The Whiteboard

user41796

2:28 AM

@WorldEngineer - a whole 7 minutes to shut that one down? You're not going to win any mod-twitchy-finger contests that way.

user20683

@GlenH7 chatting with girlfriend is a hardware interrupt.

2

user41796

I won't make any comments regarding performance then….

gnat

2:40 PM

free spam flags...

-4

A: Mono is frequently used to say "Yes, .NET is cross-platform". How valid is that claim?

Mono is as cross-platform as it helps you run C# code on the main Operating Systems. The problem with Mono is that it does not provide a productive cross-platform UI, which is however valid not only for Mono but for the Microsoft .NET implementation as well. That is why .NET is not Java yet - C#...

user55340

3:47 PM

That assembly speaker question, until he give the info about the machine I wanted to answer it with:

user55340

10 FOR L = 770 TO 790: READ V: POKE L,V: NEXT L
20 DATA 173,48,192,136,208,5,206,1,3,240,9
30 DATA 202,208,245,174,0,3,76,2,3,96
40 POKE 768,F: POKE 769,D: CALL 770

user55340

(which is how to do it on the Apple II)

user55340

4:23 PM

@gnat hmm... wonder how that question about package stuff is going to go. I was far too committed to writing an answer when I scrolled back up and saw the '2 more answers' when I was about to hit submit.

gnat

@MichaelT as far as I can tell, save for a few minor sidesteps, it is a dupe of the question I referred in the answer...

47

Q: New to TDD. Should I avoid private methods now?

I'm just now learning TDD. It's my understanding that private methods are untestable and shouldn't be worried about because the public API will provide enough information for verifying an object's integrity. I've understood OOP for a while. It's my understanding that private methods make objects...

reasoning appears to be primarily based on a premise that is covered in answers to above question

user55340

Testing really wasn't a concern back in the Java 1.0.2 days. Junit wasn't written until a year later (the early days of things moving really fast).

user55340

Heh...

user55340

> Kent: "We could have said there's a new generation of testing tools, we'll let them have the field. But our commitment is to provide the simplest tool we can imagine for developers writing tests, and we accept that responsibility. We thought we were just programming on an airplane.

user55340

user55340

4:30 PM

It used to be so simple...

gnat

@MichaelT yup, design considerations put all the modifiers in the right perspective when one ignores modern approaches to testing (as it was back then). But funny thing is, it's still not a reason to drop private access completely as OP assumes and this is covered well in the "probable dupe" question. simply put, progress made in testing approaches left that old perspective intact

user55340

Private still has its uses... and I use it a lot. 99.9% of all my fields are private because other things in the package have no business accessing the fields.

user55340

Now methods... thats another story. Being able to ask about the innards or call them, that's not a big problem exposing to the rest of the package.

user55340

Another example of default being used - HashMap grepcode.com/file/repository.grepcode.com/java/root/jdk/openjdk/… has lots of package level so that HashSet can use them.

gnat

@MichaelT sure. and when you extract method purely for structure and readability, even though it is used just once, it's a 100% guarantee it can be done private without any impact on testability

user55340

4:40 PM

As I mentioned, the fields are 99.9% private. Often I have private methods too and those are tested by calling the public interfaces to them - because that is what is exposed.

user55340

But then, we're getting into unit testing philosophy there...

gnat

5:13 PM

@MichaelT philosophy-shmilosophy. :) One of my favorite exercises is to study gaps in code coverage - that is stuff that hasn't been covered by tests. Really fascinating stuff after one learns what it is about, a royal road to discover interesting problems and solutions, including inventing testing scenarios that drill into private methods...

...these very scenarios by the way often give good indication whether it's worth to compromise encapsulation by dropping private modifiers. Other interesting stuff included too, like discovering bugs in spec / design... pretty cool stuff, detective kind of work

user55340

It is great to hit the inner methods and make sure they work through package level access. It is also interesting to see what is hard/impossible to hit through the public methods.

user55340

@gnat The question there (the dress for interview question) boils down to should we migrate dups to the other site? The answer is yes in a three fold way - first, we should migrate them to make a pointer for our users that "these questions belong on this other site" and secondly to provide an answer to the asker, and lastly to provide a link to the proper dup on the other site (that will get auto cleaned up if people don't answer it in awhile).

user55340

It isn't for us to decide if it is a dup on the other site, and we can't cross its dup. The tool we do have is migration. If they feel its a bad question: why did they answer it, and they can still delete it through 10k tools (lower threshold there).

user55340

It is a valid question, it isn't entirely crap, it is on topic on the target site... why not migrate it?

gnat

@MichaelT I was only concerned by their negative reaction on that question and that it was brought there in a bad shape. It's our luck that someone there eventually recognized a dupe, and they voted accordingly. In its initial shape, as it arrived there, it had good chances for a "rejected" kind closure

workplace.stackexchange.com/posts/17002/revisions

user55340

5:29 PM

As an aside, I did follow that migration... and I think I had the first dup flag on it.

gnat

@MichaelT oh man so you are our savior there and things were indeed more or less under control. Thanks! I am all for throwing valid "outside dupes" over our cliff :)

user55340

The migration is for the asker and the community - the asker to get the question to the proper site (and hopefully properly redirect answers there in the future) and the community to know that this belongs somewhere else.

user55340

It avoids the "but this is a problem that programmers face" things that show up on meta.P.SE because its shown that it can be handled correctly on the other site.

user55340

(mwhawhawhaw!)

user55340

5:37 PM

@MichaelT not much I think, I am a newbie to assembly language also. In the middle of my study I got this idea. I need a fast grasp. Will looking forward for your sites above, thank you — user2298766 4 mins ago

user55340

He doesn't realize I was sending him down a rabbit hole of learning and discovery. It won't be a fast grasp at all.

user55340

(I will point out that the "but this is a problem that programmers face" - if it isn't on topic on the other site is impolite as it shunts the problem to another community and gives the asker a bad experience - but if the question is a dup, it is on topic on the other site and thus this shunting shouldn't be an issue...)

user55340

@gnat is the data for this 1 day? or what range?

user55340

9

A: Regarding the Stack Overflow close review queue

Below request for data has been posted per discussion with Shog in comments I think having more data about reviewers would help to better understand what goes on. Specifically, it would be useful to see sort of "histogram" showing how many reviewers performed certain amount of reviews, like 200...

user55340

(not that you would necessarily know... you might though) - it might be more interesting to look at a normalized day view (average reviews per day by people with this many reviews over the course of 1 month) rather than a sum of "people doing reviews over the course of 1 month".

gnat

5:57 PM

@MichaelT my understanding is, this is a snapshot for a particular day. There's JSBin link there "Raw Data" that is written in a language I don't know (it says Javascript), but has recognizable array of constants (var values) that seem to be used to generate histograms (via var bins)

user55340

Hmm... I still think a "average over time" would give a more complete picture. That said, they're getting lots of people who do one review, but have trouble getting them to be part of the community / taking ownership for the site.

user55340

4

Q: How to get more people doing reviews?

We're nowhere near the point of Stack Overflow with (at the time of this writing) 65.9k questions in the close queue. I've noticed an uptick in the queue of people flagging questions. In a batch of 20, I often see a question that is not an audit that has no votes on it - this is from a flag. T...

discussion review

gnat

@MichaelT I drafted a preliminary analysis there, will see how it flies...

> Sharp drop at 10-50 likely indicates that many users find it difficult to work in review queue. Drop after 250 (silver bagde) suggests that even after substantial amount of reviews, many users still fail to discover a way to work productively (filtering 1, 2, 3). About half of ~500 users who have gotten golden badge at 1000 reviews indicated an interest to continue using queue even without badges.

overall I feel like SO queue dynamics differs too much from that at Programmers. Here, one might find it difficult to get to silver / gold reviewer badge. At SO, when one knows how to filter, it's as easy as it gets

user55340

We can keep it under control for the most part, and occasionally have the evil mods come in and completely empty the queue.

gnat

@MichaelT yup, and that leaves little room for badge-driven motivation. At SO things are much different

"freaking easy golden badges..." - for comparison, Electorate badge is harder than CV Steward (assuming filtered queue), but even Electorate gets 3-4 users a day, against 1-2 for CV Steward - why? — gnat Jul 20 at 12:45

user55340

6:12 PM

Hm... 99.6k - when will they hit 100k?

gnat

6:23 PM

@MichaelT today or tomorrow I guess. Recent related activity of SE team at MSO (I saw Shog, Jarrod, Anna) may be interpreted as they would prefer to make sort of a statement / proof that they worry

user55340

6:35 PM

@gnat btw, how did you like that meta git question?

gnat

6:53 PM

@MichaelT I simply loved it. you know, low quality of git posts has been worrying me a lot...

Oct 25 at 18:27, by gnat

@MichaelT I was trying hard to abstain from commenting with something like "why so many brainless git posts at Programmers are so brainlessly upvoted"...

...and suddenly I go to meta and find an answer that addresses it, thoroughly, in a serious, impartial manner. Woo hooo

6

A: Where does my git question go?

Question is about git the tool If you are asking how specific invocations of git such as: How do I edit an incorrect commit message in Git? How to undo the last Git commit? How do I delete a Git branch both locally and remote? These are questions for Stack Overflow. The key point here is yo...

user55340

@gnat The github question asked here was what drove me to write it... and I tried to get some example questions from each site to show how what fits there.

gnat

@MichaelT I figured that. Now we've got a sensible reference to cover this stuff. I also worried that there's no reliable way for us to figure what questions should stay at Programmers, and your answer covered it, too. Well done

user55340

Apparently, we still need some more meta site recommendation Q&A out there...

user55340

Where do questions like "how to learn C++" fit? — Blessed Geek 20 hours ago

user55340

@BlessedGeek "how to learn XYZ" questions are typically too broad for any Stack Exchange site. There are textbooks upon textbooks covering the area and then you get into the sub realms of programming languages (embedded vs game vs enterprise). Stack exchange works best when there is a single problem that can be solved in a few paragraphs. If you can imagine an entire book on the subject, its probably too broad. — MichaelT 19 hours ago

user55340

7:10 PM

@GlenH7 you wanna write that one?

gnat

9

A: What should our predefined off-topic reasons be?

Questions seeking career or education advice are off topic on Programmers. They are only meaningful to the asker and do not generate lasting value for the broader programming community. Furthermore, in most cases, any answer is going to be a subjective opinion that may not take into account all t...

would that do?

user55340

I'm thinking more a really long answer that explains why "how to learn XYZ" is always a bad question. We keep closing the questions again and again. And while we have the bit in the close reason, having a few paragraphs (more than a comment's worth of text) explaining why it doesn't work.

user55340

Might actually be useful to write a Q&A for each of the standard set of off topic reasons.

enderland

@gnat I saw that. ;)

:D

user55340

I didn't see it, but I can see it, but the raw html doesn't make quite as much sense. I think.

gnat

7:22 PM

@enderland it didn't eat html, will try raw png

gee I've got enough ammo for 7 bounties on hotness formula fix. Just 2-3 weeks to wait until the dust around CV queue / top bar settles and I can get the ball rolling...

png works :)

user55340

7:41 PM

sigh things you stick in edits when you don't have 2k rep -- meta.stackoverflow.com/posts/209644/revisions

enderland

@MichaelT yeah that's my least favorite part of going below 2k rep there :P

jozefg

Writing unit tests for a compiler is a royal pain

user55340

7:56 PM

Well, now I'm at 2016, so no more easy rep from edits.

gnat

@MichaelT while you're under 3K, use this as a chance to bump "flag weight" there. Dupe flags are dead easy at MSO; I was slowly crawling to marshal badge, being like at 2/3 way there after more than a year, when I discovered this magic after getting under 3K because of bounties. From there on, flags went as easy as it gets :)

Transcript for