I don't write in either Java or PHP -- I'm a Grade-A C-slinger -- but here are my REAL-WORLD experiences with both platforms.
C Coders perspective:
PHP - I wrote some objects for PHP5 about eight months ago. The documentation blows, I had to use gdb and a notepad to figure out some of the idiotic details for accessing the symbol table and so forth. The Horrible, horrible dangling-if-macros are terrible. Took 3 days (from "I know nothing" to "I'm done and debugged").
Java - I wrote some JNI interfaces. Actually, they interfaced to the exact same code as the PHP5 modules! (Making a useful C library, encapsulated in C++ objects usable across Java and PHP platforms). Easy stuff! I used cxxwrap. Took 1 day.
Manager's Perspective (I wear that hat, too): PHP is pretty cool, as long as you treat it like a programming language and perform proper data abstraction, code layout, blah-de-blah. "Web guys" need to learn awful fast that "Web Programming" had better be treated the same was as system programming, or large applications become difficult to manage. PHP does little to enforce this (hey, just stick some code right in the middle of the style sheet!), but good discipline will solve all of PHP's major problems.
It's also nice when PHP the guys ask for help, I say "C library function XXXX will solve your problem" or "the underlying OS call behaves this way, that's probably why you're having issues..." and it transliterates directly into PHP. (And I can look at the PHP sources and actually understand them).
Java, on the other hand -- I can't take my years of experience with the UNIX OS and help anybody coding on Java, because it has absolutely nothing in common with the underlying OS, POSIX, etc. Now, that may not be all that bad, but it's damned frustrating when you plan on doing common, every day operations that work anywhere else BUT Java, and have the platform smacks you in the face.
For example, say you need to link two different web hierarchies together (say, images from your apache server and the same images in your tomcat container). You'd make a software link, right? OH, NOOOOO, you make a soft link and then you spend the next three hours figuring out why the fuck it doesn't work, because those asshats who designed the platform didn't like them, so you instead have to hunt through cryptic XML configuration files to find out how to turn on some asshat undocumented directive to allow a BASIC FUNCTION OF THE OPERATING SYSTEM.
For fuck's sake! Now, I suppose the above criticism is more a J2EE criticism than a Java critism, but, if we want to compare apples to apples,.php vs..jsp seems right on target. So let me enunciate very clearly:
J2EE SUCKS HUGE DONKEY BALLS.
Essential redux: Each PHP guy gets more done in a day than two Java guys get done in a week.
Why? How can this be? Java solves everything except world hunger!
The Java guys spend three days a week debugging shit that's gone wrong with Tomcat on one server or another. It's always some incompatibility here, surprise-bite-you-in-the-ass-there. Two applications on the same server use the same JAR file, so the containers refuse to load. That sort of thing. Sheer idiocy.
Then they spend one day debugging shit that's gone wrong with Eclipse (or its mangling of the CVS repository, or some ant dependency problem, or)... then they spend half a day each writing code, and another half day synchronizing their changes. And meanwhile they whine that 256 megs of RAM isn't enough to edit a fucking text file (and do NOTHING else at the same time).
And Lord help you if you want to add another table to the database and want them to do something as silly as retrieve the data from it and put it on a web page. Apparently, this is incredibly difficult, because it involves creating new hibernate objects, which of course fucks everything else in the ass, well, because, something called hotspot didn't get it's monthly fucking hormone shot or som
The Movie Network in Canada has been showing Stargate Atlantis (and some other shows I don't watch) commercial-free for over a year now. That's right, baby, whole episode, 44 minutes, done. I'm not sure how they're paying for it, though -- they'd have to license it from whatever US cable network normally carries it (SciFi?)
I was going to cancel TMN ($15/mo on BEV) as I found myself not watching many movies, but I ain't gonna give up my Stargate Atlantis!!
The device works regardless of compiler implementation -- but it will be faster if it is used with a jump-table.
That said -- the jump-table optimization is SO TRIVIAL that I can't see any arbitrary compiler *not* using it. In fact, it is easier to write the switch directly as a jump table than it is to extract any other semantic meaning out of it, IMHO.
(Of course -- not all compiler writers think in lex and yacc like I do)
> but beg the question of why to do this. > Tricks like this might save, at most, microseconds > of execution time,
Duff used it for exactly the right reason: to feed a bit-blitter fast enough to be useful. Without the loop unrolling, it was 50% too slow on his machine.
I suppose he could have waited for a faster CPU to come along, or written in assembly, but I think his use of his device was exactly correct.
Subsequent uses are primarily useful as Koans on the road to C Enlightenment.
Your understanding of C switch statements is backwards.
The first case expression which evaluates as true is treated as a label to jump into the code block. The code then executes until a break is encountered, or the switch statement terminates. Think about how it compiles.
> I recall a test question in C class that asked if you should always place > a break in each and every case statement. My answer of no was graded > incorrect
You were right. Your teacher was wrong. WRONG WRONG WRONG WRONG!
Consider this code, which reflects a message parser I once wrote:
The code for start_parsing was ~20 lines. This statement most precisely reflects what actually needs to happen -- let alone being clear, consise, short, and effecient. consume_all_smail_headers() was just a 3-line loop which ate up the extra smail headers in the input stream.
Put a giant bar magnet with the north pole facing down in the ceiling.
Give each student a bar magnet. Mark the south pole "yes" and the north pole "no".
Students hold their magnets in the air to indicate the answer.
If the room moves up, the majority of the students chose "yes". If it moves down, the majority of the students chose "no". The more it moves, the more the students are in agreement.
1. Run a piece of angle iron along the front of the bench. Ground the angle iron with 10 or 12 awg wire. Touch this frequently and forget about those PITA straps. Or attach your strap to it if you're a paranoid numpty.;)
2. Grab some anti-static mats. Screw them into your workbench. This will help you avoid scratching the customer's equipment, and if you ground them, further mitigate accidental ESD.
3. Make sure your floor isn't a friggin' carpet. If it is, lay some laminate floor over it.
4. A shelf for monitors. You want a monitor ever 3.5 feet or so.
5. A couple of "test beds", matching your most common configs. These are easy to make. Take the removable part of a case that the motherboard bolts up to, bolt a motherboard to it. Grab a piece of MDF. Bolt the case back to the MDF, and then bolt a harddrive, CD-ROM, and floppy disk to it. You can put this away and take it out easily, and it won't succomb to the "oh I'll just use this..." syndrome so easily.
6. A network jack and a phone jack for each monitor (or more).
7. An internet-connected PC with a floppy drive and a CD-Burner.
There, you'll all done. Assuming you have a shelf full of spares, you can fix anything -- safely and quickly.
That's funny, I know exactly what a variac is, and I thought that testing brown-out conditions like that might not be such a bad idea. A typical PC SMPS should be able to work down to ~100V (in north america)
Slashdot Mirror
235dB???
Holy crap.
What would happen if I was scuba-diving near one of those suckers? Ear drum explosion?
I don't write in either Java or PHP -- I'm a Grade-A C-slinger -- but here are my REAL-WORLD experiences with both platforms.
.php vs. .jsp seems right on target. So let me enunciate very clearly:
C Coders perspective:
PHP - I wrote some objects for PHP5 about eight months ago. The documentation blows, I had to use gdb and a notepad to figure out some of the idiotic details for accessing the symbol table and so forth. The Horrible, horrible dangling-if-macros are terrible. Took 3 days (from "I know nothing" to "I'm done and debugged").
Java - I wrote some JNI interfaces. Actually, they interfaced to the exact same code as the PHP5 modules! (Making a useful C library, encapsulated in C++ objects usable across Java and PHP platforms). Easy stuff! I used cxxwrap. Took 1 day.
Manager's Perspective (I wear that hat, too): PHP is pretty cool, as long as you treat it like a programming language and perform proper data abstraction, code layout, blah-de-blah. "Web guys" need to learn awful fast that "Web Programming" had better be treated the same was as system programming, or large applications become difficult to manage. PHP does little to enforce this (hey, just stick some code right in the middle of the style sheet!), but good discipline will solve all of PHP's major problems.
It's also nice when PHP the guys ask for help, I say "C library function XXXX will solve your problem" or "the underlying OS call behaves this way, that's probably why you're having issues..." and it transliterates directly into PHP. (And I can look at the PHP sources and actually understand them).
Java, on the other hand -- I can't take my years of experience with the UNIX OS and help anybody coding on Java, because it has absolutely nothing in common with the underlying OS, POSIX, etc. Now, that may not be all that bad, but it's damned frustrating when you plan on doing common, every day operations that work anywhere else BUT Java, and have the platform smacks you in the face.
For example, say you need to link two different web hierarchies together (say, images from your apache server and the same images in your tomcat container). You'd make a software link, right? OH, NOOOOO, you make a soft link and then you spend the next three hours figuring out why the fuck it doesn't work, because those asshats who designed the platform didn't like them, so you instead have to hunt through cryptic XML configuration files to find out how to turn on some asshat undocumented directive to allow a BASIC FUNCTION OF THE OPERATING SYSTEM.
For fuck's sake! Now, I suppose the above criticism is more a J2EE criticism than a Java critism, but, if we want to compare apples to apples,
J2EE SUCKS HUGE DONKEY BALLS.
Essential redux: Each PHP guy gets more done in a day than two Java guys get done in a week.
Why? How can this be? Java solves everything except world hunger!
The Java guys spend three days a week debugging shit that's gone wrong with Tomcat on one server or another. It's always some incompatibility here, surprise-bite-you-in-the-ass-there. Two applications on the same server use the same JAR file, so the containers refuse to load. That sort of thing. Sheer idiocy.
Then they spend one day debugging shit that's gone wrong with Eclipse (or its mangling of the CVS repository, or some ant dependency problem, or)... then they spend half a day each writing code, and another half day synchronizing their changes. And meanwhile they whine that 256 megs of RAM isn't enough to edit a fucking text file (and do NOTHING else at the same time).
And Lord help you if you want to add another table to the database and want them to do something as silly as retrieve the data from it and put it on a web page. Apparently, this is incredibly difficult, because it involves creating new hibernate objects, which of course fucks everything else in the ass, well, because, something called hotspot didn't get it's monthly fucking hormone shot or som
Yes, of course, there is the one that was uncovered at the dig in Giza in the 1930s. But that sucker blew up on the Dark Side of the Moon.
And the other was in Antarctica. It's locked up in Cheyenne mountain right now.
What about the house committee on unamerican activities?
They don't need a search engine, just a Smell-o-scope.
Wake me when maps.search.ch covers Canada as well as Google Maps.
That's nothing. I tried to search for prostitutes in my home town, and it gave me a map to all the churches!!
You know what I really can't figure out?
The Movie Network in Canada has been showing Stargate Atlantis (and some other shows I don't watch) commercial-free for over a year now. That's right, baby, whole episode, 44 minutes, done. I'm not sure how they're paying for it, though -- they'd have to license it from whatever US cable network normally carries it (SciFi?)
I was going to cancel TMN ($15/mo on BEV) as I found myself not watching many movies, but I ain't gonna give up my Stargate Atlantis!!
Hmm, maybe that's how they're paying for it..
The device works regardless of compiler implementation -- but it will be faster if it is used with a jump-table.
That said -- the jump-table optimization is SO TRIVIAL that I can't see any arbitrary compiler *not* using it. In fact, it is easier to write the switch directly as a jump table than it is to extract any other semantic meaning out of it, IMHO.
(Of course -- not all compiler writers think in lex and yacc like I do)
> but beg the question of why to do this.
> Tricks like this might save, at most, microseconds
> of execution time,
Duff used it for exactly the right reason: to feed a bit-blitter fast enough to be useful. Without the loop unrolling, it was 50% too slow on his machine.
I suppose he could have waited for a faster CPU to come along, or written in assembly, but I think his use of his device was exactly correct.
Subsequent uses are primarily useful as Koans on the road to C Enlightenment.
Your understanding of C switch statements is backwards.
The first case expression which evaluates as true is treated as a label to jump into the code block. The code then executes until a break is encountered, or the switch statement terminates. Think about how it compiles.
> I recall a test question in C class that asked if you should always place
> a break in each and every case statement. My answer of no was graded
> incorrect
You were right. Your teacher was wrong. WRONG WRONG WRONG WRONG!
Consider this code, which reflects a message parser I once wrote:
switch (message_type)
{
mt_smail_spool_file:
consume_all_smail_headers();
mt_rfc822:
start_parsing();
break;
default:
err("invalid message type");
}
The code for start_parsing was ~20 lines. This statement most precisely reflects what actually needs to happen -- let alone being clear, consise, short, and effecient. consume_all_smail_headers() was just a 3-line loop which ate up the extra smail headers in the input stream.
What kind of DB do you want?
Berkely DB and variants (GNU DB) have been shipping with everything for about as long as I can remember.
Now, if you want an SQL RDBMS, that's a different story...
Warez.
Try pricing out oracle for a mid-high-end Sun system with eight CPUs.
You'll shit your pants.
Yes. It should analyze the porn in users' home dirs, then seek out new, but similar porn. Sort of like tivo.net, only for Linux geeks.
Wow -- if you drive that rarely, I hope you use fuel stabilizer in your tank!
You should run through a tank of gas once every three months if it's not stabilized.
Put the room on springs.
Put a giant bar magnet with the north pole facing down in the ceiling.
Give each student a bar magnet. Mark the south pole "yes" and the north pole "no".
Students hold their magnets in the air to indicate the answer.
If the room moves up, the majority of the students chose "yes". If it moves down, the majority of the students chose "no". The more it moves, the more the students are in agreement.
Best of all, the batteries will never die.
Unless you drop the answer sticks.
My dad's job used to involve, among other things, putting up guy wires for antennas, towers, and poles.
You should see him solving wierd trig problems using only his head and a slide rule. Faster than I could even input them into a computer...
My teachers would accept answers like "sqrt(c)" (except written with the square-root symbol).
Rumour has it that one year a freshman ran out of time on a trig exam, stuck on the first question, trying to write out the square root of two..
(Putting on pointy-hair wig)
MD5 is deprecated, but every web server still supports basic authentication, which uses Base64. Hmm.
64 is much bigger than 5, so it must be better.
Yup. No more digest authentication, only basic will be supported! Another security problem averted; quick: call the press!
> BTW....if anyone out there has any older Conner or
> Western Digital (pre-Caviar) 20-40-120-240mb hard drive
You mean, ST-501 interface disks?
Shit, I threw about 20 of them out... about 10 years ago.
1. Run a piece of angle iron along the front of the bench. Ground the angle iron with 10 or 12 awg wire. Touch this frequently and forget about those PITA straps. Or attach your strap to it if you're a paranoid numpty. ;)
2. Grab some anti-static mats. Screw them into your workbench. This will help you avoid scratching the customer's equipment, and if you ground them, further mitigate accidental ESD.
3. Make sure your floor isn't a friggin' carpet. If it is, lay some laminate floor over it.
4. A shelf for monitors. You want a monitor ever 3.5 feet or so.
5. A couple of "test beds", matching your most common configs. These are easy to make. Take the removable part of a case that the motherboard bolts up to, bolt a motherboard to it. Grab a piece of MDF. Bolt the case back to the MDF, and then bolt a harddrive, CD-ROM, and floppy disk to it. You can put this away and take it out easily, and it won't succomb to the "oh I'll just use this..." syndrome so easily.
6. A network jack and a phone jack for each monitor (or more).
7. An internet-connected PC with a floppy drive and a CD-Burner.
There, you'll all done. Assuming you have a shelf full of spares, you can fix anything -- safely and quickly.
But a cordless drill without an adjustable clutch is not worth the plastic it's made out of.
If I have to see one more freakin' rounded phillips screw, I'm going to stab somebody with a screwdriver.
That's funny, I know exactly what a variac is, and I thought that testing brown-out conditions like that might not be such a bad idea. A typical PC SMPS should be able to work down to ~100V (in north america)
Holy Mother of Christ!
2 6&spn=0.002198,0.003720&t=e&hl=en
6 0&spn=0.002198,0.003720&t=e&hl=en
http://maps.google.com/maps?ll=29.988463,-90.0457
http://maps.google.com/maps?ll=29.998819,-90.0374
Wow, another 50-75, this time school busses: http://maps.google.com/maps?q=new+orleans+school+b us&ll=29.954200,-90.047657&spn=0.002199,0.003720&t =e&num=10&start=0&hl=en
o n&ll=29.951737,-90.034955&spn=0.002199,0.003720&t= e&num=10&start=0&hl=en
A few here -- maybe part of the rescue op? Nah:
http://maps.google.com/maps?q=new+orleans+educati