OH, yo umean you erase all at once

?

Yeah, but it's the same thing. If the "complete code" state isn't stored, it can't recover it... If you put "StoreFrequency"->1 it should store every new keystroke, but who knows

Perhaps there's an undocumented function that can tell me if the syntax of an input cell is correct

Another one, (obvious for you) . Selecting text with the mouse and deleting is not recoverable

There's a documented one, SyntaxQ :)

yes, of course

I didn't know it

you need it to parse programs

It could be used to always save the states of correct MMA expressions

Nice

hopefully

@Rojo I think "correct" is not a valid scope for this

I don't want the program to think for me what is "correct"

I lost my previous program

I want it again

correct or not

I'm talking about extra stores

As it is now, it just dumbly stores the state of the cell

every certain character keystrokes

I think it would be nice if in every keystroke, if the cell happens to have a "correct syntax" already, such as your Plot[f[x], {x, 0, 1}], to store it... So that it doesn't happen that you end up recovering Plot[f[x], {x, 0, 1

I honestly don't know how to do this properly. I guess we could make it smarter and smarter about what to save

@Rojo I think it is better to store "action blocks"

...

I don't think I can actually catch all the ways that a user can modify a cell

if that's what you mean

I start typing, input some text (overwriting or inserting). All my contiguous typing is ONE action

I see

of course I need redo :)

that ends when

?

Redo is simple enough

Ends when you move the insertion point OR you type something that is not printable

shift-enter

F1

or when you click

or

etc

yep

I would need to catch ALL

move the insertion pooint

otherwise, it would break

Are you catching events with event handler?

Yeah

CellEventHandler

in the stylesheet

mouse events too?

No

Simple plain character events

"KeyDown"s

(and alt+click)

and you store a pre - image?

The eventts are triggered before anything changes, so

brb wait for me

if a certain criteria meets (currently overly simple), I store the cell state. Ok

sorry, someone around needed help

ustream.tv/nasajpl

Curiosity landing @ 1:30 PM EST.

@CHM hanks!

Tom!

:D

@R.M You should have pinged me... :)

Hurray! I have finally found a use for the Mathematica book!

@acl Doesn't NKS fit there too?

@belisarius I try to maintain a minimum standard in my office

@acl You shouldn't be ashamed if you own a copy. However, having two could be a symptom of some serious disease.

@belisarius I am now trying to find a use for some of the Mathematica Guidebooks (of course all these are library copies, not mine, so I could in principle just return them...)

@acl Do they have plastic covers?

@belisarius no, laminated hardcover, why?

@acl Because plastic infuse a bad taste to bbqs

@belisarius good point, I could have a bbq or more per book

(well I need to leave one to get the laptop level with the monitor)

@acl I am an expert

now I'm hungry

Sacrifices in the name of science

this is what I'd ideally go for right now

@acl Nice. What kind of meat is that?

@belisarius lamb

@acl Good lamb is very expensive here. We export almost all of them

@belisarius yes, it's expensive everywhere, but this doesn't work with pork or beef

@acl Is that like souvlaki?

it's souvla, which (in cyprus) means large pieces of lamb (not pork like souvlaki) roasted much more slowly

in greece souvla is a different thing

@acl lamb, Argentina style

@belisarius right, but that tastes quit different

@acl why?

@belisarius I suppose the cooking, I don't know, but it definitely is different

@acl It takes 2.5 to three hours to cook

@belisarius yes, souvla is also 1.5 to 2h even though the pieces are much smaller

but there are lots of tricks that different people use

@acl lamb is horrible when cooked quickly

@acl What is it that doesn't work with port or beef?

I have to say, I haven't made any for years (no equipment!)

@jVincent this kind of roasting

at least, I did try a few times and didn't like it at all

not enough fat? I don't know

@belisarius yes, also, the meat itself can be bad, unlike beef which is almost always OK. but that may depend on the country

br

brb

@belisarius off for a snack? :)

nope, work

Thinking about it I haven't ever head about people grilling port or beef over longer periods, slow roasting in a pan though can be quite good.

@jVincent I think slow roasting isn't uncommon for beef, but I am no expert (I can only cook rice, pasta and eggs, apart from roasting stuff)

@belisarius It was nice to see them all jump around, some of them crying.

That food conversation makes me hungry.

@jVincent Ever had some roast beef?

@CHM No wonder the tears. They have a few years of a nice pay ahead. Imagine that vs. going home with a bunch of burned cables in your pockets

@belisarius probably thinking about all the nice conferences they can attend to present these unique results!

"How to send a probe that doesn't fall on its ass in 10 easy steps."

Step 1: procure $2.5G

@OleksandrR. Step 2. Learn to pray.

@belisarius that's step 0. Otherwise you will be denounced as a dirty atheist and won't be able to secure funding from Congress.

@OleksandrR. The funniest thing was that the NASA interviewer asked the same question to everyone "Why is landing on Mars SO important". Nobody could articulate a reasonable answer

@belisarius I'd have replied "just because!" myself, but it is my understanding that the taxpayers will then be throwing words that aren't for polite company at you...

@J.M. Designing a parachute for a thin atmosphere is indeed very important if you plan to land in Uranus

@belisarius Bleh, I gave you the wrong link then.

The one I watched at first was /nasajpl, with the woman interviewer.

But then I got tired of it, and watched /nasajpl2, without any interviewer.

It was the view from inside the command center.

"command" center

@CHM I am joking. Thanks again for that link

@CHM He! they "commanded" my ass. Just frantically eating their nails

@belisarius NP.

@CHM BTW. Did you notice that the Landing Manager was a copycat of Elvis?

@belisarius You shouldn't be asking that question.

@CHM Why not?

@belisarius cats don't like it. They will sue you.

@OleksandrR. oh ... sorry, cats

@belisarius You're Argentinian, what do you know about Elvis?

Sorry I was eating, writing crap with one finger.

@belisarius I did yes. I'm not sure if he combed his hair for the occasion, or is always like that.

Reading this: en.wikipedia.org/wiki/Orion_(spacecraft%29

The photo behind them is the alley at my backdoor at night

frightening

@belisarius I doubt that there is a reasonable answer!

@belisarius is that the same crew that sent your balls into orbit, then?

@acl Yes, there is a big one: "They gave us $2.5G"

@acl Being Elvis there, I think my Great Balls of Fire is more appropriate

I see the discussion has turned juvenile. Again.

Getting things back on topic, here Leonid writes:

> One problem with local functions created by Module is that these symbols won't be automatically garbage-collected when Module finishes (if they have DownValues, SubValues or UpValues. OwnValues are fine), and so may lead to memory leaks. To avoid that, one can Clear these symbols inside Module before returning the result.

Can someone give me an example of this problem? It seems to me that if definitions are used outside of Module then of course they are not garbage collected, and using Clear inside isn't applicable. If they are not, they appear to be garbage collected in the couple of tests I just did, but there must be some case where they are not for Leonid to write this.

@Mr.Wizard What did you try? :D

@belisarius This for a start, because lack of garbage collection in that kind of use would be pretty disastrous.

@Mr.Wizard I was half joking. Assuring how a GC works without access to internals is not always possible

Yes, but I presume that Leonid had an example of a leak in mind when he wrote that, while I cannot recall with certainty a situation where such a leak was produced. I was just wondering if you or anyone else had had problems with this, or could illustrate how one might.

@Mr.Wizard forums.wolfram.com/mathgroup/archive/2006/Mar/msg00170.html

@Mr.Wizard I hadn't. But that is irrelevant.

@OleksandrR. Thanks, but that is simply a misunderstanding of the nature of Part assignments (evaluation being delayed, rather than instant), is it not?

@OleksandrR. Hm... maybe I spoke (wrote) too soon. Let me look at that again.

@Mr.Wizard Sorry, I don't follow. Have I missed something about Part assignments?

Behaviour seems slightly improved in 8.0.4. In 8.0.1 I had this issue, which doesn't seem to exist any more.

@OleksandrR. I mean that, for example, doing a = Range@5; a[[3]] = Sequence[]; doesn't remove the third element from a but replaces it with Sequence[] -- it is only removed after a is evaluated, and even then Sequence[] remains in a.

@Mr.Wizard I thought that's what you meant. But isn't that just a case of Set being SequenceHold, whereas List isn't? But the head List isn't seen until the whole thing is evaluated in one piece.

I don't know if it has any bearing on the Module issue.

@OleksandrR. I suppose that's a better way to look at it. I'm currently trying to see the problem that is supposed to be shown in the MathGroup post you linked. I don't understand yet.

@OleksandrR. Actually, I guess I would be more comfortable stating that the expression in which an element is changed with Part is not re-evaluated at the time of the assignment. While logical now this came as a surprise to me the first time I encountered it.

I think that's a fair way of saying it.

The general problem I suppose can be reduced to "in what circumstances do Temporary symbols fail to be collected?" -- because AFAICT this is not specific to Module itself.

To clarify, are we talking about symbols themselves or their definitions?

The symbols. Because no matter what *Values a symbol has, if it's unreferenced, it ought to be collected.

@OleksandrR. I guess I don't see the problem with that then, unless one somehow exhausts the "symbol space" of Mathematica; something I don't think I've ever really considered, actually.

However, some sorts of operations seem to introduce auxiliary references that are hard to get rid of (like Part assignment in this case).

Well, okay. But if a symbol is not collected then its *Values will not be cleared either. The symbol sticks around with Temporary on it and all the previous definitions still existing rather than just turning into a name.

@OleksandrR. I finally saw the leak described in the MathGroup thread, and yes, the problem is as you describe. I'm sorry I took so long to figure it out. <:-o

be back in 15

@Mr.Wizard and I'm sorry too, as I think I misinterpreted what Leonid was saying. As I now understand it, he is just saying that if you introduce additional references yourself by returning Temporary symbols from Module, then you can't any longer expect the GC to pick these up unless you Clear whatever refers to them. This is not AIUI referring to a bug, but just urging caution with the application of the technique.

FYI: Dominican Republic coffee is tasty! Name's "Cafe Santo Domingo", and it has a subtle cocoa aftertaste. Yummy.

Prepared in a 1 cup Bodoum, three teaspoons of coffee. 3 minute infusion time.

@Mr.Wizard Thanks for that post, I hadn't read it. I'm having trouble with a notebook that has a lot of Dynamic and Module. I might have to refactor the code to use some With and Block.

@OleksandrR. I'll be sure to ask Leonid what his intent was in writing it.

@CHM What is a Bodoum, and (how) is it better than a drip coffee maker?

@CHM interesting! My personal preference is Ethiopian. I find the South American ones too acid, but maybe I should try Dominican.

@Mr.Wizard Bodoum is a company that manufactures these

@CHM cafetière/French press

@Mr.Wizard And it's the only thing I have here. I'm not sure it's better than a drip. Actually, what comes to mind is the heavy use of plastic in drip coffee makers, whereas the bodoum is all metal/glass.

@OleksandrR. Never had Ethiopian, but I agree on acidity of SA coffee. It's good in an espresso, once in a while, but not for a big mug.

I haven't used a press myself, but I've had coffee from one and I thought it was similar to drip coffee with a metal mesh filter.

@Mr.Wizard It probably is.

Of course I have no idea if the user knew what he was doing. :-)

I was just making sure the experiment was reproducible by giving the experimental section. :)

@Mr.Wizard if you pour it immediately, there should be no difference. If you let it stand, the coffee can become off, whereas in a drip machine the grounds are obviously held away from the brewed coffee.

Thanks. I'm always curious. Unfortunately I cannot try the spit-roasted lamb discussed earlier. :-(

@OleksandrR. Ah, yes. That's why I said 3 minutes, because you can get bad taste if you let it stand. True.

Hehe. Roast lamb would be great indeed.

@OleksandrR. Ah, you mean if you leave the drip cone over the coffee so the bitter part slowly drips in, right? I always remove the cone from the pot after the majority of the water has passed.

@Mr.Wizard I think he meant that problem is avoided in drip makers.

@Mr.Wizard I mean for the Bodum. I don't think you can easily spoil the drip coffee except if you leave it on the hotplate too long.

I'm doing poorly with reading comprehension today. :-/

I guess with that I'll take my leave. See you both later!

@OleksandrR. I also noticed some difference between press and drip, although I would have thought there should be none. perhaps it's related to expectations

Goodbye

admittedly nowadays I'm back to either turkish coffee (at home) or "coffee" dispensed from a machine at work

"coffee" :)

Black water.

@CHM it's actually not as bad as I would have expected, but it's still bad

@acl the filter paper in the drip coffee catches some of the oils and other substances produced during roasting. I think the main one is acrylamide, which is toxic and carcinogenic, so I prefer to use the filter...

@acl How do they make it? The making of cafeteria coffee is more important than the substrate, I found.

@OleksandrR. could that affect the taste?

@CHM there's no they; it is literally dispensed by a machine!

(well, two machines)

@acl Oh.

@CHM there are 2 phd students whose job is to collect the coins and refill with coffee beans every day, so I got to peek inside

@acl I have never tasted acrylamide, so I don't know. Probably other aromatic substances are also absorbed.

the raw material is indeed coffee, which I think is a step in the right direction (I mean, it's not instant coffee or something like that)

Then it should be allright, especially if the machine grinds the coffee after having ordered a cup.

Ok.

@OleksandrR. that could explain it, although I am not sure how accurate impression is

@CHM yes, you can hear that happen

Ok nice

still, I usually have a cup in the morning on the way to my office and, as that is 5 min after my last cup at home, it tastes vile in comparison

@OleksandrR. LD50 = 150mg/kg

@acl there's one like that here, too. I have had a cup on a few occasions when I've run out of coffee at home, even though they charge an absurd price for it. Two impressions: it is not strong enough by far, and it tastes oddly plasticky.

@CHM sure, but that is acute toxicity. It's also a carcinogen and mutagen.

@OleksandrR. Indeed. What I meant was : "and you're better off not smelling it anyway"

@CHM the LD50 of caffeine is not much more than that...

Nicotine is substantially more toxic, though. (I don't smoke, but people don't seem to keel over after one cigarette either.)

@OleksandrR. Oh, I didn't know that. 192 it is.

That just goes to show how little I know about chemical toxicity.

I think chemists should get a mandatory toxicology education. Just make it part of the "safety in the lab" classes taught to every first years.

The most I was taught was "if you smell bitter almonds, run for your grave"

@CHM I think that is usually the case. I'm not a synthetic chemist, but I think a safe rule of thumb is that everything can probably make you very ill or kill you unless the MSDS says otherwise.

@OleksandrR. Sure, the dose makes the poison.

HCN is not as bad as H2S or diborane, which we use here from time to time. Fortunately they stink, so you can tell if you have a leak.

I just had a look in our gas store. Practically all the gases in there are labeled "severely toxic" "flammable" and "corrosive".

=)

I miss wet labs.

Hopefully this semester's labs will be interesting.

I like to write software, but I don't think I could only do that, I like to get my hands dirty too much.

That's probably the most valuable thing I've learned this summer, actually.

I think it's good to have a mix of both. I think I would have gone mad had I not had more than one thing to work on; sitting in front of a computer all day as the computational chemists do would not be for me.

Here our main interest is spectroscopy. As well as being an interesting field, I think it's a good mix of practical and theoretical work.

Gotta go. Talk later!

Okay, bye!

@acl on another note, it's widely claimed that tea tastes better with freshly boiled rather than reboiled water, for the implausible reason that freshly boiled water retains some noticeably greater quantity of dissolved oxygen. However, I can't tell the difference between tea made with once-boiled and twice-boiled water.

@OleksandrR. I could also never tell a difference

but I don't drink anywhere near as much tea as I drink coffee so am not an expert

Maybe belisarius has a view on this?

Morning

@Rojo what's up

Not much, just got home, I'm still getting home in my head

The first minutes are really enjoyable

You?

also got home half an hour ago

meant to test the undo thing but didn't have time