@acl I could post it on my blog... but the question on the site has yet to be edited to be the question intended to be asked.

The peaks I saw in the visit data were caused by R.M. posting links to mma.se on Reddit. In the meantime this earned him/her 4 golden badges. This is almost too easy...

@SjoerdCdeVries Oh, Reddit...

I'm a bit disappointed as it means some of the site growth I saw was artificial and transient.

It sucks to say this, but I don't think we'll be wildly popular (like SO for instance).

Unless for some reason wolfram.com starts linking to us, for instance...

@JM Maybe it's better that we're not going to be wildly popular. I wouldn't want mma.se to end up like c# or c++ on stackoverflow

I guess you're right.

@j.m. The Dutch distributor already mentions us to his customers.

@SjoerdCdeVries Nice to hear!

was there a question about navigating through a 3D scene, like walking through it?

can't find it

@Szabolcs It got deleted

Why??

Oh yes, that one ...

@Szabolcs Long story

Did RM say the OP requested deletion after he got answers because the code is "confidential"?

Found it

@Szabolcs That's how I understood it.

mathematica.stackexchange.com/questions/4911/…

I happen to need something similar now. I think such deletions are unacceptable.

Is the manipulate very slow for you as well?

or is it just my graphics card? (very old and not very good)

@Szabolcs Which one? The one in my answer?

@Heike Yes. I am wondering if it at all feasible to implement "walking through a scene"

The manipulate isn't slow on my computer. I guess the graphics card could have an influence.

It seems the graphics is continuously rebuilt , would it be better to only change the view parameters without touching the graphics themselves?

Rotation using the mouse is quite fast

while dragging the sliders is slow

Maybe it helps to define gr outside the manipulate.

@Szabolcs, try something like tabb = ReleaseHold[tab] in the outer With and substitute it inside

It seems faster, but my gc is very fast here, so I cannot really see a difference

I'll play a bit with it :-) Also, Heike put a nice camera cone there, which might contribute to slowing things down (and I don't need it). Perhaps performance wouldn't be a problem after all.

@Heike, since we have this example right now.. I asked here a question

which had the background to clone the 3D->2D-screen projection which is done when you display a graphics.

@Szabolcs The cone is semi transparent which usually slows down rendering.

Does anyone know what Method -> {"RotationControl" -> "Globe"} does? I can't perceive a difference.

I found it under Texture -> Applications -> Sky box in the docs

(István linked it in a comment)

Do you see any possibility, that we can extract the matrix which is used to do the camera-projection by ourselves?

AFAIK we have only ViewPoint, ViewCenter and ViewVertical. No Angle.

Actually it is bit different.

@halirutan That's quite clever actually. I wish I knew about this before. It would have made a demonstration I once wrote a bit less complicated.

@Heike Oh, I saw that. Too bad, it had nice answers...

It's funny how the deletion of this question has given is more attention than it would have had if it had been left alone.

@halirutan It should be possible to construct the projection matrices from ViewVector, ViewVertical, and ViewAngle, but I haven't managed to get it working so far.

If memory serves, the documentation never bothered to explain how all those View* options all fit together from the 3D objects to the 2D representation we see...

@Heike @Heike, with this one could eventually manage to write better pdf-export for 3d graphics.

@JM There are some examples for ViewMatrix in the documentation but they all deal with easy cases using orthographic projections

@Heike Yeah, they were trivial ones. Nothing general. Very irksome.

Do you know how to control the aspect ratio of the viewport for Graphics3D without changing the aspect ratio of the graphics? Say, I want a viewport that is rectangular with aspect ratio 1/2, not a square

It seems that an explicit ViewVector setting gives me a square

@Szabolcs Maybe you could use PlotRegion?

great idea, it works perfectly!

@rcollyer: you got obsessed with the Faux-schach blots, apparently...

@JM yes, Simon Wood's entry particularly amused me. And, I liked this image.

I see alien bugs.

The tiny image does not do it justice, though.

@SjoerdCdeVries I can understand that. I think we'll fit in the 5k - 10k visits/day range, and that will be comfortable.

@rcollyer I'm pretty sure it's ImportString...

oh wait

nm

i'm a bit slow this AM.

@EliLansey It was. StringImport isn't in the docs, so I fixed it.

@rcollyer yeah, i misread the edit order. too little sleep last night...

@EliLansey I'm very familiar with that condition. Prior to kids, I thought I understood sleep deprivation. After kids, I understand now how wrong I was. Grad school doesn't help, either.

@rcollyer the combination of kids and grad school is not good

@EliLansey for productivity or sleep. :)

@rcollyer yes

my advisor decided that i absolutely needed to finish a paper draft last night to send to a collaborator, so i ended up till 2am working on it

and, of course, he's nowhere to be seen this morning...

That's fun.

@acl I'm working on updating my old code now for Mathematica 8. I'll ping you again when I've posted it in my blog.

@JM with you back regularly, it dashes any hope I had of surpassing your reputation. :)

Hi guys, I have a quick question that is probably me not seeing something obvious... I hope you can see it

Why s[[1]] /. s -> {a} warns me that "Part::partd: Part specification s[[1]] is longer than depth of object. >>", and s /. s -> {a} followed by %[[1]] doesn't?

@rcollyer What? You guys are quick with the low-hanging fruit, and I'm a notoriously slow typist...

@JM Ah, but you out answer me in terms of volume, consistently.

@agravier At issue is the order of operations. At this stage s[[1]] /. s -> {a}, s has no depth that Part (shorthand: [[1]]) can access, so it throws an warning. But, that is all it is, once the substitution occurs, s[[1]] has meaning.

And, is then evaluated to be a.

@rcollyer I understand, thank you :)

If you run Trace[s[[1]] /. s -> {a}] you can see this happen, amid the confusion that is the output of Trace. :)

Maybe TracePrint[] is a less confusing way to see things... ;)

If you know such an error is going to occur, you can suppress is via Quiet[expr, msg], which in this case would be Quiet[s[[1]] /. s -> {a}, Part::partd].

@JM maybe. It is still a jumble. But, the level of indentation has meaning! :P

Different strokes, I suppose. I have to think harder with the output of Trace[] than with the output of TracePrint[]...

True, but I rarely use either. So, I'm still getting used to their output.

Hey, @JM that message to RM was mistaken. Can you kill it?

@rcollyer Neat, thank you. I think Quietting is dangerous while my code is not rock-tested, but I'll apply it later

@agravier It is dangerous, so it should only be applied in select circumstances.

* I mean tested and rock-solid

I understood. I have several places in my code where that message is a clear error, and others where it is just Part being annoying.

@rcollyer Okay.

Thanks.

...or a way to shirk your responsibility to code your thing so that no errors whatsoever pop up.

Here's a case for its use as the function emits a message and returns $Failed. Since messages can be slow, suppressing it and just testing for $Failed seems like the way to go.

@rcollyer Well, if you're deliberately breaking routines, I suppose...

In that case, I don't think you're "breaking" it so much as ignoring the unwanted behavior. In that case, it is like catching a thrown exception and handling it without the user ever being aware.

(Meta: My browser can't seem to be able to stay on this chat, I get "There seems to be a problem connecting to the server. Please check your internet connection and reload this page." every ~10min. Is it normal?)

@agravier On a slow connection, that happens...

@JM so true.

Although, I've used Quiet in a function I've written where I wrote the warning message...

I should get some work done. For once, it turns out a question asked has relevance to something I was about to work on ... So, see you all later.

Ok, I was using substitution rules as an equivalent of "let" in other functional languages, but I was having big problems in recursive functions, with this error message from Part + the funciton not working. Now I've discovered the wonders of Block. Yay. Progress.

@rcollyer See you and thanks again

@JM yes, please do that

thanks

Hey, @JM I don't know how much group theory work you've done, but is there any chance you can make sense of this?

I think I got ViewMatrix sussed out (sort of).

"sussed out"

hadn't heard that in years

@acl Isn't that expression used anymore?

@Heike it is, but only by native speakers

haven't lived in the UK for a while...

@acl Ah, ok. I didn't realise it was a typical British expression.

well, I have no idea if it is, but americans don't use it, apparently.

I recently learned that hooligan is considered a rather old-fashioned word in the US.

@Heike on a different note, you mentioned you write applescript so you'll feel my pain: I have spent almost an hour trying to remember how to add a string to the end of a list

I hate whoever designed it

@acl Applescript is one of the most incomprehensible programming languages I've come across.

I'd say it is the most incomprehensible language. I've been struggling with it for years and yet can't predict what will work and what will not

@acl It is also old-fashioned, and primarily it was found in the South/rural areas. You occasionally here it used in old (1950's and 1960's era) TV shows.

@rcollyer didn't know that

@Heike so actually I gave up and googled it. the answers I found don't work...

they fail with "expected reference" and the like.

Sounds like fun.

Just dropped in to clarify "sussed out." Gotta run.

bye

see you

@acl At least once you get the script working you'll still be able to tell what it's supposed to do in a year's time.

@Heike if I finish before the year's up

what are you trying to do?

transfer some tags from Media Pro to Lightroom (these are media cataloging applications that I use for my photography); media pro has fields for people, lightroom hasn't, so I am writing a script to add keywords with the names (which do get transferred between them)

and I just managed to do it! yeah!

hooray!

although I could have done it by hand faster (view all the photos tagged "john", add a keyword "john" to them). I don't have that many different people there...

definitely would have taken less than an hour

I'm thinking of switching to lightroom (I'm using Aperture at the moment)

@acl but think of all the time you'll save in future

@Heike right, every time I switch from media pro to lightroom!

hooray!

@Heike well, I have been using media pro and capture one for years, and now am switching to lightroom. I like its interface much better than aperture and better than capture one, the quality is fine, its cataloguing features almost as good as media pro.

all in all, with lightroom I have it all in one place, with no random crashes like media pro, and with a more versatile raw converter (c1 is not so good at higher ISO and I have a fairly old camera--c1 is fantastic at low iso, not so goot at higher sensitivities). and a nice set of controls too

@acl I really like the double screen feature in lightroom which is something I miss in aperture.

@Heike well, the latest version is overall very nice or I wouldn't have switched over after all these years

Also, the instructor of a photography course I'm following keeps referring to lightroom so I feel I'm missing out a bit at the moment.

@heike I wouldn't choose a program based on that

@acl No I realise that.

But since I'm quite new to either Aperture and Lightroom I rather go for the one I know I can easily get help with if I have problems.

@Heike well, that's one approach. the general idea is simple enough that I doubt you'll have any trouble once you have a think about what is being done

but, I find lightroom's controls really much easier (and actually pleasurable) to use than anything else's, even after years of using capture 1. so it's doing something right

@acl I've downloaded a trial version of lightroom so I'll give it a go.

@Heike so is this a general photography course or postprocessing or what?

@acl The course is supposed to focus more on artistic aspects of photography but there isn't really a strict schedule and the teacher is willing to discuss any questions that come up including those on postprocessing.

@Heike well, an interesting exercise is to extract the unscaled channels (ie the raw numbers) and attempt to write a simple demosaicing algorithm in mma. it's not so trivial to get good results

but not so hard to get something...

(extracting the channels can be done using dcraw)

although I am not sure this teaches much that's useful for artistic purposes!

(come to think of it, it is not trivial for me, but judging from what I see here, some of us here would probably find it childishly simple)

@acl The only time I had to deal with raw data directly was when I was imaging parasites using a fast speed camera, but that was black and white and the artistic quality of the images was at the bottom of the list.

@Heike ah yes, forgot you have also done experiments

@acl Only for a little while, but it was good fun.

@Heike Australians say "sussed out" too, but you are right, it is a bit out of fashion, I think.

@Verbeia I say "got it sussed" but i am a bit out of fashion myself :)

Have you noticed at wolfram.com/events/technology-conference/2012, that the old MathModelica (if I may say so) has an official new name "Wolfram SystemModeler" (I'm guessing the relation from its name). I'm very disappointed that this is coming as a separate product. Lately, Wolfram "all integrated" strategy has failed me on both the Finance Platform and now on this one...

What do you think? (for those that did not sign a NDA)

@PFonseca Where does it say it's a separate product?

@SjoerdCdeVries It doesn't. But if not, why a different name for something that would be a simple (or rather complex) extension of the interface functionality (and new function...). I just hope I'm wrong on my deductions...

@PFonseca me too...

I'm still living badly the second Mathematica version that goes with the name of Finance Platform. I can understand the fact that some functionality of the FP has a cost for WR (some sort of DB interface and access service paying). But, for example the Excel link that is "offered" with the FP should already exist on standard Mathematica, for a long time...

Excel is the standard in almost any production company. A strong standard interface with it, would mean a lot.

In this small details, we can see clearly which is the typical use of Mathematica...