@mustaccio Hmmmm... I was taken in by our friends with opposable thumbs - I created a tag for DolphinDB (since vanished). Maybe the unorthodox use of English should have been a heads up - their thumbs may be opposable but but their grammar and syntax isn't (maybe they can click well?)...

@AaronBertrand I took a moment to reflect on the question and answer I mentioned in the discussion about having most upvoted vs. accepted answer at the top of the stack. I still stand by my opinion that some people with outstanding reputation tend to accumulate more upvotes that other people on the site. I don't see any insult in my statement as it is just an observation on my part.

I would object to be put into the same box as J.D. as I do admire your achievements and outstanding reputation.

This is equally valid for all the other outstanding members inside our community that tend to garner more unicorn points than others.

I'd like to thank you for contacting me on the issue you had with my statement and I am looking forward to participating in further discussions in the future.

@AaronBertrand Please accept my sincere apology for having inadvertently insulted you.

Morning

Morning

Does anybody have a good RTO / Availability calculator at hand?

Or (in your opinion) is there no relationship between RTO and availability when a database goes down?

@JohnK.N. I'm not offended to the point you need to apologize, but thanks. What I'm trying to get at is that for this specific question you are put off that my answer got more upvotes than yours, and seem to think that your answer is "better" and therefore deserves more upvotes. I'm not sure it's up to you or me to declare one answer better than another because of the same subjective reasons I brought up before.

But I still think there are observable reasons why this might be the case that don't involve my name or my reputation. (a) Extended Events is a more recognizable feature; (b) as such, there are a lot more articles out there about how to consume the results; and, (c) audit just implements Extended Events under the covers, so how superior is audit really? That it deprecates C2 doesn't have bearing on its relationship with XE.

My reaction was not about the fact that gravity of name/reputation does influence some behavior... I've already acknowledged that is probably true. It was that it might be the only reason any of the upvotes happened because obviously my answer was inferior...

In the real world, we recently had to choose how to implement a DDL collection mechanism, and we chose Extended Events over Audit because of the reasons mentioned above ... also we were already using Extended Events in a framework for collecting data (perf/errors) that can't be caught with Audit. So my familiarity is there for both tangible and intangible reasons, and if a similar question came up tomorrow, my answer would involve XE again.

Audit can also do the job in this case, but there is no one objective thing that makes it an obviously better choice.

@CadeRoux Did you come to any further conclusions?

@AaronBertrand Nope. I don't care about the points. I'm just discussing the pinned accepted answer vs. floating upvoted answers feature.

:-)

Right. But as I've said before, there is no one solution that handles all cases. I think there are better examples than the one you happened to choose here that would illustrate your point better, but that doesn't make those examples magically apply to all examples.

I think the new approach absolutely improves the quality of the content in some cases. No way in hell I'm going to suggest most, never mind all.

And the impact of the change will also vary by site volume (or tag niche-ness on larger sites).

It's a different way to do things. I think the intention is good and after fighting fruitlessly for much more minor changes and eventually giving up on meta as a whole, I admire their ambition. Is it going to get things wrong on some questions? Sure. Show me a solution that you think gets it right in all cases (the existing way certainly isn't that), and I guarantee I can cherry-pick counterexamples.

I think what I'm taking from your stance is this: "This change is going to push poorer answers to the top, because some famous person has more ill-gotten upvotes than the answer the OP chose."

While I concede that, sure, that exact thing will happen, I would wager that more often it will promote a better (and sometimes much later) answer, in terms of value to the average reader (and not the OP, who may be long gone).

Even if it is 50/50, are we worse off with the change? Probably only if you disfavor change itself.

The purpose of the site is not just to solve the OP's issue, though, otherwise why let anyone else but the OP vote at all? The votes can help the OP decide because they're asking a question they trust the broader community to know more about. But the votes also help future readers who may have better judgment than the OP, and the site is about those readers.

I would prefer the accepted answer to be at the top, b/c it was the solution for that question.

Yes, I agree that the community benefits from upvoted answers.

But I think we'll have to agree to disagree that the solution for that OP is not necessarily the solution the next reader should choose.

@AaronBertrand +1

As I mentioned previously, the OP may choose a solution because it was first, or shortest, or easiest, or because they just didn't know any better.

And people still pressure OPs to accept answers (though it used to be far worse when accept rate was a single label).

And worse, "if my answer helped, accept MY answer!"

I think the feature is designed to take away the sole, binding choice of the OP to declare this is the best answer for all future readers.

Good point.

So I think on balance the OP (who, again, asked a question because they don't know the answer) is arguably less equipped to make the right choice about what is the best solution to a given problem than practically everyone else involved.

case in point

I inadvertently taught this person about SET LANGUAGE to demonstrate how unsafe his date format is, and his "solution" is to hard-code US_ENGLISH on every query.

/facepalm

Maybe there's some elaborate way to tweak the behavior such that some OPs are trusted more than others, say on a given tag, due to reputation earned. Kind of like how I can unilaterally decide to close a question in a tag where I've earned gold.

But I think that gets confusing quite quickly and again would only be relevant for certain tags on certain sites and only for OPs who are asking questions in a tag where they are obviously proficient... so is going to be relatively uncommon.

This OP I mentioned could accept his own answer and force a really bad solution to the top. Who does that help? Especially if as few as two others agree and erase the down-votes?

^ if this feature weren't already implemented on main, I mean

@AaronBertrand Hmmm. Well case in point for upvoting other answers and not putting the accepted answer at the top.

Right, that's all I'm trying to say: neither method works universally.

But I see A LOT of crappy, accepted answers pushing down much better and higher upvoted answers, and for new users of the site, this is not great for them. I haven't been looking for it, but I haven't seen as many of the cases you're talking about, where a clearly worse answer has more upvotes than the accepted answer. Those cases do exist, and I have seen them. Just far less frequently.

The "fix" for that latter case has historically been to pressure the OP to change their accepted answer from the crappy, highly upvoted answer to the newer, better answer. So the downside of the change is that that will no longer help, because overcoming the inertia of years of votes can take a long time.

More relevant on main than here, I'd think. But we probably have different opinions about the deservedness of upvotes even here.

There's probably a more optimal solution to when to promote another answer over the "accepted" answer, but I wouldn't trust SO to implement it/think it through adequately.

There's definitely a time component to things - if another answer is getting a lot of recent upvotes (or the accepted answer a lot of recent downvotes) that might be indicative the original answer isn't as applicable as it once was.

But there's never 100% a good solution here because in the land of Stackoverflow/Stackexchange the users decide what is right, and those users don't always understand the subject fully to make a determination definitively.

@AaronBertrand It is working with with OPTION(RECOMPILE) on the main INSERT/SELECT and that makes sense as a strategy anyway since the plan for a given date may vary over time (especially if you hit a weekend). The one thing I did not do was verify that the original code had a problem on my SQL 2016 instance. Unfortunately the test source data on the 2016 instance is also possibly different enough because that's a newer version of the source system. So will take longer than maybe is worth it.

To beat the dead horse that is the XE vs Audit functionality - I'd imagine the best answer would discuss both and provide the user context so they can make an informed decision. But that is split across two answers in this case. No good way to resolve that save actually having to do work.

Also, good morning

@AaronBertrand I guess the two worrying things: the hang - that's kind of my experience that when parameter sniffing bites you enough for you to notice, it bites to kill. And the WITH RECOMPILE I tried to apply to just quickly confirm it was parameter sniffing didn't eliminate the problem, so it just made it all the more confusing as to what was happening. But like you said, it could be something more nefarious, and that's always a worry.

big ooph

@PeterVandivier in dba.se? it appears cleared to me.

SO

went in to fix a link

b-b-b-b-blocked

in case anyone with 2K on SO wants to be a good samaritan, this link needs updating from master->main stackoverflow.com/a/23522945/4709762

otherwise the query param gets stripped and breaks the demo

@PeterVandivier done

I realise sponsorship is good for the mother ship because it brings extra money, but is it really a reason to be excited? Should we be excited as well?

Well for the people who negotiated it and can apply that money to their expenses, yes. For the rest of us, only in that it's a way to ensure longevity and continued resources.

This seems v. bad.

On-prem Exchange is still a thing?

So glad not to have to deal with Exchange ever again, amazing that some still do.

+1

@PeterVandivier .

@HannahVernon oopsie doodle

I guess people don't study arithmetics in school anymore...