The Whiteboard

gnat

07:49

do we have a Back it up rule? Oh wait, we don't. Quite a pity...

-1

A: What causes bad performance in consumer apps?

I think the reason is that most consumer directed apps are controlled and marketed by people who don't know anything about software, and hire developers based on their resumes or the recommendations of some no-nothing manager, as opposed to their actual skills and knowledge.

enderland

12:22

@gnat isn't the answer "inefficient programming" combined with "not enough time to do things right" ?

gnat

@enderland well I think the "answer" is as stated in close reason text... :)

> Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise.

"bad" in "bad performance" isn't defined, leaving a wide open room for opinions :)

19

A: The halting problem - or - the fallacy of "real questions have answers"

I'm going to assume that this is a serious request (excepting the last few lines) and not just a bit of leg-pulling (it's not Friday). Seems that one of the consequences of coming up with a famous quotation is that everybody either misquotes you or takes it out of context. This is what was actu...

real questions have answers, not items or ideas or opinions* ... "Real questions" don't necessarily have *practical answers, but they do have *authoritative* ones

enderland

12:58

yup

Jakub Narębski

13:23

I'd like to as an interdisciplinary question about theoretical background of robust benchmarking (i.e. finding and rejecting outliers, benchmarking with given accuracy, etc.), but I am not sure which site would be best fit: Stackoverflow, Programmers, Stats? Any advice?

gnat

@JakubNarębski take a look at a comprehensive guidance at MSO:

59

A: Which computer science / programming Stack Exchange do I post in?

Stack Overflow Implementation problems, and question on software tools commonly used by programmers. If your code or your IDE doesn't work, ask on Stack Overflow. Does not welcome subjective questions (any more). Programmers The main focus is whiteboard questions, problems that you face while...

Jakub Narębski

Thanks @gnat

Though that still leaves me with choice between Programmers, Computer Science or Stats...

user41796

13:49

@JakubNarębski - it would help if we knew more of the question you were considering asking. Don't worry though, we won't steal it for ourselves.

user41796

it also depends upon what sort of an answer you're looking for.

user41796

Comp Sci & Theoretical Comp Sci will land you answer heavy in theory and possibly light in practicality.

user41796

Programmers will land you answers with the reverse. Heavy on practicality but not always fully backed with theory.

user41796

SO answers tend to be tight and to the point; which mirrors SO's origins. It was created for fixing specific implementation type problems.

Jakub Narębski

Looks like Programmers it is.

user41796

13:53

YW. :-) Feel free to drop a link in here after you have posted.

Jakub Narębski

The question I intend to ask is about robust and relevant benchmarking in compiled languages; how to deal with code that is too fast for clock, how to deal with outliers in benchmark data (you repeat benchmark N times, and sometimes your benchmark runs when there is spike in load; those outliers make distribution asymetric because you don't get times < 0), wallclock time (sensitive to load spikes) vs. CPU clock (does not take into account I/O time and blocking), etc.

There quite a few articles about robust benchmarking of Java (but it is virtual machine not compiled code, and artices are light on math), there are Banchmark and Dumbbench packages for Perl...

user41796

Sounds like an interesting question. Be forewarned, it may take a while to gather meaningful answers. The higher level of detail behind the question, the longer it can take for an answer.

enderland

why are microsofts help forums so bad

finding out "where should I post this" is impossibly difficult

user41796

@enderland because they're forums. And Honey Badger Don't Care.

enderland

@GlenH7 that's not suppose to make me feel better is it? :P

user41796

14:07

no, sorry. I loathe heading out to forums any more. So much noise. So little signal.

gnat

I approved this (terribly incomplete) suggestion by mistake, could please someone reject or override with a better edit? programmers.stackexchange.com/review/suggested-edits/40403

enderland

@GlenH7 no. freaking. kidding

gnat

it removes the link to deleted question (which tricked me into accepting it) but leaves all the garbage that was referring to it

enderland

M$'s forums literally duplicate the "accepted" answer(s)

user41796

@gnat done

gnat

14:14

@GlenH7 thanks!

user41796

NP. Not sure if it needs another to fully sink it

user41796

@enderland Yeah.... their forums leave so much to be desired. I love how MSDN links directly to SO as well as MS forums.

user41796

I think the MS employees working the forums try hard, but the sheer volume of inane and incomplete questions that they pick up has got to be daunting

enderland

oh I'm sure they get worse than SO does in that regard

"hi this doesnt work fix please"

user41796

@gnat why do I have a feeling that you're already out of close votes today?

user55340

14:17

@gnat I didn't even your request - I rejected it for the same reason though.

user55340

@GlenH7 I drained mine last night... lots of sunday night homework questions.

user41796

Trying to decide if the boss / version control question should get a close vote or not

user41796

it's answerable based upon expertise. The conflict behind the question seems to be more of a career / office politics issue though

user55340

If you search for the text from programmers.stackexchange.com/questions/212217/… you'll see that its a standard problem... though my search didn't find any answers for it.

user41796

yeah, I can see the focus being "should I use VCS solo on a team project" and that being a relevant question for the site

user41796

14:21

how apropos that we were just lamenting how MSDN forums can't / don't keep their site clean of garbage and that they can't / don't have these types of discussions about question quality

user55340

@enderland Its kind of fun finding those questions on the cliff of auto deletion and downvoting the 2 line 'here's the answer' that got an upvote somehow to push it over.

enderland

@MichaelT yeah. on Workplace I'm about 2:1 up to downvote

woah I need to UV more

user55340

16

Q: A badge that rewards "prophetic" downvoting

How about introducing a badge that rewards downvoting, but only of the sensible kind? A measure for sensible downvoting would be a high percentage certain number of downvotes on content that has subsequently been Deleted Closed (except migrations and duplicates) Bronze, Silver, and Gold b...

vedantchandra

@MichaelT I think the answers on that question are right, its a good thought, but won't work in practice

enderland

14:39

@MichaelT already upvoted, hah

though tbh I think rewarding DVing will result in more people DVing than necssary

vedantchandra

Exactly

user41796

@enderland removing voting from the close review queue was likely a good thing. Carp questions could get hammered really hard just by rolling through the close queue. So I think you're right in rewarding down voting would have unintended negative effects

Jimmy Hoffa

@GlenH7 Which would roll them off the front page; possibly a good thing.

user41796

@JimmyHoffa I can see that argument. I'm not sure where the balance point is at.

user41796

-6 on a 1st question that wasn't necessarily all that bad is a bit harsh. I'll claim that all of the examples I'm thinking of have been deleted, but I'm pretty sure it happened more than once

gnat

14:50

@GlenH7 that's right, although I didn't CV "Boss is afraid..." for a different reason. I need some time to decide whether it's worth closing or not

user41796

generally speaking, I think -1 or -2 is enough to say 'carp question, please fix"

enderland

I almost never DV a question when I vote to close if I think it's a "needs and will get improed" Q

I DV questions which are lost causes

but eveyrone has their own voting style

whcih makes badges about this sort of thing difficult

vedantchandra

See, this is all assuming that downvoting will be done responsibly. But with the badge incentive, people might downvote not-necessarily bad posts that are flagged for closing just for the badge.

user41796

@enderland I'll put a -1 as a signal to other reviewers walking the question page that it needs more close votes. But I won't pile on with additional down votes unless it's a really bad question

Jimmy Hoffa

@GlenH7 Agreed, just playing devils advocate. I think we have a fine balance without voting in review queue so there's epidemiological evidence that no voting in queues is the right approach. Yes I did just invoke that word for that purpose.

enderland

14:53

@GlenH7 yah, same (or at least most of the time when I'm aware of it...)

gnat

@GlenH7 Marginal short answers with links has a side effect of picking extremely close-worthy questions, that's why I am often out of CVs when working with it. By the way, as of now, query finds only about 200 posts with length <200, wonder how much it would find at SO

user41796

@gnat some questions are best left unasked

gnat

don't try it at SO with length 200 since it is already been proven to time out. @MichaelT how could one estimate amount of marginal answers with links at SO having length under 200 - and avoid timeout?

enderland

@GlenH7 but NOT THIS ONE

gnat

@GlenH7 that's Mu - "Unasking" the question

Mu (negative)

The Japanese and Korean term mu () or Chinese wu () meaning "not have; without" is a keyword in Buddhism, especially the Chan and Zen traditions. The word The Chinese word wu 無 "not; nothing" was borrowed by East Asian Languages, particularly the Sino-Xenic "CJKV" languages of Chinese, Japanese, Korean, and Vietnamese. Pronunciations The Modern Standard Chinese pronunciation wu 無 historically derives from (c. 7th century CE) Middle Chinese mju, (c. 3rd century CE) Late Han Chinese muɑ, and reconstructed (c. 6th century BCE) Old Chinese *ma. Other varieties of Chinese (fangyan , tech...

the term is often used or translated to mean that the question itself must be "unasked": no answer can exist in the terms provided. Zhaozhou's answer, which literally means that dogs do not have Buddha nature, has been interpreted by Robert Pirsig and Douglas Hofstadter to mean that such categorical thinking is a delusion, that yes and no are both right and wrong.

In Robert M. Pirsig's 1974 novel Zen and the Art of Motorcycle Maintenance, mu is translated as "no thing", saying that it meant "unask the question". He offered the example of a computer circuit using the binary numeral system, …

above excerpt managed to refer two of my favorite books

guess we need to replace boring "delete" link label with 無

Jimmy Hoffa

15:07

@gnat This is new to me, I always understood this concept as the perspective of non-dual reality

Nondualism

Nondualism, also called non-duality, is a term and concept used to define various strands of thought. Historically it is linked to the Advaita Vedanta-tradition, which states that there is no difference between Brahman and Ātman. It is also found in other Asian religious traditions, but with a variety of meanings and uses. In western cultures the term is also used in modern spirituality and New Age for a "grand tradition" which sees "a primordial, natural awareness without subject or object" as the essence of a variety of religious traditions. The term is also used to refer to interc...

> By and large, Kaccayana, this world is supported by a polarity, that of existence and non-existence. But when one sees the origination of the world as it actually is with right discernment, "non-existence" with reference to the world does not occur to one. When one sees the cessation of the world as it actually is with right discernment, "existence" with reference to the world does not occur to one.

user55340

catb.org/jargon/html/M/mu.html

user55340

> According to various Discordians and Douglas Hofstadter the correct answer is usually “mu”, a Japanese word alleged to mean “Your question cannot be answered because it depends on incorrect assumptions”. Hackers tend to be sensitive to logical inadequacies in language, and many have adopted this suggestion with enthusiasm.

user55340

(realizing the old use of the word 'hacker' not the modern one)

user41796

I have interpreted "wu" as meaning "empty" or "void", so it represents absence rather than contradiction which "no" or "not" imply. @WorldEngineer may be able to weigh in on this subject. I believe he studied Buddhism to a fair degree, and the concept is integral there.

user41796

Can someone look at this tag and see if the link for excerpt is displaying correctly: programmers.stackexchange.com/tags/interrupts/info It's the same link as the body, but it doesn't look right when I click "show except"

Jimmy Hoffa

15:20

@GlenH7 Right; new term to me for it. I always think of heat and cold as the simple analogy to explain the concept: Heat is the presence of energy, cold is the absence of it; thus they aren't defined by each-other as would be dualistic, but by their nature; non-dualistic. Though I've no frame of reference for such things other than random crap I've read with no outside advisement or influences to tell me when I get shit wrong.

user55340

@gnat heh... stumbled across this page on branching - perforce.com/blog/090521/using-branch-diagrams-scare-your-ceo

user41796

@JimmyHoffa The challenge with heat and cold is they imply the presence of something. Wu (my preferred term due to Chinese influence in my kenpo) would mean the absence of both. Neither heat nor cold. Just absent.

user55340

If nothing else, you'll like the title.

Jimmy Hoffa

@GlenH7 What is cold the presence of? Wu.

user41796

Cold is the opposite of heat. Wu would mean neither heat nor cold. Just being.

user41796

15:22

Or not being. Empty is okay

user41796

The problem with Zen is that language doesn't provide the constructs to discuss it

user41796

And it tends to be pretty esoteric so people get weirded out quite quickly

user55340

Hmm... finding web pages based on neat diagrams gives fairly good results. Search google images for perforce branching and then look at the pages that have the best diagrams. google.com/…

user55340

3

user55340

This one counts too I think.

Ampt

15:39

Time to flex the old programming muscle after spending all weekend on Linear Programming. Any good blog posts in the last week? Didn't see anything from my admittedly lack-luster googling.

Jimmy Hoffa

@Ampt If you haven't read it, The kingdom of nouns is relatively expected reading

Ampt

@JimmyHoffa how was the trip to the wilderness?

Jimmy Hoffa

@Ampt awesome. Horseback riding in the morning, a manhattan in the sun by the river in the afternoon, local caught grilled trout for dinner, walk down the strip after dinner, and some pitch dark no-light hot springs time after that. Steamboat++

vedantchandra

@JimmyHoffa Sounds fantastic!

Ampt

@JimmyHoffa that sounds like a lot of fun! Glad you had a good time. Now onto this article.... He's really pushing this analogy that he's got going, and it's not exactly holding up on close inspection. But I'll finish it before I judge too harshly

Jimmy Hoffa

15:51

@Ampt Yegge's writing style is pretty in your face and he really stretches things; but it's all just to make a point. he has a lot of good thoughts, many controversial but they at least get ya thinking. Plus his writing is usually pretty entertaining which makes it worth reading anyway.

Ampt

The whole Verbs do things, Nouns doesn't, really isn't holding up for me. Sure, taking out the trash is an action, but you need someone to actually do the act of taking out the trash. It doesn't take itself out, as my girlfriend has assured me on many occasions.

Jimmy Hoffa

@Ampt You still haven't written any LISP or other functional language, the analogy makes a lot more sense then. In LISP you don't really model things, you model actions, and it's rare that the action is attached to a thing

Ampt

so this only holds up if you aren't really using OOP?

Not that OOP is the end all solution but if he's knocking the main OOP language for being Object Oriented....

Jimmy Hoffa

@Ampt His analogy is somewhat showing a distinction between Java and some less Java-OOP style approaches

Ampt

I mean he's got some good points in here about the ridiculous use of execute() and run() and so forth which is just poor programming imho

Jimmy Hoffa

15:57

You could write python with less nouns, his point is that Java has no facility to do without creating a noun that does, and he's showing what the result is when one thinks in that way; when one takes the OOP paradigm to the extreme, completely refusing to model anything other than things

gnat

@MichaelT I also like the picture

Ampt

Is it bad that I saw that and went "Well that's not even that bad really"

Jimmy Hoffa

@Ampt Nah, I think this is a great diagram. It's a flow chart for how management get's things done right?

Ampt

looks like some of the teams I've worked on. More branches than 2^(people in project+1)

not sure why everyone is so branch happy but whatever, at least they aren't overwriting my code with their crap.

Jimmy Hoffa

@Ampt Until they merge their code with your code. Then they're overwriting half of your code, and creating a system that hardly runs and you'll be blamed when your piece breaks because half of it has been misplaced by their changes

Ampt

16:01

(this is at school btw. At work we use the "MainLine" model)

@JimmyHoffa my policy is "You break it, you buy it"

Jimmy Hoffa

@Ampt "MainLine" model means so many different things around the industry. If your work actually has a consistent branching model that they actually follow, consider yourself blessed. Most companies in industry either have no model so much as "I think I need a branch now?" or "Agh branches are painful no branches!", or they have a model which they simply never follow

Ampt

nope, we have working branch->staging->mainline

staging and mainline go through rigorous unit testing

at least as far as the main system is concerned. Individual modules are less enforced

Jimmy Hoffa

That's simple, there's really two consistent easy to follow branching models I've seen work very well, everything else I've seen has people wondering where there head is (that I've seen) and that is one of the good ones

the other one I've seen work well is Feature_X->Integration->Main where each feature gets it's own branch. This is the best way to do it when you work in a company incapable of planning such that every release they want to take half the features out and move other features in etc, which when you're all working together in a WorkingBranch doesn't work because you can't just pull some features out

user55340

@gnat Thats actually how I found it... looking for professionalyish made branching images.

user55340

@Ampt Read paulhammant.com/2013/03/11/legacy-app-rejuvenation

user55340

16:06

user55340

vs

user55340

Thomas Owens

@JimmyHoffa Not just companis incapable of planning, but also customers who are indecisive. Especially if your customer is an integrator that doesn't have the resources to verify an infinite number of changes with other systems.

Jimmy Hoffa

@ThomasOwens Yes, I just mean in the situation that you for whatever reason (project managers, resource managers, marketing, clients, space-invasion-risk plans) cannot consistently follow through on a release plan, that branching model is your wisest approach.

enderland

arg working with file paths is so annoying, I can either write in all sorts of validation in this class or just make sure to code in such a way to not do things out of order...

Jimmy Hoffa

16:09

In that branching model it's oft suggested that only the leads/veterans can commit to Integration because you end up with occasionally a fair bit of manual merging work when you start pushing your features together

user55340

Thomas Owens

I prefer it in general, really. But then I'm somewhat scattered. If I'm working on feature/bug X and get stuck, I want to be able to work on feature/bug Y and not have to worry about any problems with merging the wrong thing.

user55340

(from jawspeak.com/2010/11 )

enderland

@MichaelT is that supposed to be empty?

Thomas Owens

It's not uncommon for me to have three or four items "in progress", each in their own working branch.

Ampt

16:10

@JimmyHoffa that's what we do. One guy is essentially responsible for our integration and integration testing. We have a bus factor of 1 with this guy but I don't think that will change anytime soon

enderland

r maybe adblock hti it

user55340

its flat - not empty.

Jimmy Hoffa

@ThomasOwens Agreed, I'm totally comfortable with it because I'm careful with my commits anyway and having done configuration management code merging is no trouble to me. Though I've seen engineers who get absolutely flabberghasted when they have to manually merge some conflicting changes

Thomas Owens

@JimmyHoffa It's not that I can't handle it, I want to be able to handle it as I'm drinking my first cup of coffee and minimize mistakes.

user55340

I used the trunk based model before - it works nicely when everything works nicely. However, inevitably we'd get partial feature commits in it making a mess when we wanted to do a release.

user55340

16:12

The alternative to partial commits in a trunk based model is long time between checkins - which can be very bad too.

Thomas Owens

And actually, that's been a big selling point - the ability to minimize mistakes.

Jimmy Hoffa

We're doing dev->main->release here, it works very well for us because we have 4 active coders on it and we are good about following our feature plans.

We create feature branches off dev for the uncertain features, but not for most things. Like if it's a large feature that we aren't certain will be done in time etc

But then with 4 active coders, the whole situation becomes relatively moot.

user55340

Here, with git we use feature > test > { release } > master. Always do feature branches unless its a single trivial commit (which goes to test).

user55340

release branches get branched from master, merged from test, and then remerged to master and test when the release goes live.

user55340

(that way feature > test doesn't get stalled while the release effort is underway)

Jimmy Hoffa

16:15

Though I've worked in the branch-per-feature model with 40+ coders where the project managers refused to allow team work (such a waste of resources when each coder can work on a feature entirely themselves; parallelism yeah!), we made it work but there was a ton of integration work with that many active branches that needed to come together for releases. Sometimes more devs makes the integrated development approach better (lots of integration bugs would have been found earlier)

user55340

If there's a danger of integration bugs, merge test into the feature branch to make sure everything is ok there.

Thomas Owens

I think the biggest thing is how good your version control tool's merge functionality is. Git is nice, Subversion is solid, and I've never had a problem with ClearCase. But I wouldn't want to deal with complicated merges with CVS.

2

Jimmy Hoffa

@MichaelT The problem is at release time they would be merging 12 different branches all into integration, they can't merge one down and up to the other 11 then test all 11, then merge another down and up then retest all 10 etc etc

Thomas Owens

With Git or ClearCase, I'd be more likely to not worry about complicated merges than with any other tool I've used.

user55340

We already merged them to test and did checks on them. The release branch is for bugs found while testing so that they get fixed there - and not blocking the test branch. Different branch policies on release and test is easier to handle than changing the branch policy on test.

Ampt

16:28

@ThomasOwens did you say SVN merging was solid? My experiences would disagree...

Thomas Owens

@Ampt Compared to CVS? Absolutely. Compared to git? No.

user41796

@Ampt - here's another chance to get some rep. Figured this would be a good one for you to chew on and tackle.

user41796

1

Q: Generate complex (non-convex) polyhedron UV mapping

I need to find an approach to the problem of generating texture maps for non-convex polyhedrons without using a design tool like Maya. Specifically, I am mapping simulation results data onto a 3D surface, and I need to write code which generates the maps. One generic approach uses a gradient co...

algorithms 3d geometry

user55340

@Ampt The key to using any tool is to be in the mindset of the tool. If every plays by svn's rules, its fine. When people start going off and doing things that svn doesn't like, it gets fussy.

user55340

The 'problem' is that svn allows you to do things that it doesn't like - it figures you know what you're doing and can handle its tantrums later.

user55340

16:37

The 'problem' with perforce is it doesn't let you do what it doesn't like... and people aren't willing to play by its rules - that makes it appear to be awkward to use. its actually superb once you realize how it wants to do things and do it that way.

Jimmy Hoffa

@GlenH7 I don't even understand what he's asking

user41796

@JimmyHoffa All the more reason for our resident academic to tackle it. ;-)

Jimmy Hoffa

Is he trying to display a 3d object? Is he trying to create a 3d model based on a statistical data set? Is he trying to get relevant coordinates for a 3d model?

user55340

Hey! two close votes... wonder where they came from.

user41796

@JimmyHoffa I think it's the middle option. I believe he's trying to create a surface / 3D display of graphed results

Jimmy Hoffa

16:40

Is he trying to come up with a set of 3d points based in color dimensions? (He only lists 2 dimensions, not sure what his third is..)

Ampt

We leared about convex shapes in my linear programming class. yay applicability.

user41796

@Ampt easy rep, there you go

Ampt

Actually this is exactly what we do in LP... hmm... but we aren't at the point where we are actually writing algorithms to do this stuff, just assessing the problems and turning inequalities into equalities and the like

what happened to those comments? was there anything useful in there?

user41796

@Ampt evil diamond mods would be my guess

user41796

No, not really. Another user said "go ask somewhere else" and I smacked that down

user41796

16:44

There was some back and forth between OP, MichaelT, and I about an earlier revision where he's just asking for a research paper on the subject

user41796

We got him to drop the resource request portion by pointing out people will cite them anyway as part of the answer.

user41796

I kind of wish that comment hadn't been deleted, but whatevers

user41796

hint: OP will accept an answer that's light on theory but heavy on scholarly references. :-)

Ampt

welp, I'll favorite it for later I guess. Might be able to answer it in a few weeks time.

I understand what he's asking for, but honestly don't know where to start

user41796

@JimmyHoffa - his profile gives a hint as to what he's likely doing: "I'm a software developer working for a very small company, which sells simulators to cast metal foundries."

Jimmy Hoffa

16:46

@GlenH7 Doesn't he just basically take all the fields (or a section of them which could be chosen based on an aggregate calculation to decide which fields are most relevant) then choose the points of their poles based on another aggregate that identifies opposition, then pick another arbitrary aggregate to choose directionality of the pole skew to make the data set fit? (with 5 fields, you find the adjacent point that this points field is most closely associated with, or least closely etc)

user41796

So I'm guessing he's showing the results of tests / measurements during an XYZ fabrication process

Jimmy Hoffa

@GlenH7 Weird, I would expect academia to be the main place you do something like a 3 dimensional model for a simple report.

user41796

@Ampt - The impression I got is he's looking for pointers on where to start optimizing how he's displaying things. I'm guessing the "standard" algorithms for that aren't terribly performant. But he probably doesn't know the right terms to search with

Jimmy Hoffa

If he's in industry, I'm not convinced he's trying to map a resulting data set onto the model

user41796

@JimmyHoffa You'd be surprised. :-)

Ampt

16:48

@GlenH7 and I'm not really sure what the "Standard" algorithms are quite yet

Jimmy Hoffa

@GlenH7 If you're trying to present information in a report of your results, a 3d model is 90% of the time overkill

user41796

Some of our clients are pretty demanding in how we manipulate and visualize the data. But we have a 3rd party toolkit we're using

Jimmy Hoffa

a 3d model is useful for doing result analysis computationally, but not by human eyes

(In my opinion...)

Ampt

@JimmyHoffa idk, all those 3 dimensional thinking skills I gained from playing halo have to be applicable somewhere

Jimmy Hoffa

I would generally rather see charts, explanations, numbers, 3d model perhaps is just the wow factor that markets you and yoru company

user41796

16:50

it can be handy when you want to quickly spot some minimums and maximums. And then drill into what cases create those. Cost may be a primary variable for example, but I may have cases where other parameters can overrule cost in certain scenarios

user41796

by giving me the map, I can eyeball those quickly. Writing an algorithm for automatic detection may be more difficult & not worth the effort

Jimmy Hoffa

@GlenH7 I could see doing the analysis of a 3d model as a person not being unreasonable, I would just think it's unreasonable to do the work of generating that from scratch (in your case you have a simple library so ok I guess it's worth it because it's very little effort)

user41796

yeah, I don't quite see his immediate need but that's not necessarily germane to the question

Jimmy Hoffa

I would rather stuff my dataset into reporting tools and tell it "generate this model and that model etc" and look at them until I find the ones that make the information most important

True.

user41796

Anyway, I thought it was an interesting question and would like to see it get some attention beyond a tumbleweed badge. :-)

Jimmy Hoffa

16:53

@GlenH7 The computational analysis of 3d models would be comparative analysis; this I could see being very valuable

Ampt

1

Q: How do I test a file format?

I am working on a project with a few file formats. Some formats are specified by .xsds, others by documentation on their respective websites, and some are custom in-house formats that have no documentation. Mwahahahaha. What's the problem? I would like to test these formats, but I'm not entirel...

unit-testing language-agnostic integration-tests file-handling

> I am working on a project with a few file formats. Some formats are specified by .xsds, others by documentation on their respective websites, and some are custom in-house formats that have no documentation. Mwahahahaha.

Jimmy Hoffa

@GlenH7 did my (very vaguely described) approach make sense? I like thinking about this stuff because it's so far outside of what I know anything about

(Want a working application? I'm your guy. Want it to have math? Shit my phone's ringing, I gotta get that...)

user41796

@JimmyHoffa I think that's the basics of the algorithm, but it can be inefficient if you're calculating things that would be covered up within the graph. Game developers have tackled this in the past in order improve performance, and I think that's where the now deleted comment suggesting gamers.SE was coming from.

Thomas Owens

17:08

@GlenH7 He cross posted to GameDev, but I'll let them worry about that. But don't get me started on their existence. There's only one thing that bothers me more than the existence of GameDev.SE.

user55340

@ThomasOwens It was one of their own suggesting the cross posting - I'm doubtful there was a mod flag on it asking for migration... Yep, let them deal with it (not that I've got any chips in the game).

Jimmy Hoffa

@ThomasOwens The fact that it hasn't been obliterated yet? In reality, what bothers you about it? (And really what is the one thing that bothers you more? that is such a loaded statement you just tossed out, it begs so many questions..)

user55340

Heh - evilness in Makefiles...

user55340

     foo = c
     prog.o : prog.$(foo)
             $(foo)$(foo) -$(foo) prog.$(foo)

user55340

> could be used to compile a C program prog.c. Since spaces before the variable value are ignored in variable assignments, the value of foo is precisely ‘c’. (Don't actually write your makefiles this way!)

Thomas Owens

17:16

@JimmyHoffa People who think that game development is special. 90% of it isn't unique. It's generally a combination of a lot of things, ranging from graphic arts to music to computer science.

Most technical problems in game development aren't unique to game development, and anyone who thinks that they are is probably wrong.

Jimmy Hoffa

@ThomasOwens Figured. I've always figured there's a lot to it, but I've never gotten anywhere near it so I can't frankly know. They just seem to have so much of their own design patterns and approaches and tools etc

enderland

my understanding is that most of the questions on gamedev.se aren't actually implementation/development questions but rather more meta questions

Jimmy Hoffa

It seems like a domain that would be far removed from the back end enterprise development I'm used to

Thomas Owens

@JimmyHoffa It probably is. But if you're doing some kind of 3D graphics computation, does it matter if it's in a game or not? Or trying to model ballistic physics of a bullet shot from a gub?

Jimmy Hoffa

Any argument that includes the formidable weapon known as the gub is a winning argument in my book

user55340

17:26

@JimmyHoffa It was a typo... it should have been a gRub. If you feed a bullet to one and then squeeze it just so, it will be fired from one end at your opponent who will be at a disadvantage from trying to figure out why you re squeezing magots like that.

Jimmy Hoffa

@MichaelT Ah yes, the old confuse and contrast approach to weaponized animalry.

user55340

@JimmyHoffa Thats where they got the idea for the game Worms.

user55340

Worms (1995 video game)

Mega DrivePC (DOS)PlayStationSNESSaturnJaguar | genre = Artillery, Strategy | modes = Single player, Multiplayer | ratings = K-A (Kids to Adults) | platforms = Amiga, CD32, Game Boy, Mega Drive, Jaguar, Mac, PC (DOS), PlayStation, Saturn, SNES | media = | requirements = | input = Mouse, keyboard, gamepad }} Worms is an artillery strategy video game developed by Team17 and released in . It is the first game in the Worms series of video games and was initially only available for the Amiga. Later it was ported to other platforms. Worms is a turn based g...

user55340

Granted, they had to clean it up from the original implementation and give the worms guns. They didn't like the version where they shot bullets from one end of the alimentary canal or the other.

Jimmy Hoffa

This is Thomas Owens fault

6 mins ago, 4 minutes total – 6 messages, 2 users, 0 stars

Bookmarked 10 secs ago by Jimmy Hoffa

user55340

17:32

@JimmyHoffa You should have started it one message earlier.

Ampt

@MichaelT Isn't this an example of DRY?

Thomas Owens

>_<

user55340

@Ampt The only thing that would have been better is if the source file was also named 'c'.

Ampt

foo = c
$(foo).o : $(foo).$(foo)
         $(foo)$(foo) -$(foo) $(foo).$(foo)

gnat

recent link-only answer having length 149, rare bird. Marginal answers query finds only 12 posts shorter than that (all-time at Programmers, just 12)

0

A: Easy to use JSON Web Service Hosts?

Have you seen the Google App Engine? Sample code is available at this link

Ampt

17:41

CRIKEY, Look at the size of that one! What a beaut!

gnat

though the question probably deserves answers like that "hopefully maybe someone can suggest something?"

Blog - Stack Exchange: Q&A is Hard, Let’s Go Shopping!

Jeff Atwood on November 23, 2010

Over the last 2.5 years, we’ve identified a few problematic classes of questions that tend to get asked on our sites. Many of these are documented in our standard set of close reasons: exact duplicate, off-topic, subjective and argumentative, not a real question, and too localized.

However, as we launched the great Super User experiment, a new, previously unknown class of problematic questions emerged — the shopping recommendation.

That is, on Super User we began encountering questions like: …

user55340

@ThomasOwens I'm curious - what was the order of events that lead to community taking ownership of the deletion of programmers.stackexchange.com/a/212214/40980

user55340

I noticed the user is now a nonuser... was it 'deleted answer, deleted self'?' at which point community took ownership of the deletion.

gnat

@MichaelT my guess is, this was flagged NAA / Offensive for "If you coded trash" and then dealt with by 20Kers

Jimmy Hoffa

@gnat The most striking fact to be gleaned from this blog post: Jeff still has his complete barbie set from childhood. What else don't we know about Jeff...

3

jozefg

17:50

@JimmyHoffa I've come to an interesting conclusion about teaching fp

Jimmy Hoffa

@jozefg Run? Hide? Pretend they're always after you and they can't see you under the blankets?

Oh wait, that's a totally different topic, do go on

jozefg

@JimmyHoffa Haskell is... not the best for learning fp (oh god that hurt to say). SML is better

For teaching the basics, not doing advanced applications or real world stuff

Jimmy Hoffa

@jozefg I've been wondering about that

I think the biggest problem is the community. It's not a problem for Haskell, I love it, but it's a problem for someone new to FP: That the vast majority of the available texts on the topic are beyond arcane to the uninitiated

jozefg

@JimmyHoffa Yeah... And I'm slightly guilty of doing this when I write about haskell

Plus GHCi does weird stuff sometimes.. Haskell's defaulting semantics are... interesting

user55340

@gnat NAA maybe... though I would argue that they tried. Offensive, I've seen much worse (in code and spoken) and that doesn't offend my sensibilities.

Jimmy Hoffa

17:54

I whinge every time I see someone link this to explain to people that FP isn't just arcane.

jozefg

I actually just saw hughes speak a few days ago @JimmyHoffa

Jimmy Hoffa

Good for you, I don't disagree with the text, but to the uninitiated, it's just another jargon text that proves the point they started out with: That FP is for academics, not useful for doing real work

jozefg

Yeah... I like SPJs talk about XMonad. I usually recommend that to people

It's not as exhaustive but it makes sense darn it

Thomas Owens

@MichaelT Hm. I'm not sure. It could be a user was destroyed, but a user account isn't usually destroyed unless they exist only for spam.

gnat

@MichaelT well once a non-Other flag is cast, it gets to a queue where 10/20kers can handle it however they wish, so it makes little sense reasoning how would trained moderator handle it. That's why I try to abstain from flags queue in 10K tools myself :)

Jimmy Hoffa

17:59

Either way, I could definitely see your point with SML. I think ML is important to start with, though most even among the functional circles would suggest LISP, that so rapidly just becomes a wander into list processing and macro expansion which is all neat stuff, but ML initiates you into recursion and type construction from which you can understand the list so much better anyway

user55340

@ThomasOwens as I said, I'm curious. It seems a bit of an over reaction if the user did it to himself for two downvotes. It seems a bit harsh if it came in via flags.

jozefg

Yeah, plus ML is essentially STLC with recursion, very theoretically sound

Jimmy Hoffa

but Haskell much though I like telling people to start their and just bash their head against that rock, when every other article is using terminology even PhD math students aren't familiar with, and moving through concepts so fast as you can when you see abstractions from such a high level as Haskellers, it doesn't take the time to make you see the fundamentals

Thomas Owens

@MichaelT There's no way a user can do that to himself and even if the account was deleted, it usually doesn't lead to destruction of content.

user55340

@ThomasOwens If the user deleted the answer before deleting himself, would the deletion by owner become owned by community?

2

Jimmy Hoffa

18:03

@jozefg Yeah, Haskell made a lot more sense after I forced myself to grok combinatoric calculus, but reading about Haskell didn't give me that fundamental comprehension because they glossed over it, it took a fair bit of time of me working through some concepts with simpler untyped combinatoric calculus before I began to see where Haskell was really rooted. Learning SML would have given me that fundamental comprehension far more quickly

jozefg

@JimmyHoffa exactly, Maybe this real world ocaml book will help..

Jimmy Hoffa

Though there's a problem with pushing people towards SML; to my knowledge it lacks a robust community, a robust library landscape, two other things which are important to learners. Someone learning a new language is more often than not going to want to start putting some project together using networking libraries and what have you, and finding texts will be harder...

@jozefg did my answer describing how to abstract continuations seem right to you?

jozefg

Yeah... SML has one killer app of a book: purely functional datastructures

Jimmy Hoffa

Very good point

jozefg

@JimmyHoffa Yeah, it does a good job of highlighting some of the perks of more structured continuations

Jimmy Hoffa

18:11

Perhaps regarding the library problem, people could be pushed to learn on SML.NET even though it's a dead project, it supposedly yielded good results, and .NET 2.0 though old is still a robust library to work from

though I haven't touched it myself so perhaps not

jozefg

SML.NET birthed F# iirc

And F# grew 20 features on top of SML's nice simple semantics..

Jimmy Hoffa

It did, however F# is way too imperative for teaching, and half the training stuff talks about it's cool asynchronous computations and shows "this is how you create your .NET objects"

for the most part people writing F# are just writing a more terse C# from what I've seen (F# even has first class loops as a part of the language)

jozefg

So does OCaml, though I've never actually seen someone write one without being hit over the head by a functional programmer

Jimmy Hoffa

and it birthed it in so far as they killed SML.NET because they realized they couldn't create the high level Haskell-like abstractions on the CLR, so they decided to create OCaml instead

@jozefg Yes well, functional programmers don't write F# so people are pushed towards using loops in it etc

jozefg

@JimmyHoffa Shh, if we critize F# too loudly Jon Harrop will come eat our faces.

Jimmy Hoffa

18:16

also F# has an in-line type checker and can't construct recursive types

the last part there being a killer for teaching, you can't define data List a = Cons a (List a), you're never going to teach proper FP without that

and the in-line type checker would just annoy the shit out of a learner. You have to reorder the files in your project so that they are used in-order which also means you can't have any cycles in the module dependency graph

jozefg

Wat.. eww why?

Jimmy Hoffa

and the entry point on an executable has to come at the absolute bottom of the last file

jozefg

Yeah I'm never learning F#...

user55340

@JimmyHoffa Similar to pascal days there... probably for similar reasons.

jozefg

Hurray i'm the top haskell answerer of the month on SO

user55340

18:21

All types and methods that are ultimately used by the entry point have to be declared prior to their use in the file.

enderland

arstechnica.com/information-technology/2013/09/…

Jimmy Hoffa

@jozefg It feels like they started out with F# gung-ho, then at some point realized almost no .NET developer would ever pick it up anyway, and just decided they didn't care about the fundamentals for a decent language, so they didn't bother making the type checker multi-pass and left it in-line; one definition it hasn't already parsed? Error and die. And then they went on to add a bunch of weird things like their asynchronous computations, and computation expressions for metaprogramming

user55340

One cannot have method A call method B and method B call method A - such cycles are problematic. Strict enforcement of 'must be declared prior to use' makes that cycle impossible and that problem goes away.

Jimmy Hoffa

It became their own research project for fun after they realized C# would still be used far far more than it regardless of if they got the fundamentals right or not. Alternatively maybe they never cared and were just putting it together for their own research junk

@MichaelT I get that, but it's so unbelievably counter to functional programming to restrict that...

user55340

I'm trying to remember if thats allowed in ML... and my brain goes blank. I must be blocking the ML memories.

Jimmy Hoffa

18:23

they let you do it if you put the rec keyword on the functions like OCaml

user55340

@JimmyHoffa Thats for recursive though?

Jimmy Hoffa

I don't think SML requires rec though it is eager

yes, which is what you just said

jozefg

With SML mutually recursive functions are joined with and

Jimmy Hoffa

@jozefg That's right, F# has this too

user55340

Apparently, you can in ML...

user55340

18:24

mutual recursion (forward declarations)

  fun take(L) =
     (* picks off every other list elt, starting with first *)
     if L=nil then nil else hd(L)::skip(tl(L))
  and skip(L) =
     (* picks off every other list elt, skips first *)
     if L=nil then nil else take(tl(L)) ;

Jimmy Hoffa

singular recursion is rec, mutual is rec and

jozefg

   fun someFun1 n = someFun2 (n-1)
   and someFun2 n = someFun1 (n+1)

Jimmy Hoffa

but given that, you don't need to use an in-line type checker to stop cyclical dependencies

so the in-line type checker is just laziness

@jozefg You're top SO Haskeller for the month, you don't need to worry about this now; Jon Harrop is already sharpening his knife for you on that merit alone.

congrats btw, that's a pretty hard badge to get; some really damn smart people trawling that tag.

jozefg

It's true, I should start sleeping with immutable, monad based pistol beneath my pillow @JimmyHoffa

user55340

@jozefg An immutable pistol cannot change its state. All you can do is clobber someone with it (it can't be clobbered). You're better off with a katana or the like.

Jimmy Hoffa

18:28

haha I made it on that leaderboard (the last position) for the month, I guess it doesn't mean as much as I thought ;P

user55340

Later we'll dress up like Big Oil thugs and jump Ralph Nader.

jozefg

@MichaelT Of course, it simply copies itself and then I throw the extra copy at the intruder

It's murder on the universes garbage collector though

user55340

See? Stallman - lisp fan... immutable katana.

Ampt

I could see Linus beating someone to death with nunchucks... or a semi colon. He seems like someone with some anger issues.

user55340

@Ampt I actually found a tame Linus quote when doing that signing git commits answer.

user55340

18:31

> Signing each commit is totally stupid. It just means that you automate it, and you make the signature worth less. It also doesn't add any real value, since the way the git DAG-chain of SHA1's work, you only ever need one signature to make all the commits reachable from that one be effectively covered by that one. So signing each commit is simply missing the point.

Ampt

linkify me capn edit: Nevermind

user55340

4

A: What are the advantages and disadvantages of cryptographically signing commits and tags in Git?

(This is largely based on A Git Horror Story: Repository Integrity With Signed Commits - a very good read and more information than I could put into an answer) There are a number of ways in which a git repository can be compromised (this isn't a security flaw, just a fact of life - one should no...

user55340

The link to the source of the quote (much more to it than that) is in there.

Ampt

wow that last line is not what I expected

assumed it would more along the lines of

> So signing each commit is showing how f*cking stupid you are, you ignorant twit.

user55340

The person had a reasonably thought out question - it wasn't Just Wrong.

user55340

18:33

It was misguided in his eyes, but not horrendously so.

Ampt

I just realized. Linus is the Gordan Ramsay of the software world

user55340

I worked with a guy who was the web sysadmin many years ago... He'd answer "no" before you finished the question... he left to go become a professional chef and came back ~3 years later after realizing thats even more stressful than sysadmin. Much more laid back too - he'd at least listen to the justification before saying 'no'.

Ampt

Anyone want to go to my Android Class for me today?

I think today's topic is... Fragments. Yep. Fragments.

user55340

@Ampt Thats easy. You take an android, and a sledgehammer... and you've got fragments.

user55340

@gnat only one typo? Wow.

Ampt

18:44

He did mention an interactive demo...

user55340

@Ampt See... you should go. And then express your disappointment at the end of class when the device is still in one piece.

Ampt

oh I'm going either way. Need them credits, plus it's actually a pretty interesting course and the prof knows a lot about usability

he did a lot of time building medical apps with some certain medical colleges around here

very down to earth, knowledgable guy.

user20683

Who was asking about Buddhism?

user55340

@Ampt Still say you were disappointed the demonstration didn't involve a sledgehammer and result in 'fragments'. See if you get a smirk.

This is Thomas Owens fault

Transcript for