Bloomberg is reporting that the NSA knew about and exploited the heartbleed bug for years. The NSA is, of course, denying that but if you’ve been keeping score it’s pretty easy to figure out who to believe.
If it’s true that the NSA knew about the bug and kept the information private so that they could use it, it’s a nice illustration of something that Bruce Schneier has said repeatedly. The NSA is making us less safe not more safe. It’s virtually impossible to know if a site has been exploited by heartbleed so it’s entirely possible, or even likely, that criminals or foreign state actors have been exploiting the bug. The NSA’s primary mission is protection of the country’s communications. If they failed to act to close a devastating hole in a widely used SSL implementation then they’re obviously failing in that mission. Failing on purpose.
Posted in General
The essential Mickey over at Mastering Emacs has another post in his series on
dired. This time it’s about running shell commands on multiple files from within Emacs. If you’re a Unix user and used to using
xargs to do this sort of thing, you should definitely check out Mickey’s post. Using
dired is easier and you don’t have to leave Emacs or even bring up a shell of some sort in a buffer. Unless you’re a
dired expert, Mickey’s post will teach you things you didn’t know. And while you’re there, be sure to follow the link to his post on how to work with files spread out across many directories. The two posts tell you pretty much everything you need to know to use
dired in situations that would otherwise require
If you’re like me, you probably started using
dired as a convenient way of listing files so you could pick one to load into a buffer. Later, maybe, you started doing simple operations like deleting a file or perhaps renaming it. It’s when you first realize that you can work on multiple files at the same time that you begin to realize how powerful
dired is. For example, watch Magnar Sveen use multiple-cursors to change file names in a writable dired buffer. Of course you can do this using
dired commands too but it shows how flexible
dired is and how it can adapt to your way of working. Mickey’s post shows how you can apply external commands to multiple files in a variety of ways. Very useful information.
Posted in General
Every programmer understands, at least in a general way, the relative latency of various computer operations. For example, we all know that accessing a value from a CPU register is faster than accessing that value from an L1 cache is faster than accessing that value from an L2 cache is faster than accessing that value from an L3 cache is faster than accessing that value from memory is faster than accessing that value from a disk.
We might even have an idea of what those speeds are. Typical values might be .9 ns for accessing an L1 cache versus 120 ns for a main memory access. If you’re like me, it’s hard to compare those values. They seem essentially the same: really small.
Here’s a wonderful spreadsheet that makes those comparisons come alive by normalizing the values to a one second cycle time. With that, an L1 cache access takes 3 seconds and a memory access 6 minutes. No need to do mental calculations here; we can see immediately how much faster the L1 cache access is. To continue the comparison, a disk access takes 1–10 ms; normalized to the 1 sec cycle time that’s 1–12 months. See what I mean about making the comparisons come alive?
The spreadsheet considers actions such as sending an IP packet from NY to San Francisco or to the United Kingdom and booting the system. Take a look at the spreadsheet. The values are truly amazing. As a teaser, booting the system takes 32 millennia.
Of course, normalizing the times to a one second cycle time doesn’t really give us more information, it just makes the information more understandable. Looking at the values you begin to see why things like locality of reference really do matter.
Posted in Programming
As I’ve said many times before, I’m a huge fan of The Structure and Interpretation of Computer Programs (SICP), Hal Abelson’s and Jerry Sussman’s seminal work on computer science. It’s a tremendous resource and one that’s available on the Web for free along with a series of lectures by Abelson and Sussman based on the book. It will, I promise you, teach you things you didn’t know and change the way you think about programming.
In the last couple of days, I’ve come across two articles that offer advice on why the SICP journey is worth taking and ways to make the trip easier. The first, WHY AND HOW TO START YOUR SICP TREK by Kai Wu explains what the excitement is about and what well known computer scientists have said about it. Wu explains why the effort required to work through the book is a worthwhile investment and offers some advice on how to proceed.
The second, Learn SICP, Make Friends, provides similar motivation but offers an iPad program that allows you to collaborate with other people working through the text. If working through SICP by yourself seems too daunting, this may be a good alternative.
If you haven’t yet done so, I urge you to take the trek. Be warned, though, it takes some work. The text is well written and easy to read but you really need to work almost all the exercises to get the optimum results. Watching the videos as you read the book is also a good strategy but, again, you really do need to work those exercises.
I’ve written many times about GnuPG and other programs that support the OpenPGP standard (the last time here). I really, really wish that everyone would start using it. That would certainly put a crimp in the NSA’s operations.
Over at Digital Era, a site dedicated to discussing tools and methods for protecting our digital privacy, there is a nice tutorial on using GnuPG. The tutorial assumes you’re using Linux but the information is applicable to most Operating systems that support GnuPG. It’s well worth a look if you haven’t yet installed GnuPG or similar system.
Of course, getting Linux users to start using GnuPG is a lot easier than getting your Aunt Millie to start encrypting her email but that’s what’s required to get the NSA and other snoopers out of our email. Nonetheless, it’s a necessary first step: Today email; tomorrow our phones.
Posted in General
Do you avoid left turns when you’re driving? I do. I’ll take different routes to destinations that are across the street from each other to avoid making a left-hand turn against traffic if the destination street has more than light traffic. I do that mostly because I hate getting stuck in traffic waiting to make a turn.
It turns out, though, that it’s the smart strategy and even the green strategy. UPS, who famously uses operations research to plan their routes and other operations, has studied the question and found that avoiding left turns saves time and gas even though it may result in a longer route. One UPS driver estimated that 90% of his turns were right-hand turns and said that UPS really, really hates left-hand turns.
I don’t know about you but I love this kind of thing. It’s geeky and practical at the same time. I especially like results of this sort when they’re counter intuitive as this one is. In any event, if you, too, avoid left-hand turns you no longer need feel silly; it’s the practical thing to do.
One of the joys of writing this blog is the things I learn from the comments. The other day, I wrote about fancy-narrow and that provoked an interesting discussion between Wilfred and Phil. Wilfred compared
narrow-to-region-indirect. I assumed that
narrow-to-region-indirect was a standard command or function that I wasn’t familiar with and put it in my todo queue to investigate. Phil beat me to it and discovered that it was actually a very useful piece of custom code.
The idea is that you clone the current buffer one or more times in such a way that you can narrow to different regions in each of the clones. If the utility of doing this is obscure to you, Zane Ashby has a nice post that gives one use case as well as showing code for a function to do it. As Phil remarks, before he saw that code he used to do the same thing manually.
You might think that you could open a file in a buffer and then split the buffer the appropriate number of times to do this but that doesn’t work because narrowing in one of the buffers narrows in all of them. The trick is to create indirect buffers, which can be narrowed separately. Changes made to the text of any of the indirect buffers is reflected in all the clones and the base buffer so that when you’re finished you can just delete the clones and save the base buffer.
You can find the code to automate this in Ashby’s post or a slightly different version by grayswx in this paste. If you have the need to work in different parts of a buffer at the same time—especially if you want to use different modes in each—this technique (and the function that implements it) could be just what you need.
Posted in Programming
From time to time I take note of some particularly egregious example of stupidity from the new Luddites. The last time was here; see that post for links to the others. Today, we have another example in Judith Shulevitz’s article in the New Republic on how Siri is messing up a whole generation of children.
Shulevitz starts with a vignette about her daughter playing with her iPhone and interacting with Siri. After a while, the daughter gets frustrated with Siri’s inability to hold a full-fledged conversation and tells her she’s fired. That’s not the problem, though. The danger is that Siri will get smarter and be able to hold better conversations. From there it’s a speedy descent into psychobabble. There’s the problem of Siri’s failure to live up to the “maternal expectations” of the child. Then we move onto the possibility that Siri and other modern interactive devices will engender a slave owner mentality in our children.
I won’t bother pointing out the obvious silliness in all this. You should read it and see for yourself. I’ll say it again to these people: If you don’t like the modern world, you’re welcome to ignore it but, please, leave the rest of us alone.
Posted in General
Surprise! It turns out that the NSA was looking at the phone calls and emails of Americans without a warrant after all. According The Guardian, James Clapper sent a letter to Sen. Ron Wyden admiting to the practice. What’s surprising is not that they are doing this—we all assumed they were—it’s that Clapper admitted it.
The problem for Clapper is that he’s been caught lying to Congress on the public record. When he denied collecting information on millions or tens of millions of Americans, Wyden knew he was lying but couldn’t say so because the information was classified. Now Clapper has put the truth on the record.
That should result in his indictment for lying to Congress but of course it won’t. Even if Congress refers the matter to the Justice Department, the DOJ will ignore it just as they did when Congress referred Attorney General Holder for contempt. These people are corrupt and drunk on power. And why not? They obviously have nothing to fear.
Posted in General
Jean-Philippe Paradis over at HexstreamSoft has a nice page about getting started with the Common Lisp Hyperspec. If you’re already familiar with the Hyperspec, you’re probably with most of the material. If you’re not, it’s a nice introduction.
One thing that I found really helpful was detailed instructions for installing a local copy of the Hyperspec on your computer. If you often find yourself working off line, that can be a real lifesaver.
Paradis has a lot of Common Lisp material on his site so if you’re there for the Hyperspec introduction, take a look around at some of his other offerings.