Some time ago I tried an image search engine which works similarities in various kinds of features of the images. The demonstration version was quite impressive, but unfortunately they had only indexed an internal image collection. I can't find the link right now, but I seem to recall that the project was hosted by a Swedish university.
I don't mind EU spending a bit of research funds on developing search engine technology, but the PR around this project is a bit unfortunate.
If USA hadn't kicked a whole country's DNS off-line before, the promise in the Economist might be credible. But since but ".dk" has been taken off-line once by USA for not paying for the domain, I don't really trust such a promise.
/Jacob
PS: No, the countries do not pay for their domains.
Which is bad. But I still find taking the logarithm of a line count worse, and by any sensible scientific measure undefined.
Now, I'm all for emperical data, but that is just bistromatics and totally insane.
That's not fair to bistromatics.
I have no doubt that if you take a piece of code and does a before&after check after some major rewriting it may tell you something.
I doubt that. I could probably hack a script, that could optimize the "maintainability" of the code after the major rewriting. Just adjusting the number of comment lines can apparently do wonders to the maintainability of a program.
But comparing two different pieces of code with this formula is just plain bogus.
Using that formula for anything is plain bogus. You may be lucky that there is a correlation, but a function just selected for optimal fitting to a few data-points without any (apparent) regard to simple things like keeping units straight and some version of common sense, has no place in science. I'm actually quite disqusted that ACM Queue would print that article.
Besides the general problem of not keeping the units straight, I find it very hard to imagine that a single formula can be used to measure the maintainability independently of programming language. And how can anybody argue that a comments to code ratio different from zero is universally optimal? (zero is trivial to argue for, but I have a feeling it isn't correct)
Of course there are some tasks for which C or C++ are the still best choice for other reasons, so I still use both frequently and applaud any attempt at adding better security scanning to the compiler.
It would probably be stupid to disagree completely
with that, so I will just disagree a little bit.
One of the arguments I have heard most often for
using C or C++ instead of Java (my acquaintances don't seem to be interested in C#) is that they
gives you better access to the hardware.
While that's correct, I like to use Ada, where I both get a fair deal of checking done by the compiler and reasonably easy access to the raw hardware.
I can't help thinking, though, that even in those cases a language with the granularity of C but with built-in strings (UTF-8), arrays that are checked by default but with an override, with fixed built in data types (e.g., a 'byte' type that isn't signed in some places and unsigned in others), and yet without all of the massive baggage of C++, would go a long way to improving C's bug proneness without removing its power.
I am not sure I understand you completely (exactly what do you mean by "with the granularity of C"? that everything doesn't have to be a class?), but I think you might appreciate languages like Ada and Eiffel (although none of them have UTF-8 encoded strings built in AFAIK).
I haven't, but I believe in freedom of speach no matter what they happen to show on al-Jazeera.
If not you should really take a look before slamming me. It is one thing to report the news. Quite another to make news and incite riots. Al-Jazeera is running an "America The Great Satan" commercial every half hour broadcast.
Freedom of speach is not limited to saying nice
things. It also allows people to say rather
horrible and unpleasant things. But since we
don't all agree on what is horrible and what is
not, it has to be like that.
Jacob
PS: IMO freedom of speach is not the same as freedom from responsibility.
PPS: Considering the current actions of USA, UK and DK (did I leave anybody out?), I find it hard to disapprove of people who call USA "The Great Satan".
Slashdot Mirror
Some time ago I tried an image search engine which works similarities in various kinds of features of the images. The demonstration version was quite impressive, but unfortunately they had only indexed an internal image collection. I can't find the link right now, but I seem to recall that the project was hosted by a Swedish university.
I don't mind EU spending a bit of research funds on developing search engine technology, but the PR around this project is a bit unfortunate.
JacobIf USA hadn't kicked a whole country's DNS off-line before, the promise in the Economist might be credible. But since but ".dk" has been taken off-line once by USA for not paying for the domain, I don't really trust such a promise.
PS: No, the countries do not pay for their domains.
Yes. Unfortunately.
Which is bad. But I still find taking the logarithm of a line count worse, and by any sensible scientific measure undefined.
That's not fair to bistromatics.
I doubt that. I could probably hack a script, that could optimize the "maintainability" of the code after the major rewriting. Just adjusting the number of comment lines can apparently do wonders to the maintainability of a program.
Using that formula for anything is plain bogus. You may be lucky that there is a correlation, but a function just selected for optimal fitting to a few data-points without any (apparent) regard to simple things like keeping units straight and some version of common sense, has no place in science. I'm actually quite disqusted that ACM Queue would print that article.
Besides the general problem of not keeping the units straight, I find it very hard to imagine that a single formula can be used to measure the maintainability independently of programming language. And how can anybody argue that a comments to code ratio different from zero is universally optimal? (zero is trivial to argue for, but I have a feeling it isn't correct)
Jacob
It would probably be stupid to disagree completely with that, so I will just disagree a little bit. One of the arguments I have heard most often for using C or C++ instead of Java (my acquaintances don't seem to be interested in C#) is that they gives you better access to the hardware. While that's correct, I like to use Ada, where I both get a fair deal of checking done by the compiler and reasonably easy access to the raw hardware.
I am not sure I understand you completely (exactly what do you mean by "with the granularity of C"? that everything doesn't have to be a class?), but I think you might appreciate languages like Ada and Eiffel (although none of them have UTF-8 encoded strings built in AFAIK).
</ada-propaganda> :-)
I haven't, but I believe in freedom of speach no matter what they happen to show on al-Jazeera.
Freedom of speach is not limited to saying nice things. It also allows people to say rather horrible and unpleasant things. But since we don't all agree on what is horrible and what is not, it has to be like that.
Jacob
PS: IMO freedom of speach is not the same as freedom from responsibility.
PPS: Considering the current actions of USA, UK and DK (did I leave anybody out?), I find it hard to disapprove of people who call USA "The Great Satan".