Archive

Archive for November, 2013

Ordinary Least Squares is dead to me

November 28, 2013 12 comments

Most books that discuss regression modeling start out and often finish with Ordinary Least Squares (OLS) as the technique to use; Generalized Linear ModelLeast Squares (GLMS) sometimes get a mention near the back. This is all well and good if the readers’ data has the characteristics required for OLS to be an applicable technique. A lot of data in the social sciences has these characteristics, or so I’m told; lots of statistics books are written for social science students, as a visit to a bookshop will confirm.

Software engineering datasets often range over several orders of magnitude or involve low value count data, not the kind of data that is ideally suited for analysis using OLS. For this kind of data GLMS is probably the correct technique to use (the difference in the curves fitted by both techniques is often small enough to be ignored for many practical problems, but the confidence bounds and p-values often differ in important ways).

The target audience for my book, Empirical Software Engineering with R, are working software developers who have better things to do that learn lots of statistics. However, there is no getting away from the fact that I am going to have to make extensive use of GLMS, which means having to teach readers about the differences between OLS and GLMS and under what circumstances OLS is applicable. What a pain.

Then I had a brainwave, or a moment of madness (time will tell). Why bother covering OLS? Why not tell readers to always use GLMS, or rather use the R function that implements it, glm. The default glm behavior is equivalent to lm (the R function that implements OLS); the calculation is not being done by hand but by a computer (i.e., who cares if it is more complicated).

Perhaps there is an easy way to explain this to software developers: glm is the generic template that can handle everything and lm is a specialized template that is tuned to handle certain kinds of data (the exact technical term will need tweaking for different languages).

There is one user interface issue, models built using glm do not come with an easy to understand goodness of fit number (lm has the R-squared value). AIC is good for comparing models but as a single (unbounded) number it is not that helpful for the uninitiated. Will the demand for R-squared be such that I will be forced for tell readers about lm? We will see.

How do I explain the fact that so many statistics books concentrate on OLS and often don’t mention GLMS? Hey, they are for social scientists, software engineering data requires more sophisticated techniques. I will have to be careful with this answer as it plays on software engineers’ somewhat jaded views of social scientists (some of whom have made very major contribution to CRAN).

All the software engineering data I have seen is small enough that the performance difference between glm/lm is not a problem. If performance is a real issue then readers will search the net and find out about lm; sorry guys but I want to minimise what the majority of readers need to know.

R now has its own shelf in Dillons

November 25, 2013 No comments

I was in Dillons, the one opposite University College London, at the start of the week and what did I spy there?

Programming language books

There is now a bookshelf devoted to R (right, second from top) in the programming languages section. The shelf would be a lot fuller if O’Reilly did not have a complete section devoted to their books.

A trolley of C/C++ books was waiting to refill the shelves near the door.

Programming language books

Being adjacent to a university means that programming language books make up a much larger percentage of software books.

Programming language books

And there is O’Reilly in the corner with two stacks of shelves.

Programming language books

And yes, this is a big bookshop, the front is a complete block; computing/mathematics/physics/chemistry/engineering/medicine are in the basement. You can buy skeletons and stethoscopes in the medical section a few rooms down from computing; a stethoscope is useful for locating strange noises in computer cases without having to open them.

Programming language books

Readers a bit younger than me probably know this shop as Waterstones.

What is the error rate for published mathematical proofs?

November 17, 2013 2 comments

Mathematical proofs are sometimes cited as the gold standard against which software quality should be compared. At school we rarely get to hear about proofs that turn out to be wrong and are inculcated with the prevailing wisdom that all mathematical proofs are correct.

There are many technical and social issues involved in believing a published proof and well known established mathematicians have no trouble pointing out that “… it is impossible to write out a very long and complicated argument without error, …

Examples of incorrect published proofs include Wiles’ first proof of Fermat’s Last Theorem and an serious error found in a proof of a message signing scheme.

A question on mathoverflow contains a list of rather interesting false proofs.

Then, of course, there are always those papers that appear in journals that get written about more frequently on Retraction Watch than others.

What is the error rate for published mathematical proofs? I have not been able to find any collection of mathematical proof error data.

Several authors have expressed the view that because there so many diverse mathematical topics being studied these days there are very few domain experts available to check proofs. A complicated proof of a not particularly interesting result is unlikely to attract the attention needed to check it thoroughly. It should come as no surprise that the number of known errors in such proofs is equal to the number of known errors in programs that have never been executed.

Proofs are different from programs in that one error can be enough to ‘kill-off’ a proof, while a program can contain many errors and still be useful. Do errors in programs get talked about more than errors in proofs? I rarely get to socialize with working mathematicians and so cannot make any judgment call on this question.

Every non-trivial program is likely to contain many errors; can the same be said for long mathematical proofs? Are many of these errors as trivial (in the sense that they are easily fixed) as errors in programs?

One commonly used error rate for programs is errors per line of code; how should the rate be expressed for proofs? Errors per page, per line, per definition?

Lots of questions and I’m hoping one of my well informed readers will be able to provide some answers or at least cite a reference that does.

Testing compiler semantics with minimal manual input

November 11, 2013 3 comments

The 2011 revision of the C++ Standard added lots of new constructs to the language and in the past few months both the GCC and LLVM teams have been claiming that the next release of their C++ compilers will fully support the 2011 Standard. How true are these claims? One way of answering this question is to run both compilers over an extensive test suite. There are commercial C++ compiler test suites available, but I don’t have access to them and if I did the license agreement would probably not allow me to talk in detail about the results. Writing compiler tests cases requires a very detailed knowledge of the language; I have done it often enough in previous lives to know that more than a year or so of my time would be needed just to get my head around the semantics of the new C++ features, before I could produce anything half decent.

Is there a way of automating the generation of test cases for language semantics? Automated statement/expression generation is very effective at finding problems with optimizers and code generators. Can this technique be applied to check the semantic requirements of a language?

Having concocted various elaborate schemes over the years I recently realised that life would be a lot simpler if I was willing to accept a very high percentage of erroneous test programs (the better manually written test suites usually contain around as many test cases that intended to fail to compile as tests that pass, i.e., intended to compile correctly; the not so good ones have few failing tests).

If two or more compilers are available the behavior of each of them on a given source file can be compared: differential testing. If both compile a file or fail to compile it, they may both be right or wrong; either way this shared behavior is not interesting, but is likely to be the common case. The interesting case is if one compiles a file and the other fails to compile it; this could be a fault in one of them, or one of those cases where the Standard permits compilers to do their own thing.

I hereby jump to the conclusion that behavior differences is a good proxy for compiler conformance to the language Standard (actually developers are often more interested in all compilers they are likely to use having the same external behavior than conforming to a Standard).

Lets implement this (source code here)!

First we need to generate lots of test cases. The process I used is based on program templates, such as the following (lines starting with ! are places where various constructs can be inserted):

int v;
 
! declare_id 2
 
int main(void)
{
! declare_id 2
 
! ref_id 2
 
}

the identifier after the ! is the name of a file containing lines to be inserted at the given location (the number after the identifier is the maximum number of lines that can be inserted at that point; default 1 if no number given). The following is example file contents for the above template:

declare_id

int i;
enum {i, j};
enum i {x, y};
struct i {int f;};
typedef int i;

ref_id

enum i ev_i;
struct i sv_i;
typedef i tv_i;
v=i;

It is then simply a matter of permuting through all of the possible combinations of lines that can be inserted in the program template, creating a stand alone file for each possibility (18,000 of them in the above example); I used the Python Natural Language Toolkit to do the heavy lifting.

A shell script compiles each source file and compares the compiler exit codes. For the above example there were 16,366 failures, 1,634 passes and no differences (this example contains well established C constructs and any difference would have been surprising).

Next, a feature new in C++11, lambda functions!

Here is the template used:

! declare_xy 2
 
int main(void)
{
 
! declare_xy 2
 
auto foo_bar =
! define_lambda
;
 
return 0;
}

I cut and pasted some examples from the Standard to create the following optional lines:

define_lambda

[](float a, float b) { return a + b; }
[=](float a, float b) { return a + b; }
[=,x](float a, float b) { return a + b; }
[y](float a, float b) { return a + b; }
[=]()->int { return operator()(this->x + y); }
[&, i]{ }
[=] { decltype(x) y1; decltype((x)) y2 = y1; decltype(y) r1 = y1; decltype((y)) r2 = y2; }

which generated 6,300 source files of which 5,865 failed, 396 passed and 39 were treated differently by the compilers (g++ version 4.7.2, clang version 3.3).

How should the percentages be calculated? If we take the human written numbers for well written test suites containing (roughly) equal numbers of pass/fail tests, then we have around 800 tests of which (say) 40 gave different behavior, giving us a 5% fault rate. Do we share that 5% equally between both compilers or assign 3% for both being wrong and 1% for each being uniquely wrong?

Submitting a bug report to both compiler teams pointing out that their behavior is different from the other’s is a sure fire way to make myself unpopular. Any suggestions for how to resolve this issue, that does not involve me having to study the tiresomely long and convoluted C++ Standard, welcome.