Comment by kragen | Hacker News Reader

Comment by kragen | Hacker News Reader

Parent Full thread

haldar's summary is completely incorrect, as you would expect from someone who says 'tl;dr'

the paper finds that faults (and wmc (methods per class), cbo (coupling between object classes), response for a class (rfc), number of methods added (nma), etc.), correlate to class size (non-comment source lines of code), not program size

it only looked at one program so it could not measure any effects due to program size

moreover it wasn't measuring code quality in the sense of 'defects per kloc', it was measuring defects (whether a bug had been detected in a class in the field or not)

stripping away the acronyms, what they found was that classes that contained more code more often had at least one bug, and also had more methods, but that having more methods without having more code didn't make classes significantly more likely to have a bug

and similarly for the other complexity metrics like how many different methods a class calls and implements (rfc)

this is unsurprising, since things like the number of lines of code in a class and the number of different methods in a class are just alternative metrics of class size, and everyone knows that in general more code means more bugs

that's why we measure code quality in defects per kloc and not total defects. the paper didn't even try to measure code quality in that sense

that doesn't mean the paper is bad. if the paper's authors are correct that many other papers have failed to control for size in their defect metrics, they have identified a serious shortcoming in the existing research literature; haldar merely totally failed to understand them

so the paper haldar tried to summarize doesn't measure either program size or code quality

(some of their references did look at program size tho)

why are none of the other comments pointing this out

are they all just commenting without having read the paper

now i am sad

>as you would expect from someone who says 'tl;dr'

...What? This is hilarious.

How is using an acronym to provide a brief summary of something indicative of correctness?

> How is using an acronym to provide a brief summary of something indicative of correctness?

I’m struggling to understand how this is the point that’s most worth debating here. What do you propose we’ll learn?

it turns out they thought they were being criticized so they got defensive, explaining their otherwise shockingly aggressive and even mendacious comment thread here: https://news.ycombinator.com/item?id=33567361

not sure that's a good excuse for their total failure to contribute anything substantive though

'tl;dr' is an abbreviation for 'too long; didn't read'

people who don't read long things sometimes know a lot about things like how to weld, how to comfort the grieving, and what makes them happy, but they are always profoundly ignorant when it comes to book learning

that's what happened in this case, where haldar's summary claims that this paper shows that a variable they didn't attempt to measure is the best predictor of another variable they didn't attempt to measure

evidently he just looked at the diagrams and guessed what the words meant

>people who don't read long things sometimes [...] but they are always profoundly ignorant when it comes to book learning

You seem to have fundamentally misunderstood how this acronym is used.

When provided by an author, it doesn't mean that the author didn't read... It's literally just a substitute for "Summary:". An indication to the reader that if they find the article too long to read, they can read the immediately following subsection to get a summary of the article.

Do you also think that people who say "Summary:" are profoundly ignorant? Or are you just against acronyms?

probably you haven't noticed this, but using 'tl;dr' instead of 'summary' or 'abstract' is a group identity marker by which the author shows solidarity with their (presumably profoundly ignorant) readers. it's the same sort of gesture as describing the paper's authors as 'eggheads' or 'boffins'

and evidently in this case the blog post author really didn't read the paper they were 'summarizing'

i find your contributions so far to the discussion of this paper profoundly disappointing

>probably you haven't noticed this

>their (presumably profoundly ignorant)

>i find your contributions so far to the discussion of this paper profoundly disappointing

I'm not sure why you are so hostile, and why you think everyone except yourself is profoundly ignorant, but it's clear that you have some sort of world view where you're superior than anyone else you interact with. I doubt anything I say would convince you otherwise.

tl;dr: I found this conversation disappointing, too.

i frequently interact with people i admire and who know more than i do (and sometimes they are the same people), and, as i said above, even people who are profoundly ignorant about book learning are often wise about many other things

rather than being hostile, i have taken the time to explain in more detail the things you asked about in my previous comments because they were unclear to you

i am surprised that you have stooped to launching personal attacks on me in this comment, since so far i have criticized only your low-quality comments (and the original blog post) rather than any presumed personal attributes of yours

try to do better please

>i am surprised that you have stooped to launching personal attacks on me in this comment, since so far i have criticized only your low-quality comments (and the original blog post) rather than any presumed personal attributes of yours

You've called me "profoundly ignorant" several times in this chain. How you can say this with a straight face is baffling.

i haven't called you profoundly ignorant even once. please do not lie about what i have said; there is no point in doing so in any case because anyone reading this thread can look half a page up or click the 'context' link and see that you are lying

i said that people who don't read long things are profoundly ignorant, at least within the sphere of book learning, though often not in other ways. that's not a personal attack on them, it's just an obviously true statement if the phrase 'book learning' refers to anything at all

you haven't said even once that you don't read long things, so there was no reason for me to even suspect that what i said applies to you, even if it were a personal attack (and, well, some people are sensitive about being ignorant because they're used to being criticized for it)

From now on I will strive to only write "abstract" instead of "TL;DR" so I can avoid showing solidarity with the presumed profoundly ignorant. It's a matter of principle that I must appear to be the most intelligent person in any room I enter, virtual or physical.

maybe you could show solidarity with them in a way that doesn't reinforce a dichotomy between them and people who read books

for example, by emphasizing the praiseworthy attributes you have in common rather than the misfortunes that have befallen you and the self-destructive choices which, in cases like this one, perpetuate those misfortunes

i mean it might be useful to read a book from time to time, or at least a 25-page article, and that's less likely if you turn it into an identity threat

we're all born profoundly ignorant but we don't have to stay that way

if you're the most intelligent person in the room maybe you're in the wrong room. this can easily be reversed to provide a plan of action if being the most intelligent person in the room is what you value most. beware, though, some people are smarter than they look

Some of us know how to weld, how to comfort the grieving, and have good book learning. Still trying to find the balance between happiness, satisfaction, and pleasure.

agreed, i'm not trying to say it's a zero-sum sort of thing, just that using 'tl;dr' is not an indication of profound ignorance in all possible ways, just the forms of ignorance that result from not reading long things

Reply on news.ycombinator.com