I think my install is acting a little flaky. Over the past couple of days, I've noticed it take a long time to do simple tasks. First, there's the complete lack of syntax highlighting for symbols from packages. Second, some tasks seem to send the kernel into an infinite loop. The most recent one: at some point it will be unable to Import a "TABLE". I've run a check on the hd to verify that it wasn't the hd itself that was at issue. But, it is getting frustrating.
Closing down seems to work, for a time. Is there any way to verify my installation? Should I ask this on the main site?
Never mind about Import. It appears I have a resource leak in my code, to the tune of ~31000 InputStreams. Yep, you read that right. Have to figure out why they weren't closed. :(
For the curious, I was using StringToStream to Read in an array of numbers in a line loaded from a file, and was forgetting to Close the Stream. So, for each file, I kept building up more and more streams, and likely Import was having a hard time getting its hands on one. StringToStream is evil, and a better method for converting a String to a number has to be devised. Or, just ensure it is wrapped appropriately. :P
@MrWizard Perhaps it'd be useful to include a link to that "cautionary tale". It's not at all obvious that it's so important to close streams. He was using StringToStream exactly for the same purpose, and that's how he ended up with a gazillion open streams.
@CHM Simple, sometimes the only way to easily process a file is to read it in line by line, and determine what to do with it then. Unfortunately, there is not a great way to convert 1.19273e-10 to a number MMA can understand, hence StringToStream and Read.
@CHM In general, I search for stuff with *`*foo* for anything that might have foo in it. Since Mathematica is pretty verbose, it is very likely that a few trials will point you to interesting stuff
@CHM I'm reasonably certain that it can only read mx files that you've created on your system... but then again, I don't tinker much with these, so don't take my word for it
@rcollyer well, I'm sure they have their reasons. Also, we were notified of it in advance, so we should be patient. Besides, not all network moderators have the sites they moderate as their primary activity site :)
I don't know if Sjoerd and Mr.Wizard are flooded with flags and mod stuff to do, but if they are, I'm sure they'll let SE know to add more. In the mean time, there's plenty of us to do the routine closing + deletion.
It's a pity that this functionality doesn't work on ListPlot or DateListPlot.
I just contributed the last remaining delete vote for this question but I felt a bit bad about it, given the highly voted answers on it. I'm wondering if we shouldn't have some guidelines on what to do with marginal questions that attract good answers.
@Verbeia ah, so that explained the drop in rep :) I was confused, because the rep history didn't show "removed"... maybe it just hasn't updated its cache
@Verbeia well, it only takes 3 2k/4k users to vote to undelete, so we'll take each one as it comes... guidelines for such cases only makes it messy and to be honest, will be fringe cases. Either way, I am not voting to undelete this, nor do I disagree with the deletion.
@RM I lost a small amount of rep too, at just the wrong moment to send me down the rankings. I had managed to maintain position #10 for quite a while despite travelling all last week and not really seeing any questions I can answer for a few days.
So I suppose my delete vote will be seen as above board :)
I don't know why anyone's surprised at the lack of traffic on econ.SE. First, there are no right answers in economics (but lots of wrong ones). Second, the people best placed to give certain answers are legally not allowed to. (e.g. central bankers with strict policies on what they may say publicly).
@Verbeia BTW, my internet might fail. If it does, I'm going to bed. If I stop answering, you'll know why xD
Hmm..
Well, I'm surprised that there's a "Research Economics" proposal, as well as an Economics beta.
Just like what is going on with Theo.Phys and Phys.
But about the site statistics... I've thought about compounding every indicator into one, like this: Questions per day * Percent answered * (avid users/total users) * Answer ratio * Visits per day
Which means, nothing probably, but it might be an indicator of site health if the number is divided by total days in beta.
I should modify my code to compile the list for all betas... just too lazy now.
@CHM the indicator of site health is better obtained as a derivative (i.e. evolution of the last N days divided by N), because a site may be healthy even if it started very slowly
well, tomorrow Mathematica.SE will have spent 90 days in beta
@rcollyer @rcollyer I import files with strings like yours (#.###e##) and have just used "TSV" as the import method `Import["filename.txt","TSV"], and then added my own wrapper to strip out our usual headers and column info.
@tkott That's why I was using Import[..., "TABLE"]. The issue with the particular file was there was a lot of formatting that I wanted to preserve.
@CHM Physics gets a lot of low-level/crackpotish questions that is off-putting to serious researchers. Not that I can answer questions on Theoretical Physics, but I prefer it over the other one.
@Heike For some reason, Area51 started recording towards the end of the first day, not at the beginning.
@rcollyer I've found "TSV" to be much more reliable than "Table" for importing data. Then again, there's no formatted type things that I need to preserve. What kind of info are you storing, out of curiosity?
@rcollyer another approach I've sometimes used is to use a regex to replace "e" in scientific format numbers with "*^", which can then be understood directly using e.g. ToExpression. (Seems to me that "*^" is really not widely used at all in the community? Personally I use it all the time.)
@tkott For the files I was using Import[...,"Table"] on, they're just the results from a density functional calculation: one line per calculation with all data being floating point numbers.
Also some fields of computer science publish conference papers rather than journal articles. The standards for presentation are a lot lower there than in the natural sciences, it seems.
@OleksandrR Right, I guess you can't write "1.0e10" in MMA, but it does understand it when importing files using "Table", "CSV", or "TSV", which is all I care about :)
@rcollyer most likely. To be honest I found it a bit annoying how their typesetting resulted in highly uneven line spacing when certain lines contain characters with accents (e.g. overtilde) or large superscripts or subscripts.
@tkott I wish I had known about it before creating nearly 31000 open string streams. It's nice to know, I can still find creative ways to bring MMA to its knees.
@rcollyer Oh, I seem to do that just by forcing it to use multi-GB of ram and it shutting down the kernel. Pretty common occurence when trying to do matrix equation solving and simplifying.
@rcollyer I haven't graduated yet to the need for quirky ways of bringing down MMA :)
@tkott Oh, I'm getting to that part of the data processing stage, again. But, it will have to wait a few days while I get other calculations up and running. Then, look out.
That does a very good job of bringing mma to its knees. :)
Although, that was on v.6. I hope v.8 fairs better.
To get that working, I had to downsample the function. I'd prefer to use an InterpolatedFunction, but that doesn't seem likely.
@rcollyer then Physics is like Biology (crap). I just hope the upcoming Chemistry beta doesn't fall prey to stupidity.
@rcollyer I figured (as many others have, obviously) ultrasonography would be a good way to measure blood flow. Turns out the gods of Biology don't like ultrasonography.
@Fx Agreed, but I'd have to learn how to use the SE API... gneh. From the couple of calculations I made, the number was around 2.9 K.
It seemed roughly stable, before adjusting for time in beta.
Here's an interesting idea for an IDE. The Table format reminds me of the Notebook interface in Mathematica. Personally, I'd like the ability to visually trace through my code like in the link, though.
Talking about cutting it close. The disk drive in my macbook broke today. Luckily it's covered by applecare but apparently that's expiring in 4 days time, so I'll have to find an apple shop tomorrow and hope they can repair it within the next three days.
I had my screen, keyboard and HDD all replaced in about 1h30, in one night. The guy told me I'd have to come back in two days. He called me back once I was getting in my house "Uh, sir, we've repaired your computer"
If customer service wasn't excellent, I'd see absolutely no reason to pay a premium to get an Apple product. I'd get a cheaper PC with linux. Customer care is priceless, especially when your computer is a workhorse.
But then, my cousin called me yesterday: she bought an MBP at a re-seller. She also bought Office Mac. The clerk asked her if she wanted them to install Office, she said yes, and paid the 70$ he asked for the installation.
@CHM I would've thought a more likely reason you were downvoted is that blood flow through a certain blood vessel and total blood volume aren't directly related, and if you know the dimensions of all blood vessels (an unlikely proposition if you include all capillaries etc...), you already know enough to calculate total volume without caring about flow rate.
@Fx I'm not really familiar with TeX; I gather not a lot of people in chemistry departments are (and most of us seem to use Windows too!). For TeX submissions to AIP, do you get a template, like with APS journals, so you can lay it all out as you like and make sure the typesetting is okay? Or are you still at their mercy to convert what you give them into a correctly typeset article?
@EliLansey thanks, I feel really frustrated about that one :)
@OleksandrR the template is called RevTeX, and it's somewhat standard TeX class; but they still perform some translation into their own markup language, and errors can be introduced (I remember from an earlier article)
@CHM you would presumably have to assume a single time constant for recirculating all of the blood through that blood vessel (highly unrealistic), ormeasure a hierarchy of vessels (quite involved). Also a single blood vessel doesn't have a constant volume; they expand and contract depending on the dynamic blood pressure.
@CHM do you get to keep the previous (broken) hard disk? I've never been inclined to take up warranties on hard disks; I'd rather securely destroy them myself. (The ones with glass platters are nice for that...)
@OleksandrR no. I was frustrated about that too, but I figured I had mostly school stuff on there. I lost a year's worth of lab reports, which was heart breaking, but nothing else really important.
I had an old back up (I back up regularly now) with all my music and pictures. When I reinstalled OSX, I did it using their security configuration guides
I'd feel less bad if my disk were to fail me again.
@OleksandrR I think the only thing the AppleCare warranty doesn't cover is the battery. It's pretty extensive.
@CHM I guess you can encrypt your data on disk to begin with and then delete the decryption key before you hand it over for service. But even so, it's not ideal.
@CHM sounds good. Personally I don't use a laptop and Mac OS confuses and frightens me, so not really something I've looked into.
@CHM probably if you do everything at the terminal. It's the GUI that I find confusing. Everything about it just seems unintuitive. On the other hand, maybe it's that I haven't had enough experience using it to give it a fair chance.
Not sure if the same would be true of Ubuntu's new Unity UI or Gnome Shell, neither of which I've used either. Standard Gnome and KDE seem usable enough, although 99.5% of what I do on Linux is at the terminal over ssh anyway.
