@Jens come in the chat if you want.

@ian.milligan hi

Guys, i have an output in Scientific Notation Form. How to show it on normal form?

AccountingForm[]

Should work

@GustavoBandeira take a look

Thanks.

Did it work?

(It should, I'm just making sure :P)

Yep.

Good.

=)

It works

But there's a error

AccountingForm::sigz: In addition to the number of digits requested, one or more zeros will appear as placeholders. >>

Take a look at AccountingForm[](reference.wolfram.com/mathematica/ref/… documentation.

I've seen before.

But it's strange.

The application seems ok

Looking at possible issues now

Placeholder zeros may be needed if the requested precision is small:

Ah.

What?

About the placeholder zeros.

I've spent the past hour or so extracting information from PDFs with Mathematica but now there's some that's missing. Damn.

Extracting information from PDF's? What kind of information?

Using the TextRecognize[] function on 87 600 DPI GIFs.

CAS registry numbers, from a NIST document.

Is this tool an OCR?

They're chemical identifiers.

Yes.

Didn't know Mathematica had something like this.

It works very well, but it's not optimal.

Do you know some OCR system which is optimal?

Never found one.

Like, the input is: "You may have clicked an expired link or mistyped the address. Some web addresses are case sensitive."

@GustavoBandeira I found out earlier today, when I needed such functionality. A lot of functions I discover only when I need them :P

It gives "Bananas in the truck are like the carameloes of cameltos of the chaos"

o.O

Haha. Which makes me wonder what OCR you're thinking about?

Ok. My example was TOO extreme, i know.

Haha I figured.

No, try TextRecognize[].

Try it with different DPI settings, and file formats.

You'll see how it fares.

I've had success with 600 DPI GIF

300 DPI was just a bit too low, it confused Mathematica on some occasions

Here's the document I'm working on: ois.nist.gov/pah

It's a database of chemical compounds, but it's in PDF format (it kind of blows my mind that someone would want a database in PDF... must be an old illiterate who compiled it!)

xD

Why don't you send an e-mail: "Hey, don't you have that database on a decent format?!"

Already done.

But it's been two days, no news.

So I thought I'd give Mathematica a try :P

PDF DB FTW

Gneh.

Good night..or morning.

Night.

@acl I am not sure, is it a coincidence or did you buy it because I posted the link?

anyway, if you like it, I'd advise this one:

amazon.com/Molecules-Murder-Criminal-Classic-Cases/dp/…

by the same author, deals with organic poisons

@Fx I saw it from your link and bought it. It's good.

I'll probably take a look at this too

@Heike this was posted this past weekend. A prof of mine just posted it on his facebook page.

articles.nydailynews.com/2012-03-24/news/…

@rcollyer Ironically that a post about someone who's supposed to be hard to find keeps popping up again and again.

@CHM ... have you tried pdftotext to extract the text data ?

@image_doctor pdftotext only works if text is present as strings, not on scanned documents

@chm thanks, sorry I just skimmed the conversation and read this "It's a database of chemical compounds, but it's in PDF format..."

oops sorry that comment should have been to F'x

anyone here familiar with Einstein notation?

I have an equation which says $f(\mathvec{a}) = S_{ijkk} a_i a_j$

(a being a vector, and S a rank 4 tensor)

but I don't know what to make of k, as it's not on the LHS; is it supposed to be summed over?

@Fx sorry way above my pay grade

@Fx The link to the .pdf file that @CHM posted wasn't scanned.

@Heike oh, only pages 2 and 3 were scans

sorry then, yes, pdftotext (from xpdf or poppler packages) is definitely the way to go

@Fx I think so? I personally hate Einstein notation for that very reason

i think any clarity gained by not writing explicit summation signs is lost by the vaguenesses.

@EliLansey I thought that in the official version you're supposed to use subscript-superscript notation.

@Heike yeah, for co/contra-varaint things

but people have taken to use the term "Einstein notation" to mean "i'm too lazy to write a summation sign, so just sum over all repeated indices"

@EliLansey I remember coming across an exception on the rule where you should sum over repeated indices as an undergraduate but I can't remember the details.

@Heike yeah, there are tons

so you wind up with sentences in books like "from here on we sum over repeated indices, unless we mention it otherwise"

with the implicit "or forget to tell you, and sit back and laugh as you try to figure out why the hell we're summing over something here"

@EliLansey I guess you could tell from the context whether you the result should be a scalar or a tensor

@Heike sometimes. but the problem with this "implicit summation" convention is that it's an extra pain in the neck

@EliLansey I must have some benefits or otherwise it wouldn't be used anymore

@Heike it's less work for the person writing the paper/book who already knows the material :)

@EliLansey I do agree, which is why I'm so bad at reading it :)

Maybe they could use a different font for indices that should be summed over.

@Heike that's a pretty good idea, actually

some people use different sets of indices (i.e. greek letters, {i,j,k}, or something else) but I think that would be clearer (although harder to replicate while writing by hand)

You would only need a handful of alternative indices.

@Fx In Einstein notation, all repeated indices are to be summed over. So, in this case, this is equivalent to $f(\vec{a}) =\sum_{ijk} S_{ijkk} a_i a_j$.

@EliLansey I personally never understood contra- v. co- variant. I think of it in terms of the vector space and its dual, i.e. $\langle \Psi \vert \Psi \rangle$ is equivalent to $\Psi^i \Psi_i$ where $\Psi^i = (\Psi^\ast)_i$ is the dual of $\Psi_i$.

@Heike See my comment to Eli, above.

@EliLansey Most groups use Greek for 4-vectors and Latin for 3-vectors.

@rcollyer yeah, I had finished to convince myself ot his (with some help on Physics.SE); thanks!

At first site, it made my eyes cross, too. But, I found it is very useful, at times. Other times, I specifically add the summation back in just to be sure that I'm doing it correctly.

In particular, I found Einstein notation extremely useful for dealing with this equation: $$\mathcal{Q}^{(\text{M})}_{i j} = -\frac{1}{6}\int(3 x_i x_j - \delta_{i j} x_l x_l) \partial_k M_k\, \text{d}^{3}x$$.

The site is getting sizable. I'm starting to get "phantom" rep, i.e. days/weeks/months later, I am getting rep from stuff I've posted.

@rcollyer me too. Actually that's been happening for a while already, but it's certainly increased of late.

@OleksandrR In the past 1/2 hour it has occurred twice on an answer that is off the front page. So, people are searching for things. Encouraging sign.

@rcollyer hard to say whether that's the case or if it's just users catching up after not having logged in for a while. I got an upvote on this one just now that I think is probably the latter, but I got another on this, which is probably more searchable.

However, the site stats definitely show us that there are visitors who aren't necessarily regular participants here, which is highly encouraging.

@OleksandrR definitely.

Also I think there's a strong tendency to link between related posts. Of course that's a good thing, but it perhaps means that older posts are more visible than one would think by looking at the front page.

@OleksandrR If you don't mind me asking, how did you find the info required for the first answer, i.e. the BLAS info?

@OleksandrR Good point. I try to bring up related posts when I'm writing a question/answer.

@rcollyer I just used Trace. Nothing sophisticated. That's my usual approach; Trace until you find a way in, then hook functions if necessary, etc.

@OleksandrR I actually don't use Trace all that often. Maybe that needs to change.

@rcollyer I think you can be excused for that; the traces usually are pretty overwhelming (especially with TraceInternal -> True). (Kind of annoying that TraceInternal is now undocumented, but at least us long-time users remember it.)

@OleksandrR I've used it since around v.3. That's the first version I owned. I missed all of v.4, though. The company I worked for didn't use mma whatsoever.

And, I didn't use Trace at that time. It didn't even occur to me then.

@rcollyer I've been using it since 4. But, it took me a long time to become comfortable enough with how the system works to do any serious spelunking.

@rcollyer I also prefer vector/dual thinking to the super/subscript index. but at least with super/sub it's more apparent what's a dot product, etc.

@OleksandrR What does TraceInternal do?

@EliLansey basically what its name suggests; it makes a lot of internal, undocumented, not-supposed-to-be-user-visible functions appear in the output of Trace (such as the BLAS function I used in my answer, for example).

Without it a lot of kernel functions act as black boxes to Trace, but with it one can (sometimes) see what they're doing internally.

@EliLansey It wasn't until that occurred to me that GR started to make some sort of sense. Note: it only makes a little sense to me, not much more than that.

@OleksandrR interesting. a new toy...

@rcollyer ditto. tristan needham is working on a new book "Visual Differential Geometry"

which i can not wait for

@OleksandrR I've been able to "make it dance" for years now, but I've only started to think about digging into its internals to figure out how it works.

@EliLansey There's also a chart in the 3rd edition of Greenstein's chapter on relativity which discusses different metrics. For materials, it makes the statement that the metric varies, and with that I started seriously thinking about the distinctions between vector/dual. Not that I've gotten very far, but it helps.

@rcollyer I think there's still a big difference between knowing enough to do what you want to do, and being familiar with all the various techniques and being comfortable exploring the limits of what's possible. My knowledge of graphics, for example, is workable but I would hardly say it's that well developed, and as for e.g. typesetting a book in Mathematica, I'm not really sure I'd know where to start as I have no familiarity at all with notebook programming.

@OleksandrR I know enough about typesetting in mma to have written up my homework in it, once. If you were to ask me to write anything lengthier, I'd have to have to pass.

@rcollyer I haven't yet started writing my thesis (although, I probably should have...) and for a while I considered writing it in Mathematica. But, I'm not sure that's really a good idea given my unfamiliarity with Mathematica as a publication medium.

@OleksandrR I'm going with LaTeX and dealing with the learning curve. But, it is easier than mma.

@rcollyer Also doing it with LaTeX for mine

Lunch is calling. Talk to you all later.

bye

@rcollyer they are clearly living in sin

in particular:

while

@Fx @imagedoctor @Heike I've used Apple's Automator to extract text from my PDF, but Mathematica's TextRecognize[] worked better. It caught more CAS numbers.

@acl Clearly.

I flagged this for migration. Anybody want to add their own flags to this?

@rcollyer do you really want to move your answer? Why did you answer it then?

@SjoerdCdeVries I answered it without thoroughly reading through the question as pointed out in the comments. I think my answer/the ensuing comments are worthwhile attached to the question, though, as they help clarify.

I thought about just deleting it.

@rcollyer OK, migrated to comment

I think it was mentioned before that we could have our own code repository, or a blog for that matter

Are people still interested?

I think, for a site like ours, it would be an excellent complement.

@SjoerdCdeVries sorry about the extra work.

@rcollyer No problem

@CHM I think a blog would be interesting, but we'd have to ensure it was populated weekly or biweekly (at a minimum), otherwise it isn't as much use.

A code repository would be interesting, also, but where?

@rcollyer And what do you want people to write in the blog? "I've been able to do this/that , here's how"?

Well there's choice: github, sourceforge are two easy picks.

But at the same time, I figure a blog is more convenient for posting the kind of code we write, since it could give readers a direct look at what the code does (pictures, maybe CDF) and be accompanied by an explanation.

Or I guess a google + page would be fine too

@rcollyer check this out

I've set that up but I'll probably remove it. I thought multiple users could contribute to a google+...

@Verbeia was it gradually increase (I think not)? and if not, do we know where the new traffic came from?

The number has been increasing steadily since the first week of March

Graphs are now updated on meta

@Fx It's a bit hard to tell because the number is the 14-day moving median. There was a sudden jump up to the new level over the course of a day, but that just means the traffic has been rising enough that the median rises, which it naturally does as a step change. That happened two weeks ago as well.

I think some of the new traffic comes from the fact that we are regularly on the StackExchange front page with "hot" questions. I am not sure what the algorithm is but it seems to be when a new question gets a lot of votes and multiple, upvoted answers in a short period.

We are the only beta that gets on the front page regularly.

We have now vaulted over Quantitative Finance, Writers and Software QA in the traffic rankings

@CHM As I said on the meta post, I think a blog is a great idea, but I would rather use the existing blogoverflow mechanism. That will be another way to drive traffic to the site.

Fitness & Nutrition and Christianity both have blogs and are beta sites. I don't think we are quite ready yet but we will be soon.

@Verbeia I've answered your comment.

@All I'll edit my opening post later, as the ongoing debate is not quite the one I wanted to start.

Edited.

@Verbeia BTW, did you use some of my code to gather the Area51 stats?

Or do you still gather it by hand?

@CHM I confess I'm still gathering it by hand. I'm building up the time series, so each morning over my coffee I note down the stats. It makes sense for me because I travel a lot, and I can note it down after viewing the page on my iPhone or iPad. I don't need Mathematica with me then. I have stopped taking my laptop on business trips and just take the iPad now. (The iPad screen is amazing)

@Verbeia right, I see.

You must have the new iPad then?

yes, bought it last week. I had a version 1 before. I think it's changed the lives of lots of people who travel a lot on business.

I do at least 9 international trips a year, and I live in Australia, so it's a lot of time in planes.

Nice.

Do you enjoy traveling for work?

sorry I've been busy :(, I can't believe I've found 2 gcc bugs in 2 days

gotta update to 4.7 and see if I still have the problems, this is insane and I want to get back to MMA :P

@Verbeia ah, does the new screen really make that much of a difference? I am tempted but can't really justify yet another gadget lying around...

@CHM 6-8 trips would be fine. much more than that is a bit taxing, but it's part of the job and I love the job.

@acl yes, definitely. check one out at the store. it's incredible.

@Verbeia hm, maybe I shouldn't actually check one out :)