

sigh

user41796

Bad enough that I was tempted to save off a copy for myself. Easily one of the worst I have ever seen.

Jimmy Hoffa

I never used a cover letter for years, but I've actually come up with one that sells pretty well- I've gotten lots of good comments on it even though generally they're a shit thing to include IMO. One of the interesting things I've had people tell me is when it's come across their desk a second time for another position, they remembered it clearly. A good thing I guess?

enderland

it makes me sad, how low the bar for most people's resume/job applications are

user41796

@JimmyHoffa Make sure to tweak it for each application you're going after.

enderland

3:09 PM

how much time do you actually spend reviewing a cover letter?

Jimmy Hoffa

@GlenH7 Nope. those are shit - useless garbage. My cover letter serves an entirely different purpose. It doesn't say what I think about the company, rather it's just a small personal blurb about me, to sell me, rather than selling my loyalty which is what most people try to put in their cover letter

user41796

@enderland They're rare enough that I'll at least skim them

Robert Harvey

Eh, ultimately I did away with cover letters during my job search. I decided I needed to favor more job applications over specific company research and tailoring. Everyone knows it's a dog and pony show anyway.

Jimmy Hoffa

Yeah, but I don't tailor my cover letter. It's not about the company, it's about me. Which is not what most people put in their cover letter, but it's actually gotten a fair bit of notice and appreciation. I think it helps my resume stand out a touch. Maybe. Who knows. It's as much luck as anything I suppose so meh

Speaking of resumes gads we need to pick up some people..

enderland

I'd guess that would have very mixed results; HR folks will dislike it but actual hiring mangers probably appreciate the lack of BS :)

enderland

3:29 PM

I want to run an agile project

►

user41796

Algorithms question: I have a tree of items (Foos).
Events can be created against any Foo that indicates that the Foo and its descendants are unavailable.
Events are also specified as a range of time. ie. start & end dates.
The lowest nodes of Foo in the tree also have time series data associated with them that indicates the Foo was producing something.
Lower level Foo production rolls up to higher levels in the tree. So even though the parent Foo doesn't actually produce something - it has the rolled up value of all the Foos below it.

Thomas Owens

Is everything really a Foo? I mean, if a Foo produces something, maybe it's parent should be a Bar. If you have a fixed number of levels, you can have a parent to Bar and a parent to that. Each one is responsible for maintaining or computing state of its children. You still have some loop central, but it's spread out logically.

user41796

@ThomasOwens Foo has three types. Bar, Baz, Qux. Bars are the root Foos and has Bazs underneath. Quxs are below Baz

user41796

Foo is essentially a category of things, and I have a tree of different types of things representing a hierarchy

Thomas Owens

OK. That makes sense.

user41796

3:43 PM

At the moment, I have a fixed level of things. Although the client can ask about production at any level in the tree.

Thomas Owens

I do something similar. If you ask the root node about it's information, all it does is query each of it's children and returns that cumulative data. It goes down 4 levels.

If you want a particular child of the root's information, you ask that child (root.getChild(index).getData()).

user41796

Yes. Exactly. The odd bit about this is that the root node can have an event assigned to it that affects the reported production from the lowest levels.

Thomas Owens

When I get an event, I flow it to every node in the chain that makes sense. Not every child will get it, but the root will get it, then the next child, and the next child, until the end. Each node determines if it is relevant or not and if it is, adjusts internal state.

user41796

If it helps, it's a relatively small number of objects in total. One thing I have considered is building an "unavailability" map for each of the lowest levels of Foo based upon that Foo's events and any parent events

Thomas Owens

If you need to correlate events (like one event followed by another event matters), then it gets harder.

user41796

3:47 PM

Do you store the expanded event information against all of the nodes, or do you flow them out only when you need to run your production calculation?

user41796

In my case, events are manually entered and subject to editing

Duga

Stack Overflow is for help with specific problems. More general questions like this are likely do do better on programmers.stackexchange.com — Jacob Raihle 57 secs ago

Thomas Owens

As the event happens, it flows down from the root to each possible affected node. The parent is responsible for determining which children (if any) may care.

user41796

I'm currently only storing the single instance of the event and not flowing it down to the lower nodes in order to make updates easier to control.

Thomas Owens

But I don't need to deal with editing events.

user41796

3:49 PM

Yeah, the edit makes things difficult. I can't have two versions of the same event being applied.

user41796

But it's cheap enough to flow that down at calculation time.

Thomas Owens

You could cache all events and edits in the parent and reflow on calculation.

Or just not flow until calculation anyway.

So when you edit, you only edit in one place. Then, when you calculate, you just flow a whole bunch of events to everyone.

That would require your root node to have your entire public interface. The client wouldn't be exposed to anything other than the root node.

user41796

That makes sense. Some of the event stuff has already been done by another developer.

Thomas Owens

I guess your root node is like a repository?

user41796

@ThomasOwens The client wants to be able to add events to anywhere in the tree. And as currently coded, the events get stored to the node in the DB.

Thomas Owens

3:52 PM

That style of interface, if you're querying and manipulating data.

@GlenH7 You can add events anywhere in the tree, I think. Programatically, you do it through the root node, though.

user41796

@ThomasOwens Yes, essentially it's a logical container as opposed to a specific physical thing

Thomas Owens

@GlenH7 Yeah.

So if you had a publishEvent method, it would need to know what node willl get the event as an argument. But the root node would actually just hold onto it until you did calculations (or refreshed a backing data store or whatever). The root node is actually a caching proxy repository?

That's the best description I have for it. It caches incoming events until it needs to work with it, it provides a proxy interface to your collection of nodes, and it's a representation of this data repository tree.

user41796

@ThomasOwens I can't quite switch to that approach as it is semantically different from how we do things in other places in the code. Our client is given enough awareness to know what the child nodes are, and in general we record things against the actual Foo instead of at the root Foo.

user41796

So I agree with what you're saying, but I would introduce a new semantic if I did that. And the team would be unhappy with me. :-)

user41796

But I'm going to use the flow down approach at calculation time. That's an easy enough way of balancing referential integrity needs during the edit case along with avoiding complex queries in the production calculation

user41796

3:58 PM

So the flow down results will live in memory only, which is okay in this case. They don't need to be persisted

Thomas Owens

4:09 PM

@GlenH7 Let the team be unhappy for a while. For the good of the product!

user41796

I wish. As entrenched as that approach is, all I'd be doing is signing myself up for an unpleasant 1:1 with my boss.

psr

@enderland I find it difficult to support my weight on my pseudo-pods.

gnat

folks, do we have team-spirit? I just noticed someone is cleaning this tag. I think it would fit the tag name if we help them with votes to close / delete. There are 6 questions, none is locked

Jimmy Hoffa

4:38 PM

@gnat can I just go through an retag them all teen-spirit instead?

@psr browserify bundles all your modules together into a single file or some such, aye?

gnat

4:51 PM

@JimmyHoffa only if edit tag wiki to refer relevant content...

Nirvana - Smells Like Teen Spirit

►

user41796

~~Down votes please to push it below an aggregate score of 20~~ Boom!

gnat

@GlenH7 team-spirit!

user41796

Yep. The ones that are left are too new to be taken down

durron597

5:41 PM

0

Q: Should I flag link only answers to tool request questions?

I know that, on most sites throughout Stack Exchange (except Software Recs), tool / resource requests are off topic. I know what to do with these questions, flag / vote to close them on that particular site. However, if I come across a link only answer, and then notice that most or all of the ...

discussion flags link-only-answers

user41796

5:52 PM

@user568458 - Welcome to The Whiteboard

user41796

You've got a number of threads that we could address, which part do you want to go after first?

user568458

The underlying problem is, I have a question which is on topic to your community, and which people in your community can answer. What hoops do I need to jump through before you'll allow people to answer it?

user41796

For those wondering, here's a link to my meta answer that's in question

user41796

@user568458 But it's not on-topic, that's what we're trying to explain

user568458

The more general version would be "How can we phrase questions asking about empirical research so that people who haven't worked in or with published research don't mistake them for laundry lists"?

user41796

5:55 PM

@user568458 You can't, which is the problem. We spun that one around for a long time but it boiled down to an intractable problem.

user568458

Every other site on the network manages it fine!

user41796

Asking for empirical evidence (aka research on a subject) will inherently consist of links off-site. That puts Programmers in the position of being research librarians, which isn't terribly interesting to the community.

user568458

Look at any question on ANY of the science sites.

user41796

@user568458 That's not germane. You have enough experience on enough sites to know that each site has it's own measure of culture and rules. Other sites allow it, but we're not those sites.

user41796

Heck, I'm a mod for a site that allows it within a tightly bound space.

user568458

5:58 PM

I've answered lots of such questions myself. They're very interesting to answer.

user41796

So I completely get that argument. But those requests don't work well for Programmers

enderland

@user568458 did you read the linked meta threads? such as meta.programmers.stackexchange.com/a/5455/52929

user41796

@user568458 I understand the perspective, but the community came to the consensus that they're not interesting

user568458

They follow a format like "The research generally indicates X, here is the most compelling study and here, based on my experience and expertise, are the objective reasons why it is a particularly strong study. Here are the most important and relevant caveats, and here's an example from my experience of why one is particularly important@

user41796

And part of the back story there is we were seeing a lot of crappy one or two sentence questions of "gimme teh research."

user41796

6:00 PM

@user568458 Again, I understand what you're saying. And your question on main was well written. It's just a question that doesn't fit on Programmers. Doesn't mean it's a bad question, it just doesn't fit based upon the community and the problems we've had in the past.

user568458

enderland - I feel like I'm going around in circles. "What's the state of play on research on X" is absolutely completely different to "please recommend a book [or paper] on X"

enderland

@user568458 that doesn't answer my question though

there's not a meaningful difference between "give me a book' and "give me references"

user568458

[enderland linked to yet another "why we don't allow opinionated laundry lists of book recommendations] thread

I'm asking for an expert-based summary of the research. That is NOT the same as "give me references"

user41796

> Asking for information about statistics, or more generally, asking us to search for some data for you is off topic because of these problems:

No actual problem to solve
Not drawing from expert knowledge of the community
Will suffer from link rot or become out of date

user41796

I pasted that to make sure we're talking about the same section of the meta answer

enderland

6:02 PM

I feel like you are trying to inform this site about how its community consensus about what is on/off topic is wrong, which I don't understand - can you help me understand what I'm missing here?

user568458

Look at any good answer on Skeptics.SE. It's an expert-based summary of research, WITH references, not a list of references.

Look at any good answer on biology, chemstry, physics, cogsci - even good quality answers on UX

enderland

@user568458 Skeptics is not Programmers

user568458

They're summaries of research, by people with relevant expertise, not lists of references

user41796

@user568458 As currently phrased, you're not. Look at the title and look at the first sentence of the fourth paragraph. "To clarify, I'm looking for the results of objective research in this area." The rest of the 4th paragraph backs your statement about looking for a summary, but understand how that could be misread based upon the title and the first sentence.

Jimmy Hoffa

@user568458 You have jumped through that hoop! Really! That hoop? Join The Whiteboard, and ask it here. Seriously, if something is uncertain or you just want some help, ask in here (during business hours), we're always happy to help you structure your question so it works on the site if that's possible, and even if it's not possible: We'll try and help in here best we can.

user41796

6:05 PM

And as a random aside, I thought Brooks provide evidence in the footnotes to his book. I thought there were multiple evaluations backing his claims.

user568458

So empircal research is not off topic? Great. I can carefully, walking-on-eggshells-style, edit out anything that could be misread as "gimme references", and it'll be re-opened, and I can continue with my life and pretend this debacle never happened?

enderland

@user568458 being intentionally obstinate isn't going to make anyone here support your demands

user41796

@user568458 We are trying to help you understand how the site works. One of the reasons why we took a black-and-white stance is because it's very difficult to explain gray areas to newcomers. The community wasn't willing to sign up for that.

Jimmy Hoffa

unfortunately, requesting resources just can't work: You're asking for references to studies, we really don't do that here. I realize some sites like History.SE or Skeptics, or others can act as places to ask for references to studies, but that's not really what P.SE is about.

enderland

It sounds like your actual problem is your management won't listen to you, and so your resolution to that is a list of studies/references/etc which you can drop on them and say, "see! research indicates I am right"

but there's not a clear problem - other than this - which is why your question as written comes across so much like "give me a list of resources/references"

user568458

6:10 PM

My actual problem is very specific and quite complicated. Lots of politics between an idealistic soc-ent startup guy, a government department, a university that wants to dump interns somewhere... it's messy. My immediate problem is I want something more robust than "a popular book said in the 1970s" when convincing the idealistic startup guy that he needs to slow down...

...The politics side of it is fine, I've got that. What I need is to quickly get up to speed on the facts of what exactly the real risks are if they carry on as they are.

enderland

What you need to do is create some numbers for your developer hours and assign efficiencies to them

user41796

FWIW, the 2nd ed of MMM was published in 95

enderland

and breakdown your tasks per week

maybe your core devs can do 30 hours of dev work at 100% efficiency right now

and your new devs can do 20% efficiency for 2 months, then 40% for 2 more, etc

user41796

Pretty certain that was the version I read, and I want to say there were a number of footnotes with backing evidence

enderland

but your core devs will have to lose a lot of their time to training purposes, so quantify it

user41796

6:13 PM

I thought Brooks breaks down an example of that as well

user568458

GlenH7 you know that would actually be an okay, +3 sort of answer to the question? And if someone with more expertise came along, they could post an expert-based summary of the latest evidence, complete with an experience-based and expertise-based explanation of what was missing or lacking in Brook's 1995 edition, and that would be an excellent page of content

user41796

@user568458 It would work as an answer on another site, yes. :-)

user41796

Don't facepalm too hard on that one, sorry

user568458

I editted the question based on the earlier comment. How is it now?

user41796

Again, I get what you're saying. But there's a lot of back story to explain about Programmers and where it came from. More than you care to hear, I promise you.

user41796

6:15 PM

Short version is the site was a cesspool of crap questions. SE said "turn it around" and the community has been digging out ever since then.

user568458

I editted the question based on the earlier comment. How is it now?

user41796

We're 4 years past the proclamation and we're still cleaning up the crap from before

user568458

Acch, double post

user41796

Still reads as "I'm looking for research papers" which doesn't work for the site.

user41796

We do appreciate that you're trying to put together a constructive question and you're following SE protocol with asking on Meta about how to edit a post.

user41796

6:20 PM

But I can't think of a way to rephrase the question to where it will work on the site.

user568458

How about now?

user41796

That edit might be workable.

user568458

Great.

What now?

user41796

@ThomasOwens Care to review the most recent edit

user41796

@user568458 Still working through "what is the problem that's trying to be solved."

user41796

6:24 PM

Not being able to find research isn't a problem the site can / will address

user41796

I'd rather turn it into "how do I quantify the risk" but that's too broad

enderland

@GlenH7 the core problem is, "how can I convince my management that adding more devs won't result in a faster delivery time"

Thomas Owens

It's still essentially asking for research. I don't think you can write a question that eliminates a set of good answers that are based on experience and knowledge. The sharing of expertise plus links to sources of various levels of rigor is what makes SE SE.

user41796

@enderland right, which is where I'm struggling to find a constructive aspect to that

user568458

How about now? I've added a whole paragraph explaining the problem I'm trying to solve.

user41796

6:27 PM

@ThomasOwens I keep coming back to this as well

Thomas Owens

I wonder if this is an X-Y problem.

672

Q: What is the XY problem?

What is the XY problem? When asking questions, how do I recognize when I'm falling into it? How do I avoid it? Return to FAQ index

discussion faq asking-questions terminology

user41796

In a way it is

user568458

It eliminates all answers that are not backed by actual research.

user41796

@user568458 I'd suggest it makes things worse as it opens another angle to the question.

Thomas Owens

@user568458 That's bad.

user41796

6:29 PM

X is "Convince management they're making the wrong decision."
Y is "Demonstrate that adding team members to a late project will make it even more late."
Z is "Is there any real issue / risk beyond the project being late."

enderland

@GlenH7 it absolutely is an XY problem

user568458

I'm not asking about X. I am only asking about Y.

user41796

@user568458 - So what if the project is late?

Thomas Owens

@user568458 How do you know they are making the wrong decision?

user41796

Not trying to turn things - but a project being late means time and money are lost. All in all, those aren't horrible things and happen to projects all the time.

user568458

6:32 PM

None of this is relevant to the question, the question is, how can I use facts to defend an argument based on Brook's Law. That's it

enderland

@user568458 no, you are basically saying "I need a cohesive reason adding team members to a project late in the project isn't going to make it delivered faster" which is the thrust of your post, but it's the answer to, "how to persuade my boss/management?"

user568458

We could turn any P.SE question into a philosophical "Why do you even want to ship your product" if we wanted

enderland

your post assumes the answer and asks for justification for it - not asking a tangible problem and soliciting answers

user41796

@user568458 Oh, I've moved beyond your question on main. I'm trying to focus on the problem from your perspective. In other words, I'm trying to understand the "So what?" from your viewpoint.

enderland

I mean heck, the order you wrote your post is basically exactly what I said - you talk about your solution, and then added (belatedly) a footnote as to what the problem/question really was

user568458

6:33 PM

You're imagining things that aren't there. I am NOT asking how to convince anyone. In fact I'm not even trying to neccessarily CONVINCE anyone, I just want to be sure that, when I raise this as something to consider, I have facts to back up its being something to consider

user41796

That rule within MMM is fairly sacrosanct. No one has proven it wrong to my knowledge. And based upon my comments earlier about the 2nd ed, I think the "how to" portion is laid out in the 2nd ed for you.

enderland

right, that's my entire point - your question on the site isn't a question it's a request for justification. Your mind is made up and decided at this point

user568458

I really don't understand why that's so hard to understand. There are about 30 answers on this site that are little more than references to MMM. Surely you can understand that "does this stand up to scrutiny" is a legitimate question?

user41796

@user568458 There's 2 types of project management. Those that buy into that rule from MMM and those that ignore it / aren't aware of it.

user41796

@user568458 It's a legitimate question, but as phrased it's not a good fit for the main site.

user41796

6:36 PM

Yes, we understand other SE sites can work with those types of questions. Progs is different in that regards.

user568458

THAT'S WHY I'M HERE. You're the experience people here, you KNOW what I'm trying to ask, please tell me how to phrase it so that it meets your (highly unusual, as previously established) criteria.

enderland

I don't see a way, since the core thing you want, ultimately, is "give me a list of references written nicely into a post"

Thomas Owens

What is the problem that you are trying to solve? Convincing someone else isn't a problem. Not finding references isn't a problem, either.

user568458

(and my mind is NOT made up about anything, other than that I want to warn my partners that what they are doing appears to not be good practice and appears to be a very common mistake, and I want to do so with the relevant facts at hand)

user41796

@enderland Even if it isn't a list, it's still an off-site reference for the backing evidence

Thomas Owens

6:39 PM

Wanting to warn someone, to raise a warning flag, is a solution. What's the problem?

enderland

Also this page has a lot of good guidance about this - programmers.stackexchange.com/help/dont-ask

user568458

The problem I am trying to solve is - I am in the process of convincing some partners to not try to solve a tech dev problem by simply hiring and hoping - something they are keen to do for political reasons. I'm aware that, according to Brook's writing, this is a bad practice. But...

...

I want to be sure that this really is something accurate and based on fact, and if it is, I want to understand what the evidence for and against it is and what important caveats have been discovered in research. I want to be sure that I'm not arguing for something that is out of date or disproved, and I want to be able to present an evidence-based, robust case, not simply hearsay.

That's the latest version of the final paragraph.

Whoops, I left in the trigger word "convince". Apologies! I'll edit it...

Thomas Owens

That's not a problem.

user41796

Apropos of nothing, this conversation reminds of one of the death marches I was on. Lead PM knew they shouldn't but upper management insisted on throwing more bodies at the project. Project ended up being cancelled.

user568458

(of course what I mean is "convince them to consider an alternative" but anyway...)

enderland

6:44 PM

I feel like we're just going around and around in circles with this conversation..

user568458

an Andres F. comment puts it well: I find it curious that nobody made the effort to research this, since it's one of the fundamental insights of software engineering. Anecdotes remain just that -- anecdotes. Research work that documented hundreds or thousands of cases where this purported law holds would go a long way. Research is very important for software engineering, otherwise we will remain in an age of cargo cult methodologies and processes :/ –

user41796

@user568458 It's non-trivial to research in a scientifically valid way

Thomas Owens

@user568458 I would agree with that.

Ixrec

wow I missed a lot

user568458

Yes, we are, and it's infuriating. I keep being accused of all sorts of things, when all I want to do is know what the evidence on a widely quoted piece of IT best practice is

Is it really such a novel idea, to want evidence to back up idiom?

user41796

6:46 PM

A proper experiment requires a control group and a variable group. Ideally, you would have identical teams tackling identical problems so you can control the extra variables in the research

Thomas Owens

@user568458 On SE, the back-it-up approach refers to answers. It says that you are supposed to provide experiences and/or references to claims that you make in your answer.

user568458

Please, let's not get into a discussion of research methodology, there are enough people who are experts who have considered how to do research where double-blind isn't possible

user20683

@user568458 if you can ask via the stringent rules that Skeptics requires, you might be able to get a useful answer there

user41796

@user568458 Just pointing out that it's difficult to research

user568458

World - That's actually quite a good idea. It would fit on skeptics, and those guys aren't terrified of evidence. The problem is, the relevant experts are on this site, they're just not allowed to answer right now

enderland

6:48 PM

@user568458 I suspect no one on Programmers.SE has a background in research for human studies of the sort you are looking for

user568458

There's a whole DISCIPLINE of management research. Many of your users are probably experts on it

user41796

There's overlap between the sites' users

user568458

Why not? If they're product owners, project managers, they may well have familiarity with management research. My question doesn't require professor-level expertise, just relevant experience, and knowledge.

Thomas Owens

Have you read a copy of the 20th anniversary of MMM?

user568458

Seriously, I've answered every question. We got to a point where it was clear that it wasn't asking for just references, where you just wanted clarity on the underlying problem. I've given that. What is the remaining obstacle?

Have we reached some kind of "user has done everything we asked, but we can't be seen to back down, what else can we ask for' point?

Thomas Owens

6:53 PM

You haven't done everything that we've asked for. You are still asking for reference requests.

user41796

@user568458 No, we haven't because I don't think you've listened to what we're trying to explain with why your question won't work on main.

Thomas Owens

Have you considered that your question simply isn't, and won't be, a good fit for Programmers?

user41796

Several of us have looked at it. We're good at "salvaging" a question to make it fit if it can be done. We can't find a way to make it work with yours. It's off-topic for main. It might work very well on skeptics

user41796

We are also trying to help you solve your actual problem, but that needs to be done here in chat.

user568458

Come on. You've tried everything from XY to no underlying problem to user has already made up their mind to book request to reference request. I clarified that it wasn't a reference request about half an hour ago. I explained the difference between reference request and evidence summary. There were no complaints - Thomas simply said that what was lacking was clarity on the underlying problem.

enderland

6:56 PM

You aren't actually explaining the difference though, you are just repeating that they are different over and over again

user568458

Where I explained it has dissappeared of the top of chat. How do I scroll up beyond where it stops?

Ixrec

there should be a "load older messages" button

Thomas Owens

@user568458 An evidence summary is a reference request. You want us to go out, find papers, and summarize them. That doesn't require the expertise of professional software developers.

That's not a good fit.

Maybe some sites allow that, but we don't.

user568458

I DO NOT. I explained that above.

enderland

12 mins ago, by user568458

Yes, we are, and it's infuriating. I keep being accused of all sorts of things, when all I want to do is know what the evidence on a widely quoted piece of IT best practice is

can you help me understand what you mean by that then?

user568458

6:58 PM

I want someone familiar with the area to come along and say something like "There has been research on this in [x] discipline. It fits the results in Brook's work except in that [exception, cavaet]. This fits my experience too in that [example]"

Thomas Owens

That's a reference request.

You want someone to summarize someone else's work.

user568458

It's an answer, like any other, except it's not just examples from exerience, it's examples from experience plus evidence

I want someone to summarise their expertise on the body of research.

That's the difference.

Transcript for