aichrist

• 2012-05-17

  Why I Use Owncloud

  I have previously posted about how I use the program unison to distibutrute and synchronize my files across machines that I use. This has been working out well on all of the Unix-like machines that I interact with on a regular basis but has been lacking on Windows systems.

  In the meantime, I have been using both dropbox and Google drive to store a small working set of files. This has obvious drawbacks such as space limitation, cost, and security. Despite the limitations of this approach, it has worked quite well to sync with my windows machines and as such has been a tolerable addition to my data management arsenal.

  Over the past months I have been following some other options which aim to be open source replacements for dropbox-style services. One that I quickly skipped over in my initial survey was Owncloud. The sole reason I skipped it over so quickly was that it only had web and web dav interfaces and as such was a centralized solution. I absolutely require distributed solutions as I do not want the extra time burden of maintaining a central server and, more importantly, the burden of backing it up and periodically verifying the restore process. In other words, I require passive solutions which will not leave me in a bad place should I neglect a server for five years or so.

  Although Owncloud was subject to these critical flaws in my initial survey some months ago, its developers have since added sync clients for many platforms including the ones I use regularly. This progress immediately prompted me to install the Owncloud software on my linux server and the sync client on a few of my desktops. At first, the requirement of a central server seemed to conflict with my requirements but ultimately it does not because if the central server were to die for any reason I will retain a working copy of my data on each machine which has the client installed on it, which is exactly the type of redundancy and distribution I require.

  In my initial testing the Owncloud solution has been working well alongside my dropbox and gdrive clients. I will slowly try to use it more and will ultimately migrate to it should it continue to function.

  Furthermore, this software is so central to my daily computer usage that the Owncloud developers might have earned a new patch contributor.

• 2012-03-26

  drt054 asked: Hey Andy.. I heard you have been busy, just wondering if you have made any progress with cm9 on LG Revolution? PlayfulGod got his hands on a revolution and one other guy in the forum offered help. Disregard all of the complaining if you go on to XDA to read anything.. idiots. Thanks!!

  Hi drt054… I haven’t made any progress on the build recently, though I’ve done some research into possible ways forward. For me, the main issue is a lack of time.

  I’m very happy to see that some others have picked up development some. We really need a community of people working toward it rather than just an individual or two.

  -Andy

• 2012-03-07

  Using the UNIX fold Command

  Today, while writing some technical documentation, I encountered the need to limit a large amount of log output to be under 80 characters in width. Traditionally, at least when dealing with natural language, I would use the fmt program for this task. I discovered that the fmt command is not apt for this task when it began joining unrelated lines into single lines. The output was made to fit in 80 columns of text, but the semantics of having separate log lines was lost.

  After some time, I discovered that the UNIX fold command exists and is much more suited to the task of limiting large amounts of programmatic text where distinct lines are of semantic importance. What fold does is merely wrap lines that are greater than the limit. The limit can be set via the command-line switch -w, but the default sufficed for my purposes. The output of the fold command ends up being equivalent to what you would get if you took a terminal emulator and reduced its width to be less than the width of the content. This is exactly the behavior I required in order to limit the width of log output.

• 2012-03-01

  Using Haskell, Pandoc, Markdown, and Mathjax for a Brilliant Literate Programming Solution

  I originally posted this as a reply to someone on stackoverflow:

  If you intend to share programs on the internet, I’ve found a combination of literate haskell in markdown style with mathjax to be a great combination. The program “Parsec” is seriously brilliant for taking this “markdown+lhs” to any format you desire, including PDF or HTML. If you tell Pandoc to output to HTML, you can use the -mathjax (or other similar flags if you prefer) to have your latex math formulas render.

  When using this style, I find bird style to be preferable because it just is more readable to me and seems to fit along with the markdown style better.

  The great thing about using Pandoc with markdown is that you can add citations to your code, math formulas, and have a really portable format. You can build something that resembles a scientific research paper but is executable and can also be posted to blogs/wikis/websites.

  To give an alternative point to Norman where he says that literate programming is useful for clearer code arrangement, it could be argued that Haskell is expressive enough that the problems you solve with the code are actually interesting and can really benefit by being surrounded with explanatory text. Think of a mathematical research paper. Good papers in pure mathematics have a lot of text to explain the motivation or higher-level interpretations of what the mathematical notation means. In a paper about the Navier-Stokes equations, for example, it would be super useful to surround the notation of the equations with text explaining how it relates to Newton’s conservation of momentum.

  In summary, I have had good success with, and recommend, using markdown+lhs style, dollar signs to embed latex math formulas, bird style, and pandoc. I would recommend writing programs as if they were research papers and treat the haskell itself as you would mathematical expressions in a research paper.

• 2012-02-28

  Last.FM Scrobble from Soundcloud in Google Chrome

  It’s possible, even easy, to scrobble from Soundcloud when using Google Chrome.

  The key is to use the Last.FM scrobbler script from userscripts.org. The non-obvious part of this is how to get Greasemonkey scripts in Google Chrome. That’s easy enough with the Google Chrome extension called Tampermonkey. The best part is that if you have your browser set up to sync with your Google account, your extensions and user scripts will automatically be there, regardless of which computer you access the internet from. I find this to be a rather satisfactory solution to the problem, and the solution is one that can be used for many other purposes as well.

  

  EDIT: It may not be true that the Tampermonkey scripts sync via Google Chrome. The issue “Scripts synchronization” and its accepted state led me to believe this was the case. It may just be that it isn’t working in my case, but it’s something to keep an eye on.

• 2012-02-27

  Example “Shell” Script in Haskell

  Synopsis

  This is an example of how to do something like shell scripting with literate haskell. It is like shell scripting in that it exhibits composition of standard UNIX utilities, but it is unlike shell scripting in that it is done in Haskell, which has a very powerful type system. This means that issues with escaping quotes and even many runtime errors are minimized.

  Dependencies

  The primary dependency is HSH:

  > import HSH.Command

  Main Program

  > main = do runIO $ "ls -l" -|- "grep hsh"

  Execution

  This document is a literate program, so to run it, copy the text to a location L and then execute runhaskell L. It’s that simple.

• →

  Cygwin’s Terminal Emulator

  I updated cygwin today on my Window 7 machine, and I noticed something new: the “Cygwin Terminal.”

  I’m not sure if this is new or if it’s something I somehow have missed up until this point, but I quite like it. See, previously cygwin would, by default, provide a batch file which would start a bash process inside of cmd.exe. Windows’ cmd.exe is quite a horrible terminal emulator when compared with the likes of urxvt, xterm, putty, or Terminal.app on Mac OS. This new “Cygwin Terminal” supports transparency, supports arbitrary window sizes, supports copying by selecting with the mouse as in typical UNIX terminal emulators, and so on.

  I’m not sure how, when, or why this got here, but I am really happy to have it.

  EDIT: I’ve been informed that this terminal is mintty http://code.google.com/p/mintty/

• 2012-02-06

  MIA’s Middle Finger Offends Me

  MIA’s middle finger offends me.

  While M.I.A. has offered no official comment on why she threw up a middle finger during Sunday’s Super Bowl halftime show, a source close to the rapper told ABC News Radio that her actions were caused by “adrenaline and nerves,”  a result of getting “caught up in the moment,” and “weren’t an attempt to make any kind of statement.”

  http://abcnews.go.com/blogs/entertainment/2012/02/m-i-a-upstages-madonna-with-middle-finger-at-super-bowl/

  The gesture offends me because it apparently had no “kind of statement” behind it. I suspect this response that the gesture had no statement is a lie motivated by PR, but that would imply that MIA hires PR that censors her. That would directly contradict the persona she presents in her other work.

  A highly-public, highly-risky offensive public display really ought to have a good reason behind it. At least, that’s my opinion.

  But as critics come after M.I.A. for making a minuscule hand movement during Madonna’s performance, more people — as always, as she has said — are missing the message. Never mind wardrobe malfunctions or Madonna’s medley of hits performed on Sunday. There was more to M.I.A.’s mannerism than a silly pseudo-statement. This time it was a “fuck you” to America, and once again the media is missing the message more than ever.

  http://rt.com/usa/news/mia-super-bowl-finger-623/

  RT seems to get it. They don’t, however, cite any direct evidence that this gesture did have meaning. They’re inferring the meaning based on all the public information about who MIA is and what the message behind her art is. That’s a reasonable inference, but again “a source close to the rapper” said there was no meaning.

  MIA has a few choices:

  1. Do nothing and remain inconsistent.
  2. Renounce her gesture and remain inconsistent.
  3. Own up to her gesture and double-down on what she believes is right.

  Hopefully MIA will choose option 3. Then I won’t have to complain to the FCC about how offensive her gesture was.

• 2011-12-27

  Why I use unison and dropbox together

  Ever since I began doing serious work with computers I’ve been in search of better backup and data management solutions. Many automatic solutions exist, but even automatic solutions require active monitoring to ensure that data retains its integrity. Thus, I have come to accept that important data must be actively managed for as long as it remains important. Data management is much like security in this sense.

  For many months now the tool I rely upon for active backup and synchronization of data has been unison. It is much like rsync with the primary difference being that unison synchroizes in two directions rather than the one direction you get with rsync. So as long as I remember to synchronize my files between all of the computers that I use on a regular basis, which happen to be physically distributed, my files remain backed up and continually verified for integrity as used. This has been successful so far as I remember to sync my files just as I remember to lock my house up when I leave home.

  The data under active management via unsion is as secure as my ssh keys and local security. No external computers or organizations need to be trusted.

  Dropbox adds utility to this setup by supporting automatic realtime synchronization of files that I am actively working with. I do not rely on dropbox for long term storage or for storage of large datasets for a few reasons. Firstly, dropbox costs money. This is because it is a service and is provided by real people who get paid. I stick with the base free service because the storage quota is sufficient for my working sets. Secondly, entrusting ones entire long-term data storage is a risk that need not be assumed. Thirdly, moving large datasets offsite over an Internet connection is just too slow and wasteful.

  While using dropbox storage for short-term convenience, I still sync the working set via unison. I do this by simply configuring dropbox to use a directory within a directory synced by unison. This setup has proven to be highly convenient and reliable after months of real-world usage.

• 2011-12-21

  Why I Sold GE @ 17.51

  Today I sold the shares of GE that I mentioned in the previous post. The reason I sold them at 17.51 is that the price rose beyond the limit order I had in place to take profit in the case that the price rose enough to yield at least a 15% annualized return.

  I originally ended up getting into GE at 16.75 because the order I put in after the previous blog was fulfilled after the price had dropped. At the time of purchase, I put in a limit order to sell at 17.51, good for 60 days. Had the order executed at the last minute of that 60-day time frame, my net yield for the trades would have been exactly 15% annualized. Since the limit was hit after only nine days, the annualized yield ended up being a bit over 100%.

  One thing I would have liked to have done is set a stop-loss rather than a limit order for the early exit. The reason I did not do this is that the GE purchase was meant to be a long-term dividend-yielding investment at worst, and an early opportunistic exit at best. The range of return on the investment was most probably going to be in the range of 4% (the dividend yield on a long-term hold) and 15% (the appreciation yield on the early exit). Because the early-exit was an opportunistic action, I would have only entered the stop-loss once the price got very close to the opportunistic exit price of 17.51. As far as I can tell, Etrade does not support this type of dynamic order, at least with my account level. I did not want to rely on my own reflexes to enter the stop loss at just the right time, risking the price moving below the opportunistic exit price. After doing some limited investigation, it looks like the “think or swim” platform by TD Ameritrade supports more complex automated trading of this sort. Of course, there are many other ways to do this as well, but that seems to be one of the simpler ones.

  Next up, I am looking into T as a possible buy. I will, however, wait for more cash to flow into my high-risk account so that I can suffer a lower commission cost relative to the investment yield.