Agree on this. I remember a talk Lenat gave at IJCAI 87. Funny... he said something along the lines of, "if we feed enough common sense knowledge to CYC, it will get to the point it will be able to read books and newspapers on its own; then its learning process will become automatic". He set a time frame of 10 years to reach that result. I wonder if he really believed it at the time.
As a young kid, I certainly was thrilled! Lenat was a hero, with his work on AM and Eurisko.
Sigh.
Well, in Italy we have this ISP Fastweb who gives us a 10 Mb/s connection. I have fiber right up to my desk! I found it *does* work up to nearly 900Kb/s when connecting to a fast site, even overseas. For instance, when downloading from Microsoft or Adobe.
The downside is that we Fastweb users are behind a NAT firewall, and can't accept incoming network connections. Unless you pay quite a bit more...
Alas, PHP has features designed to make programming quick and easy that also introduce lots of gotchas, like the loose-typing. You can take random input from the user, like $username = $_POST['username'], and execute that *inputted text* like a function if you screw up and put extra parentheses after it:
$other_var = $username();
How likely it is that the program works at all if I make that error? Dangerous errors are things that generally work as intended, but have exploitable flaws.
Handling user input well is critical for security; here is where programmer competence is needed, no matter what programming language you use.
it's not languages that are insecure; no matter what language you use, you need to have competent programmers.
Php is safe and robust, simple and elegant. You can build small or big applications. It is technically *simple*, and that means you don't need books like the one we're discussing here. Just learn the language and you're all set.
Now take an incompetent programmer, and they will build you an insecure system, no matter what language you use.
So it seems you can invent a whole hierarchy of n-dimensions Rubik's puzzle... for n >= 3. But is it possible to invent a 2-dimensional version? Just for people like me who could never figure out the 3D
But, why not do without both EJB and JDO? I find JDBC is a fine interface. You don't really need an object-relational mapping. And for real applications you need to design the database by hand, and design the SQL transactions by hand. You can't get enough speed otherwise.
I have used Smarty for a while before I realized it's a totally unnecessary tool. PHP *is* a templating language already. Instead of loading a Smarty template, just include a php script. You can very easily rewrite Smarty-syntax templates to plain PHP.
The quality of the lines in Italy is very good now. The reason everyone here has got a cellular phone is it's sooo useful... It's a pager on steroids. You can send text messages. You can send voice messages. People use it as a substitute for email. People use it as a PDA. As a substitute for the old answering machine.
The real question is, how come most other rich countries like the USA are so slow in adopting cellular phones?
(Well one thing that never worked very well in Italy is public payphones. That hasn't changed:/
the proper way to code the compare function is something like
int compare(const void *a, const void *b) {
int x = * (int*) a;
int y = * (int*) b;
if (x < y) return -1;
if (x > y) return 1;
return 0; }
Your compare functions were all flawed; the resulting array was likely not sorted at all.
Also, some people complained that casts to int* from void* are inefficient. This is certainly not true! The binary representation of a pointer to "void" does not change when the pointer is cast to int. The casts are just a way to manipulate the C type-check system; in the case of casts from a pointer type to another no change of representation is needed.
Finally, the reason why the templated C++ sort _should_ run faster than the C version is that in C you have the overhead of a function call for each comparison. The beauty of the template mechanism is that if "<" is really the native less-than-between-ints, rather than an overloaded operator, then it will be inlined.
Slashdot Mirror
Agree on this. I remember a talk Lenat gave at IJCAI 87. Funny... he said something along the lines of, "if we feed enough common sense knowledge to CYC, it will get to the point it will be able to read books and newspapers on its own; then its learning process will become automatic". He set a time frame of 10 years to reach that result. I wonder if he really believed it at the time. As a young kid, I certainly was thrilled! Lenat was a hero, with his work on AM and Eurisko. Sigh.
Well, in Italy we have this ISP Fastweb who gives us a 10 Mb/s connection. I have fiber right up to my desk! I found it *does* work up to nearly 900Kb/s when connecting to a fast site, even overseas. For instance, when downloading from Microsoft or Adobe.
The downside is that we Fastweb users are behind a NAT firewall, and can't accept incoming network connections. Unless you pay quite a bit more...
Am I the only one who sees the Caldera logo and thinks "Mickey Mouse"?
Dear Billy Troll,
it's not languages that are insecure; no matter what language you use, you need to have competent programmers.
Php is safe and robust, simple and elegant. You can build small or big applications. It is technically *simple*, and that means you don't need books like the one we're discussing here. Just learn the language and you're all set.
Now take an incompetent programmer, and they will build you an insecure system, no matter what language you use.
So it seems you can invent a whole hierarchy of n-dimensions Rubik's puzzle... for n >= 3. But is it possible to invent a 2-dimensional version? Just for people like me who could never figure out the 3D
But, why not do without both EJB and JDO? I find JDBC is a fine interface. You don't really need an object-relational mapping. And for real applications you need to design the database by hand, and design the SQL transactions by hand. You can't get enough speed otherwise.
I have used Smarty for a while before I realized it's a totally unnecessary tool. PHP *is* a templating language already. Instead of loading a Smarty template, just include a php script. You can very easily rewrite Smarty-syntax templates to plain PHP.
The quality of the lines in Italy is very good now. The reason everyone here has got a cellular phone is it's sooo useful... It's a pager on steroids. You can send text messages. You can send voice messages. People use it as a substitute for email. People use it as a PDA. As a substitute for the old answering machine. The real question is, how come most other rich countries like the USA are so slow in adopting cellular phones? (Well one thing that never worked very well in Italy is public payphones. That hasn't changed :/
Ahem, just for the sake of being pedantic...
the proper way to code the compare function is something like
int compare(const void *a, const void *b) {
int x = * (int*) a;
int y = * (int*) b;
if (x < y) return -1;
if (x > y) return 1;
return 0;
}
Your compare functions were all flawed; the resulting array was likely
not sorted at all.
Also, some people complained that casts to int* from void* are
inefficient. This is certainly not true! The binary representation
of a pointer to "void" does not change when the pointer is cast to
int. The casts are just a way to manipulate the C type-check system;
in the case of casts from a pointer type to another no change of
representation is needed.
Finally, the reason why the templated C++ sort _should_ run faster
than the C version is that in C you have the overhead of a function
call for each comparison. The beauty of the template mechanism is
that if "<" is really the native less-than-between-ints, rather
than an overloaded operator, then it will be inlined.