The Whiteboard

5:22 AM

@Shog9 we started protecting questions featured at Ars Technica about half year ago, after we discovered that these get polluted with low quality content. If you think we were doing it wrong, I would like to understand why and what we should have been doing instead

Shog9

@gnat Protect when there's a problem, not when one is anticipated.

gnat

@Shog9 there is a problem, as I said we started protecting after we tested that unprotected featured questions are polluted. Here is a proof by the way, just in that recently featured question:

-1

A: Why should passwords be encrypted if they are being stored in a secure database?

I think there's a fundamental flaw in the question. The fact is, there's no such thing as a "secure database". It should be taken as a given that there are flaws in your system somewhere that could allow malicious users access to arbitrary data on your servers. Heartbleed is an excellent case in ...

@MichaelT auto-protect has been already tested at Workplace and at Programmers. So far so good

Shog9

@gnat Look, I have a huge backlog of stuff to get through yet tonight - I'm not gonna re-hash this same "all questions are equal, all attention is bad" argument again. Protect isn't magic pixie dust that makes all the bad things go away; it's intended for one specific purpose - please reserve it for that.

gnat

5:43 AM

@Shog9 all right I will open meta discussion to re-evaluate approach we were successfully using for half year before you noticed. Meanwhile I protected that particular question back, following guidance given by Jeff: "attracting a lot of drive-by noise answers" - per my reading 5 of 15 answers are like that

Blog - Stack Exchange: New Protected Question Status

Jeff Atwood on June 06, 2010

At the behest of the Super User moderators, we’ve instituted a new question status of protected. A protected question is like a protected wikipedia article — it no longer allows additions by anonymous users.

Protected questions are indicated in the standard question footer like so:

This question is protected to prevent “thanks”, “me too!”, and spam posts from new users. To answer it, you must have more than 10 reputation.

We needed this because some of the more popular Super User questions attracted a lot of noise from random drive-by users who didn’t und …

gnat

11:53 AM

0

Q: Questions featured at Ars Technica

Per my understanding, for a while there was a sort of informal agreement to protect questions that are going to be featured at Ars Technica. Recently this approach has been challenged, so I would like to have a dedicated meta discussion to sort this out. Does it make sense to protect such questi...

@Shog9 ^^^ there you go - feel free to chime in

^^^ @MichaelT @GlenH7 - I would appreciate if you take a look too

MetaFight

12:52 PM

So I was recently told in an interview that the company in question didn't use ORMs for performance reasons. I was reluctant to bring it up, but aren't those concerns pretty much outdated? These guys were using SQL Server 2012. Add EF on top of that and, in most cases, I'd argue performance would be just as good.

they did mention, however, that some of their stored procs required temp tables to boost query performance. I guess that's an instance where a stored proc is still just fine (or even required). But even then, you can still invoke it from your ORM.

Have I drank the ORM koolaid, or is there really still an argument for using Stored Procs and hand written DALs?

(for extra context, this was a .Net technology stack company. SQL Server 2012, Windows Server, .Net apps)

Tom W

there are probably more voices of dissent against ORMs than you think

we're not using one on our (major, public sector, complex) project

MetaFight

Also, I ask because according to them I would have been a great fit if it hadn't been for my weakness in DB optimization. Admittedly that's an area I don't focus on... but because I thought it was less and less valuable.

@TomW Was that an informed decision, though?

Tom W

the way I've seen it put a few different places is that ORMs don't really save you any time at all, because you spend just as much time banging your head against the configuration and mapping as you would against handcrafting DALs

yeah, these people know what they're doing. Nobody could say they're not opinionated, but I think they've experienced ORMs not really making anything easier at all

Johan Larsson

I'm so glad I have no db in my daily grind

MetaFight

I found EF to be a pain in the ass the first time I [mis]used it. After that, it was all easy sailing.

Tom W

12:59 PM

as for efficiency - dunno, really

I suppose if you're dragging enough columns out of a DB to make writing data access tedious, then that's the bottleneck, not ORM performance

MetaFight

I've always seen it as a great tool in Agile setups as well. When Data requirements are changing often, ORMs typically let you adapt faster than traditional Stored Procs. And once your requirements have settled you can always switch to a hand written DAL.

(adapt faster, and with lower risk)

gideon

2:05 PM

hey guys, I'm wondering what are the merits of jumping from learnng one language to another as a hobby. My Case: I picked up perl about a year ago, mostly to get hired somewhere. It didn't work out, but all in all I made a few web apps, contributed to a web framework and even made a small doc fix inside perl code ;) I want to now move to Python. I've used it before but I want to build web apps with it.

And so I'm caught in a dilemma of pursue perl, but perl is a little crazy! OR study python

Johan Larsson

I have not learned much F# but thinking about functional stuff has made me better at C#

Python is a language that I rarely hear anything bad about.

Haskell, C# & F# also get little hate.

gideon

@JohanLarsson yea I'm seeing how C# gets some hate. I've been using it since I was 16 so most of my professional experience is there.

@JohanLarsson I have looked into Haskell a bit also. But I was thinking of making python my goto language, now that I have mostly *nix systems.

but is it a bad thing to sort of leave Perl behind, lately I've been hearing too much that I need to be a master with one and not a jack of all trades.

Johan Larsson

I have no idea how useful perl is but I imagine python is more useful

Tom W

why must you do one or the other?

ratchet freak

writing perl is easy; maintaining cowboy perl is impossible

gideon

2:19 PM

@TomW to eventually try and move into a linux/unix shop.

and I also hobby/hack at home and Perl is ok but it can be really funky, and python seems neat and tidy for the most part.

Tom W

I mean why do you need to choose only one?

you posed the question as perl OR python

gideon

@TomW well I know Perl, and now I feel like moving on to python. I'll know both then. So I can choose whichever for the job. The question though is: Is sticking to one thing for a while better than jumping from one language to another ?

Tom W

I think there are arguments for depth and breadth; I'd never say to anyone 'don't study language x because it might distract you from y'

gideon

Hmm. I see. @TomW I've always believed in a sort of checklist for that language, so for perl I developed things, I contributed to OSS projects etc etc. So I crossed everything on my perl checklist. And now I'm making a python checklist.

Tom W

I guess that's quite a good way to make sure you keep progressing

gideon

2:32 PM

yep. But I've got like heat from a couple people lately telling me I'm a jack of all trades etc etc.

So I was sort of rethinking everything.

Johan Larsson

on the other hand many programmers are jack of no trades :)

deuces?

Tom W

those people are tools.

learning lots of languages is great

helps you to see past the fiddly bits in the language and look at concepts

gideon

haha true. @JohanLarsson I've kinda always wanted to understand how everything ticks. Which is why I also studied the linux kernel, sent in patches, and then even did a quick stint with micro controllers and electronics.

But with all the rethinking I'm doing I

wondering how useful/useless it is when trying to get hired, that I do a wide spread of things like this..

:/

Johan Larsson

how is the job market where you live?

I'd say knowing enough to get a job and then go for mastery while being paid sounds pretty sweet.

gideon

It's not all that great, I'm currently a tech lead in a C# Product Shop. But down here an attitude for good work and learning is never there. I mean, even telling my developers to stick to a coding style is hard.

I mostly want to move a company that uses *nix systems. I'm seeing that all the hackers are in the *nix world.

but since 2013 I've had no luck!

gnat

3:10 PM

how strong is our adherence to Don't Repeat Others rule? I made a feature request at TWP meta where it's enforced rather strongly, now wonder if we can make sense of it at Programmers...

3

Q: When there are many answers already, help me check that mine won't repeat others

When posting my answer I check that it doesn't repeat others as required by our FAQ. In current UI, this becomes difficult when question already has many answers. I would want that questions with 10 or more answers have additional UI allowing me to review each of previous answers individually p...

feature-request answers

Johan Larsson

nice, I'd +1 it bt to lazy to create account :)

Robert Harvey

5:38 PM

What prevents you from using the existing ui to review previously posted answers?

Johan Larsson

6:02 PM

nothing but wall of text is tl; dr;

Robert Harvey

7:03 PM

Some pretty good question on p.se this morning.

gnat

8:01 PM

@RobertHarvey imagine you want to add an answer to this question (16 answers), how would you check? Regular UI shows "wall of text" there indeed

Robert Harvey

8:29 PM

@gnat Why does the question have 16 answers in the first place? Does it need to be closed?

Shog9

8:55 PM

@MichaelT, you might find the numbers here interesting:

1

A: Questions featured at Ars Technica

Protection racket Your own list illustrates how pointless this sort of proactive activity is: in most cases, few or no new answers were posted regardless of whether the questions were protected. Indeed, the most-answered question of the lot was immediately protected. It's not the raw traffic t...

That covers a pretty large % of the entire traffic from Ars over the past 2 years; the most striking aspect is just how little activity of any kind it generates on the site.

user55340

11:23 PM

@Shog9 Numbers are always interesting. From the point of an active community non-moderator moderator, we've got so little insight into whats going on, we need to act on hunches sometimes.

user55340

The 'worst' was when we noticed it with (what felt like) a deluge of poor answers on one question (though admiralty my memory may be wrong there).

Shog9

@MichaelT Would like to add more information to the 10K tools; if you think of something that'd be broadly-useful, don't hesitate to suggest it.

user55340

I'm actually missing the 'anon and low rep feedback'. I realize that was a huge drain on SO's with its queries... and I've gone and poked at data explorer... but that doesn't give me the last day of "5 people hit this from google and thinks it sucks in the past 24h"

user55340

Maybe tossing that anon feedback into a activity panel of the 10k tools instead (with all the associated cacheing - don't need up to the minute results there)

Shog9

@MichaelT I suspect the problem was that so few people used it... Makes caching difficult. But I would definitely like to see it back in some form - perhaps as "most anon feedback in the past N days" lists on the stats tab.

user55340

11:33 PM

@Shog9 Highest and lowest - just like the regular votes.

Shog9

right

user55340

And yea... I'm not surprised that few people used it given how long this bug probably existed before I caught it.

user55340

6

Q: Navigating pages in 10k post feedback clears date range filter

Note: post feedback page has been removed and will no longer be available Under post feedback (anonymous and low rep post feedback. last of the links on the tools page), the filters (underrated, overrated, most helpful, least helpful) and the date range (day, week, month, year, all) are not ...

bug status-completed 10k-tools post-feedback

user55340

another traffic bit that would be useful would be instead of (in addition to) 'visits in time range' - 'non-SE referring page visits' so that we have a chance on getting a heads up on something from HN or Reddit rather than just "unusual activity"

Shog9

@MichaelT not sure how feasible that is. Outside of referral links, we don't track views by referrer.

(and the vast majority of all views are from Google, so...)

Point being, I don't have that information, much less the mods, much less anything we could put into 10K tools.

user55340

11:41 PM

As a "hmm, I wonder if its useful" I'm not so much interested in that HN or Reddit or the name of the blog that did it, but rather "there are 100 views to this question that came from a referring page other than Stack Exchange."

user55340

No matter which way someone came, be it google or HN... there's still something thats drawing people to that question for some reason. It would be great if someone is getting a publicist badge to go with it... but the unusual additional views is the "hmm" part that I'm curious about.

user55340

And yea, many times with this the numbers are just pieces of the bigger puzzle of "what draws the eyeballs here" and we're just wondering whats going to be the next thing.

user55340

Sometimes you can just feel that a question is going to get popular (and thats the internal referring thing that we've got an idea of how to watch).

user55340

But the external referring visits don't always show up on the hot list (and they can be on quite old things).

Shog9

Referral-link stats + anon votes might get us 90% of the way there.

user55340

11:46 PM

Quite possibly. We're just working with what tools we have - theres two parts to that: the information on what actually merits the use of the tools (the 10k stats), and the tools themselves (protection, close votes, flag a mod for locking). Some of those are rather blunt and corse in the visibility.

user55340

You see something and find the only thing you can hit it with is protection of the question... it may be a bit too blunt of a tool... or we just can't see that the problem isn't as big as we think it is because of the coarseness of the information tools.

user55340

Things like "someone bumps an old question that resulted in a bit of activity on it 2 years ago with another answer... we can't unbump it... so to prevent another bunch of low quality answers on it, protect it... (for a day or two until it doesn't sit on the front page tempting every new user to answer it with their $0.02).

user55340

I'd so be using a "protect for 24h" or "protect for 48h" if it was available for those questions.

Shog9

My issue with the use of Protection is that it's actually not a very big hammer; more like a stick-on patch, really. So folks see a problem, and they go looking for a solution, and pick Protect... And then wonder why the question is still getting crap answers. Well, shucks, you blocked up a mouse-hole but the elephant in the room is still stomping on everything.

user55340

Its not that they should be closed (though sometimes that happens next...) but without someone posting an epicly good answer to set the expectations, an answer on the poor side of 'meh' draws more.

Shog9

11:52 PM

Editing, close, lock, delete... In not necessarily that order... Are gonna be far more effective outside of scenarios where something is legitimately a honeypot for unwanted outsider attention.

user55340

Protection certainly isn't a big hammer. We don't always have the close votes to close it without a mod (I've occasionally come across a question were I had an expired close vote from December or so... and asking for other close votes find out that Gnat and Glen also have expired votes... and well, thats most of the close voters in the review queue...

user55340

When I had the anon feedback I would go edit the questions that I found in there (or close vote them).

user55340

Finding 4x "not what I'm looking for" on a low rep feedback is very helpful to go in and look at and say "yep, thats a misleading title for its question..."

Shog9

Good point

user55340

If you look at my stats you will certainly see that I'm a down voter. At times I wonder if it would be useful to expose more of the 'anon feedback' and allow (down) voters to participate in that.

Transcript for