Tavern on the Meta

I went to write a document in LibreOffice just now, got frustrated with its random assignment of fonts, indent levels, and line spacings, opened Google Docs and nailed it, then it dawned on me how much more functional Google's weird little web-based word processor is than all of these desktop applications.

SmokeDetector

[ SmokeDetector | MS ] Pattern-matching website in answer: How to install Ubuntu on the Samsung Galaxy S3? by cjen on askubuntu.com

angussidney

9:10 AM

@JasonC if I'm just writing text, I find Google Docs the best, but if I need to do anything else with my document I find Word better

SmokeDetector

9:25 AM

[ SmokeDetector | MS ] Link at end of answer: Play video with 2x speed in MX player kindly help by opera mani on android.SE

Bart

Does anyone have experience with Adview? Is the OP here publishing personal information? stackoverflow.com/questions/42758319/…

Stijn

@Bart No experience with it, but I don't think the long numeric string is critical sensitive info

    // Create a banner's ad view with a unique placement ID (generate your own on the Facebook app settings).
    // Use different ID for each ad placement in your app.
    AdView adView = new AdView ("YOUR_PLACEMENT_ID", AdSize.BANNER_HEIGHT_50);

Bart

Okay, thanks :)

Shadow Wizard

uh oh, when @Bart is so polite it's time to worry....

:D

@JasonC they probably just spend more developers on this. (and money)

Stijn

9:41 AM

@JasonC people don't use LibreOffice because it's good (it's not), they use it because it's free

I recently read that they'll finally add an optional ribbon interface

I tried LO some years ago but couldn't deal with the old style menus

Bart

9:53 AM

@ShadowWizard /me makes note of this on his checklist ...

Journeyman Geek

@Stijn and not everyone trusts ms or the cloud

Shadow Wizard

10:16 AM

stealing @Bart's checklist

SmokeDetector

[ SmokeDetector | MS ] Manually reported question: The Riemann Hypothesis Proof and The Mutiple Zeta Functions by Faria on mathoverflow.net

M.A.R.

10:52 AM

> Were supposed
by Donald J. Trump
NEWS media lied about. Very nice!
Press Conference:
Done during my term(s) in office.
Index Surges by Most Since 1980'

Why? I will study this dumb deal!
Was hacked? Can that be possible?
Amnesty! Vote outsider Navy SEAL
Me to do the right thing! Terrible!

Failed policies. #BigLeagueTruth
And want to Get out & VOTE…
Chicago, Illinois. #BigLeagueTruth

Shinzo Abe in the United States! …
Making up stories & sources!
Picked up additional votes!

Shadow Wizard

Is this a song?

M.A.R.

@ShadowWizard Click the reply button, see where it leads ya

> Jersey underneath
by Seth Rogen
It's gonna be like Spaceballs.
I'll straight up tweet that shit!
Of July! Remember: Grills Kill...s.
That game was bullshit

Expose myself to my own opinions.
As the experience itself at times
2 of for about a million reasons.
While wearing elaborate costumes.

Like it when mom was a cashier.
Nobody does like New Orleans...
Come to the premier!

Imma report your ass.
Of line leaking info to the press!
Just wait dumbass.

Derpy

11:23 AM

> I've got to find a way
To make this typescript class type-safe...
I can't believe this small mistake
Could've caused so much heartache...

SmokeDetector

11:36 AM

[ SmokeDetector | MS ] Bad keyword with email in answer, email in answer: What took Bella so long to turn into a Vampire? by lee on scifi.SE

Shadow Wizard

11:51 AM

@M.A.R. to @Jason

M.A.R.

@ShadowWizard See his message

Shadow Wizard

@M.A.R. poems (?) by @Nick

M.A.R.

@ShadowWizard It's a site thingy where you put someone's twitter ID and they make a poem out of it

Shadow Wizard

@M.A.R. very clever... open source?

M.A.R.

poetweet.com.br

Dunno, didn't check

Shadow Wizard

12:00 PM

oh, PHP

> Your twitter we coundn't find.
You have to try again, if you don't mind.

lol

So polite

Derpy

1:09 PM

.... obviously, IIS does not support subdomain wildchars in Access-Control-Allow-Origin. After all, it is pretty new... only 10 months....

random hissing noises

Stijn

How do they not support it? Do they have validation on the custom response headers you specify?

Derpy

1:28 PM

@Stijn <add name="Access-Control-Allow-Origin" value="http://*.somedomain.com:port" />

^ this was introduced by w3c 10 months ago. IIS as expected still hasn't supported it

why I need it?

short version: SharePoint is evil

long version: every time you debug an app from Visual Studio, the url changes.

Since I also have <add name="Access-Control-Allow-Credentials" value="true"/>, that disables the wildchar support for access control origin

Stijn

1:49 PM

@Derpy I'm confused, that line just tells IIS to always include that response header. Isn't whatever problem you're having caused by the server or client application?

Or does IIS actually complain about using a * in the value?

Just found out about space.stackexchange.com; what's the point of having a site in beta for nearly 4 years?

SmokeDetector

2:26 PM

[ SmokeDetector | MS ] Bad keyword with email in answer, email in answer: Is the second vaccine dose necessary if seroconversion occurred after the first dose? by Bright Gibson on health.SE

Andrew T.

@Stijn Board & Card Games is 6 years old and still in beta...

Derpy

@Stijn you cannot use the "*" "allow any origin" wildchar if you also use the "Allow credentials" switch. So, you have to specific an url for the Access-Control-Allow-Origin instead

since every time I run the debuger my app url CHANGES because Microsoft... I tried to use the new "partial url" option for the Access-Control-Allow-Origin. IIS seems to not like it.

anyway, resolved with a rewrite module that sets the correct header based on a regex.

Not optimal, but it is only a "Don't get mad" development time workaround

in the production environment, once the app is installed the url is defined and stays that way, so I can just use that in the service config.

user315433

@Stijn Because "beta" no longer means "testing phase", that role went to extended private beta, up to the appointment of pro tem mods. Once the site is established enough to have pro tem mods selected, the "beta" label is like "Division II", as opposed to Division I teams/sites.

Jason C

3:10 PM

@M.A.R. Lol I did like 5 for trump as soon as I found that link, haha.

M.A.R.

@JasonC That guy never disappoints to be funny

You can trust some monkeys to always be entertaining

Shadow Wizard

Monkeys are the new cats

ProgramFOX

@bjb568 I'd rather be efficient than 'kool' then :P

@JasonC Yeah I should probably filter words out.

Anyway I don't think I'll get immediately started on this, I'm going to be quite busy the next two-three weeks.

Jason C

7

Q: What's an efficient way to calculate covariance for a large data set?

What is the best algorithm for computing covariance that would be accurate for a large number of values like 100,000 or more?

15

Q: How to do SVD and PCA with big data?

I have a large set of data, about 8GB. I want to use machine learning to analyze it. So I think I should do SVD then PCA to reduce the data dimension for efficiency. But MATLAB and Octave cannot load such a large dataset. What tools I can use to do SVD with such a large amount of data?

bigdata data-mining dimensionality-reduction

I may not have to find that old coworker after all.

@ProgramFOX ^ cc @bjb568 if you're interested.

ProgramFOX

cool, thanks!

Jason C

3:20 PM

Disclaimer: I haven't reviewed the answers I just found those right now.

PS the PCA one doesn't apply to the text algorithms but I got excited when I found it because that was my old problem.

ProgramFOX

I think I'll have to look up what 'SVD' and 'PCA' mean exactly before I can understand the second question :P

Jason C

SVD = magic, PCA = principle component analysis = geometrically, you're basically looking at a collection of N-dimensional points then figuring out the best fit collection of axes to those points, which is a really poor explanation I guess. Not sure how to translate that. It finds a set of eigenvectors and eigenvalues of a dataset which you can use for all sorts of awesome stuff, where the eigenvalues represent the variance the corresponding eigenvector accounts for.

You can do fun things like, do PCA on a collection of a zillion faces. Then the components with the largest eigenvectors are the ones responsible for the most uniqueness between faces. Then you can recombine the components in arbitrary linear combinations to generate completely new faces based off the data set. You can also drop out "noise" by discarding the lesser components and find "typical" faces.

It was sort of our general catch-all algorithm for all sorts of stuff when I was in medical robotics, we had a super cool algorithm for finding bones in ct scans based on a database of existing models of bones. That's what that lady was working on.

I wanted to write a synthesizer based off of input samples of different kinds of drums, with sliders to adjust the weights of the principal components in the generated sound. That's the later project that I couldn't get working with a large dataset. I might still revisit it just for shits although that was like 15 years ago when it was still cool to write your own audio synths.

It's used a lot in face recognition: youtube.com/watch?v=SaEmG4wcFfg

ProgramFOX

3:52 PM

Interesting @JasonC, thanks for the explanation!

Derpy

4:14 PM

@ShadowWizard ponies are the new cats

Nov 11 '15 at 14:47, by Derpy

SmokeDetector

4:42 PM

[ SmokeDetector | MS ] Link at end of answer: Uso de jQuery en prestashop by Jose Clemente Pulido Rodriguez on es.stackoverflow.com

user315433

6:10 PM

What-do-they-call it
by Normal Human

People find the lost Programmers
In my Developer Story headline
Tags with non-English characters?
On Stack Overflow in decline?

Lost after Developer Story release
Wt am I meant to be reviewing here?
Meaning of this review audit case
Score link-only answers are there?

And weekly votes after unupvote?
Language and moderator abuse
Not updated after retracting a vote

- Is there an SE for Cell-phones?
Follow up on Area 51 commits?
And not the earlier ones?

user315433

NormalHumanSE is a decent source for SE-themed tweet-poetry

Shog9

this room is getting more emo by the day

I blame JasonC

4

user315433

> Guess what: if your site is full of crappy questions, your site sucks meta.stackexchange.com/a/220099

user315433

But... Sturgeon's law.

Shog9

yup

@ShadowWizard this is the formula

note that it doesn't make any mention of which questions are considered

I guess I could add that

user315433

6:20 PM

Manually picking up and throwing out crappy questions is a very near equivalent of Sysyphean task, and generally a recipe for burnout.

user315433

"Keep them all, Google will know His own"?

Jason C

6:41 PM

@Derpy Kittens are the new cats.

Shog9

@Shog9 added.

Feeds

7:25 PM

Stack Overflow Blog: Stack Overflow Podcast #104 – Jenn Schiffer Talks to Us about Fog Creek’s New Glitch

posted on March 13, 2017 by Kaitlin Pike

Welcome to the Stack Overflow Podcast #104, recorded Thursday March 9, 2017 at the Stack Overflow HQ in NYC. Today’s podcast is brought to you by Nightmare at 20,000 Feet, the classic episode of The Twilight Zone starring William Shatner. OooOOoo scary. Today’s podcast is also brought to you by Stack Overflow Jobs. In honor of Women’s History Month, Stack Overflow Jobs is showcasing opportuniti…

SmokeDetector

[ SmokeDetector | MS ] Bad keyword in answer: Recursively delete files of a certain extension - File name is too long (Win Server 2012) by fostersanders on superuser.com

dorukayhan

7:38 PM

> [b]Long Path Tool[/b]

feature-request: should the textboxes on stack exchange scream "this is not a forum goddammit" when someone tries to format prose with bbcode?

Shadow Wizard

@Shog9 yes, that was the missing part. Thanks.

@dorukayhan nope, nobody will read it

-1

A: Hover text (title attribute) missing in question title

We took the titles off as part of mitigating a DDoS attack on our servers - removing the title attribute has reduced the page size dramatically, meaning we could cope with the attack better. There are no plans for re-enabling it in the short term.

I bet many users want to downvote but are afraid.

user315433

@dorukayhan No, because spammers don't care.

user315433

(This is a spam post, and I am surprised it's not yet obliterated from the site.)

user315433

For those who doubt: an account with the same username posted identical text a month ago. metasmoke.erwaysoftware.com/post/57115

user315433

Which was nuked in 4 minutes, along with the user.

Shadow Wizard

7:45 PM

@Shog one more question, for how long exactly is the HNQ list cached? I mean, suppose a question is now in the list and being closed. After how much time at maximum it will be removed from the list?

Shog9

@ShadowWizard I don't remember off hand

probably two different caches involved though

maybe 3 or 4 if you squint

Shadow Wizard

@Shog9 how come?

Isn't it taking from the list of 100 Hot Questions on stackexchange.com?

Anyway, if you can dig a bit to find some numbers I would be really grateful. :)

Shog9

@ShadowWizard sure, which itself is cached. And the questions in that list are pulled from a list of every question on the network which is synced... On some schedule that may or may not be instantaneous.

Shadow Wizard

@Shog9 oh, I see. Complicated.

Well, I took the safe side and wrote "up to one hour"... ;)

Shog9

I know I wrote more about this somewhere on meta. Probably in an answer to some gnat question that I immediately regretted answering.

2

Shadow Wizard

7:48 PM

lol, "gnat question". :D

Is there also "Shadow question"? ;)

Shog9

in any case, I should note that I've about exhausted the time I can afford to spend on HNQ today, so hopefully the information I provided above suffices

Shadow Wizard

@Shog9 oh well, will use my search-foo skills to find it soon-ish. :)

ɥʇǝS

wait, are we still complaining about HNQ?

Do I need to write an HNQ blocker userscript?

Shadow Wizard

@ɥʇǝS uh? Who? Nope.

(except gnat of course, who always is)

ɥʇǝS

@ShadowWizard oh ok. good.

Shadow Wizard

7:50 PM

(this is all inspired by someone who asked if closed questions are removed from the HNQ list.)

John

8:42 PM

sd why

SmokeDetector

@John [:5849727] Body - Position 131-145: Long Path Tool

rene

8:56 PM

Hmm, maybe it not wise to annoy Nick with my handy userscripts ... all seems fine now!

Shadow Wizard

9:18 PM

@rene well, he didn't have to interfere... ;)

rene

Yeah, well ...

I didn't expect him to mention the localStorage, specially because I tried to be as conservative as possible

But his comments makes sense and I've improved the script now.

Shadow Wizard

Well, he knows a thing or two about web development... ;-)

user315433

SU election is on, three candidates already.

user315433

9:39 PM

> Personally, I like my gods having shallow, petty reasons for acting the way they do. worldbuilding.stackexchange.com/a/73873

user315433

10:04 PM

I have a hunch Grace Note doesn't like DST

user315433

Biology is up for their second election

Jason C

10:39 PM

Lol...

meta.stackexchange.com/questions/292291 noobs, noobs everywhere

Could use some NAA flags on both those answers, lol.

user315433

> @StackPodcast PRO All bread products taste better when cut diagonally. True for sandwiches, toast and especially bagels. -- Andrew Wilkinson at 11:22 AM - 7 Mar 2017

user315433

Meh. This is the correct way to cut bagels

Jason C

@NormalHuman Nobody likes DST.

Turkey killed it this year, adding yet another fun rule to zoneinfo.

> It is much more fun to put cream cheese on these bagels than on an ordinary bagel. In additional to
the intellectual stimulation...

Ah, the little things in life...

user315433

11:06 PM

Pro tem mod nomination accepted on DevOps, so presumably the site will get a green light to public beta.

user315433

And then we get a break from new site launching for a while; nothing else on Area 51 is close to completion.

Transcript for