In combination with the DOS/4GW dos extender, it provided a flat 32-bit address space for programming. It also produced the better code than any compiler, by a wide margin. Back then, Microsoft's compiler for DOS was broken (it produced buggy code when optimizations were enabled). Borland's Turbo C++ was cool, and some games used it, like Wolfenstein 3D and Jazz Jackrabbit, but for games for the 486 like DOOM, nothing compared to Watcom. It also had really nice tools, like a cool VI like editor for DOS, an excellent debugger and profiler, and lots of other cool stuff. Not to mention that it would compile for DOS and Windows in both 16 and 32 modes, and OS/2. It was the first C/C++ compiler that I really started using heavily back in the day, and I must say that it was a very excellent product, worth every penny (I even got the student version for $99).
There is probably a compiler or two now that produces better code for x86 (especially PIII and Athlon), like Intel's compiler, but being that most Linux distributions are compiled for 386 or possibly Pentium, there would be a ton of stuff that would benefit from a Watcom compiler for Linux.
Encryption can partially solve this problem. Updates are already signed by Microsoft, the same way device drivers are signed. Presumably, the automatic updater would only updates that had a proper signature. Though, there was that story a while ago about someone else getting a Microsoft certificate out of VeriSign...
But I'd like to point out that the exploits on IIS are due to out of the box configuration problems. If you follow the security checklist for IIS, it doesn't get infected, even if unpatched.
You'd also need the broadband adapter, which is not exactly cheap. The total cost would be more than a $100 off the shelf broadband router. And what good is a router with only a single interface? Plus it's probably only 10mb. If you have access to old hardware (a Pentium, say) then it would be just as cheap to use a real PC.
The solution is trivial, ftp the package to your drive, then type pkg_add foo{tab}. All done!
But the point is that with Debian, you don't need to FTP it manually. If I want to install Apache, I just type "apt-get install apache". Having to log into the FTP server, find the file and download it, or open a web browser and look around on a page takes a lot longer.
Don't get me wrong. I have used FreeBSD for the past year at work and I love being able to compile stuff from ports with custom options (although with Debian, there are packages avaiable for most common options). But compiling takes a long time, especially on slower hardware. Building the minimal X libraries to install PHP with GD can take several hours. Being able to "apt-get install php-gd" is very nice. If you could combine apt with the port's system of making it easy to compile things custom and build packages, you'd have the perfect packaging system.
Re:Training and Planning are the keys.
on
KDE 2.2.1 Up
·
· Score: 1
Debian's package management does transparent upgrades. The apt-get tool automatically calculates dependencies, and downloads and installs the appropriate packages. Conflict resolution is automatic. It is very similiar to FreeBSD's ports in some aspects, and it is likely that at least one of the projects used ideas from the other. I wish FreeBSD did upgrades as as nicely as Debian, and I wish Debian made it easy to build from source with different options (i.e. present a menu).
For a desktop or workstation, I can't see an easier way to do package management. Debian was the first Linux distro that I started using, because my friends who used Linux used it, and I can't imagine switching to anything else. I've tried RPM based systems, and installing anything is a mess. Getting the right version and the right dependencies must be done manually and is a chore. I'm sure there are tools now to do it for you, but Debian has had for at least several years and it works great. I run sid (the unstable distribution version) and upgrade my system daily with the two commands shown below. Everything upgrades seamlessly, and when it needs to do something like upgrade a configuration file, it gives an option to tell it what to do, but otherwise it just does it's thing without requiring any intervention.
If the monitor isn't running, a computer shouldn't use more than about 10-20W. A hefty power supply is only necessary for an AGP graphics card that uses a lot of power, or when spinning up the disk drives.
The google cache only caches the HTML pages, not images or any other embeded objects. HTML is maybe 5-10% of the total bandwidth used for a site (much less if it is a porn site).
I attended his speech at the conference. I could barely understand what he was saying, and I imagine it was like that for everyone else there too. Unlike most of the speakers there, he wasn't really giving a speech, but rather reading a pre-written speech. My thoughts at the time were that it could have been a better presentation if a native English speaker were to have read his speech, with him answering questions. It was obvious when he was reading and when he was not.
There was a short television interview with him available online a few weeks ago, and his English was very poor.
I agree with the original poster that he will not be able to articulate his defense as well as a native speaker.
Is arguing always a bad thing? We have about 10 programmers at our company, and the only two that ever argue with me on issues are the ones that are on the same skill level as me. Everyone else just agrees with everything. While there's not always a right and a wrong way to do something, I think it's important to be in agreement with other good programmers about things. Either we just have a different opinion about something that could be done multiple ways, or one of us is wrong. If the other programmer has a different opinion then me, it should be resolved. I want to know why he feels a certain way about something. Eventually we will have to decide how to do something, and everyone should understand why the decision was made. I can't work in an environment where something is done poorly because "that's the way it's always been done" or "that's just the way things are".
I agree with you on the issue of confusing people with nonsense programming questions. That seems to be a common practice at university course tests, even though it makes no sense. I took the second level C++ course my first semester freshman year in college, and found it to be relatively easy. Just straight forward object oriented programming, not even getting into STL or templates. Judging from talking to many people in the class and knowing how long it took people to get the assignments done, I know I was in the top 5 programmers in a class of around 150. I'd had about three years of C++ experience, mostly self taught by programming and from books, along with high school AP Computer Science thrown in there (got a 5 on the AB test). The point being I considered myself to be a good C++ programmer. But some of the questions on tests had questions that demonstrated some of the more archane C/C++ language "features". I wasn't able to be 100% sure that I answered them right. I have no idea how people who were new to programming and C++ were supposed to know these things. You certainly wouldn't get a good grade if you wrote code like that on an assignment, yet it was on the test questions.
Perhaps they will use wireless, although I would guess that many more people have ethernet cards than have 802.11 cards. The article did say that "MSN is still testing the best technologies to achieve this" and didn't specifically mention ethernet.
That's incorrect, because BSD (at least FreeBSD, don't know about the others) will run Linux binaries. So that statically linked Linux binary will work on BSD. It's also not a fair comparison, because each BSD is really a different operating system (different kernel) while each Linux distribution uses the same kernel (though the original poster did not make that distinction).
The guy obviously didn't know a lot about PHP. The first example made Perl look better than PHP, but that PHP example can look almost exactly like the Perl one:
<?php
$title = "My first Script";
$greeting = "Welcome to my first script.";
print <<<EOF
I think it all depends on what you are doing. As for it being faster, the page still has to be parsed just the same, and with Zend, the difference in speed would most likely be negligable. If you have several hundred kilobytes of HTML on a page, then not outputting it through PHP might make a noticable difference in speed, but otherwise it shouldn't make a difference.
If PHP is just used to output a few variables, then it is easier to read to not have the HTML inside the PHP code. But if most of the page is generated dynamically, then it ends up being cleaner to make it all PHP code. Most of my scripts have only one set of in them. However, for any real site or application, it makes sense to use templates for any HTML generation, thus allowing the designers to change the look of the page without ever having to touch the code, along with being able to internationalize the output without any extra work besides the actual translation.
You can't really call PHP an interpreted langauge anymore, and I believe the same is true of Perl. PHP 4 uses the Zend engine, which compiles PHP before executing it, the same way that Java must be compiled before it can be executed. I would assume the same is true with most other scripting languages, such as ColdFusion and ASP. The difference is that the compiling is done at run time, so to update some code it is only necessary to save the file. PHP (with add-ons) and ColdFusion can both cache the intermediate compiled code, so that it doesn't have to be compiled everytime. I wouldn't be surprised if Perl (with mod_perl?) and ASP had this sort of feature.
I agree completely that a real comparison would be useful. Coming from someone who's day job is doing web applications, I'll say that programmer time is much more expensive than hardware, so if it takes less time to develop and maintain an application, having to add an extra server, more ram or a faster cpu is going to end up being cheaper in the long run.
The problem is that AOL can't add anything to their client as a security feature. If a majority of the existing clients don't already have it, then it won't work. Just mimick the last version of the official client that didn't have the security feature.
Slashdot Mirror
In combination with the DOS/4GW dos extender, it provided a flat 32-bit address space for programming. It also produced the better code than any compiler, by a wide margin. Back then, Microsoft's compiler for DOS was broken (it produced buggy code when optimizations were enabled). Borland's Turbo C++ was cool, and some games used it, like Wolfenstein 3D and Jazz Jackrabbit, but for games for the 486 like DOOM, nothing compared to Watcom. It also had really nice tools, like a cool VI like editor for DOS, an excellent debugger and profiler, and lots of other cool stuff. Not to mention that it would compile for DOS and Windows in both 16 and 32 modes, and OS/2. It was the first C/C++ compiler that I really started using heavily back in the day, and I must say that it was a very excellent product, worth every penny (I even got the student version for $99).
There is probably a compiler or two now that produces better code for x86 (especially PIII and Athlon), like Intel's compiler, but being that most Linux distributions are compiled for 386 or possibly Pentium, there would be a ton of stuff that would benefit from a Watcom compiler for Linux.
Encryption can partially solve this problem. Updates are already signed by Microsoft, the same way device drivers are signed. Presumably, the automatic updater would only updates that had a proper signature. Though, there was that story a while ago about someone else getting a Microsoft certificate out of VeriSign...
Or pr0n :) When you have racks of servers pushing 10-80mbit constantly, a few hundred extra hits a second for the "Slashdot Effect" is nothing.
Sounds like you do have a lot to do :)
But I'd like to point out that the exploits on IIS are due to out of the box configuration problems. If you follow the security checklist for IIS, it doesn't get infected, even if unpatched.
You'd also need the broadband adapter, which is not exactly cheap. The total cost would be more than a $100 off the shelf broadband router. And what good is a router with only a single interface? Plus it's probably only 10mb. If you have access to old hardware (a Pentium, say) then it would be just as cheap to use a real PC.
But the point is that with Debian, you don't need to FTP it manually. If I want to install Apache, I just type "apt-get install apache". Having to log into the FTP server, find the file and download it, or open a web browser and look around on a page takes a lot longer.
Don't get me wrong. I have used FreeBSD for the past year at work and I love being able to compile stuff from ports with custom options (although with Debian, there are packages avaiable for most common options). But compiling takes a long time, especially on slower hardware. Building the minimal X libraries to install PHP with GD can take several hours. Being able to "apt-get install php-gd" is very nice. If you could combine apt with the port's system of making it easy to compile things custom and build packages, you'd have the perfect packaging system.
Debian's package management does transparent upgrades. The apt-get tool automatically calculates dependencies, and downloads and installs the appropriate packages. Conflict resolution is automatic. It is very similiar to FreeBSD's ports in some aspects, and it is likely that at least one of the projects used ideas from the other. I wish FreeBSD did upgrades as as nicely as Debian, and I wish Debian made it easy to build from source with different options (i.e. present a menu).
For a desktop or workstation, I can't see an easier way to do package management. Debian was the first Linux distro that I started using, because my friends who used Linux used it, and I can't imagine switching to anything else. I've tried RPM based systems, and installing anything is a mess. Getting the right version and the right dependencies must be done manually and is a chore. I'm sure there are tools now to do it for you, but Debian has had for at least several years and it works great. I run sid (the unstable distribution version) and upgrade my system daily with the two commands shown below. Everything upgrades seamlessly, and when it needs to do something like upgrade a configuration file, it gives an option to tell it what to do, but otherwise it just does it's thing without requiring any intervention.
apt-get update
apt-get dist-upgrade
If the monitor isn't running, a computer shouldn't use more than about 10-20W. A hefty power supply is only necessary for an AGP graphics card that uses a lot of power, or when spinning up the disk drives.
If you live in Arizona where it's 105 degrees on a good day, then air conditioning really is a NEED!
I hope that comment is sarcastic. Do you realize how many hours in developer time can be lost by something like this?
The google cache only caches the HTML pages, not images or any other embeded objects. HTML is maybe 5-10% of the total bandwidth used for a site (much less if it is a porn site).
I attended his speech at the conference. I could barely understand what he was saying, and I imagine it was like that for everyone else there too. Unlike most of the speakers there, he wasn't really giving a speech, but rather reading a pre-written speech. My thoughts at the time were that it could have been a better presentation if a native English speaker were to have read his speech, with him answering questions. It was obvious when he was reading and when he was not. There was a short television interview with him available online a few weeks ago, and his English was very poor. I agree with the original poster that he will not be able to articulate his defense as well as a native speaker.
Is arguing always a bad thing? We have about 10 programmers at our company, and the only two that ever argue with me on issues are the ones that are on the same skill level as me. Everyone else just agrees with everything. While there's not always a right and a wrong way to do something, I think it's important to be in agreement with other good programmers about things. Either we just have a different opinion about something that could be done multiple ways, or one of us is wrong. If the other programmer has a different opinion then me, it should be resolved. I want to know why he feels a certain way about something. Eventually we will have to decide how to do something, and everyone should understand why the decision was made. I can't work in an environment where something is done poorly because "that's the way it's always been done" or "that's just the way things are".
I agree with you on the issue of confusing people with nonsense programming questions. That seems to be a common practice at university course tests, even though it makes no sense. I took the second level C++ course my first semester freshman year in college, and found it to be relatively easy. Just straight forward object oriented programming, not even getting into STL or templates. Judging from talking to many people in the class and knowing how long it took people to get the assignments done, I know I was in the top 5 programmers in a class of around 150. I'd had about three years of C++ experience, mostly self taught by programming and from books, along with high school AP Computer Science thrown in there (got a 5 on the AB test). The point being I considered myself to be a good C++ programmer. But some of the questions on tests had questions that demonstrated some of the more archane C/C++ language "features". I wasn't able to be 100% sure that I answered them right. I have no idea how people who were new to programming and C++ were supposed to know these things. You certainly wouldn't get a good grade if you wrote code like that on an assignment, yet it was on the test questions.
Perhaps they will use wireless, although I would guess that many more people have ethernet cards than have 802.11 cards. The article did say that "MSN is still testing the best technologies to achieve this" and didn't specifically mention ethernet.
But if it costs you more money to write it than it does to buy it, then it can make more sense to buy it.
That's incorrect, because BSD (at least FreeBSD, don't know about the others) will run Linux binaries. So that statically linked Linux binary will work on BSD. It's also not a fair comparison, because each BSD is really a different operating system (different kernel) while each Linux distribution uses the same kernel (though the original poster did not make that distinction).
You're right, because on x86 values are stored backwards internally.
The poster above made the original post on the kernel mailing list...
The guy obviously didn't know a lot about PHP. The first example made Perl look better than PHP, but that PHP example can look almost exactly like the Perl one:
<?php
$title = "My first Script";
$greeting = "Welcome to my first script.";
print <<<EOF
<html>
<head>
<title>$title</title>
</head>
<body>
<h1>$title</h1>
<p>$greeting</p>
</body>
</html>
EOF
?>
I think it all depends on what you are doing. As for it being faster, the page still has to be parsed just the same, and with Zend, the difference in speed would most likely be negligable. If you have several hundred kilobytes of HTML on a page, then not outputting it through PHP might make a noticable difference in speed, but otherwise it shouldn't make a difference.
If PHP is just used to output a few variables, then it is easier to read to not have the HTML inside the PHP code. But if most of the page is generated dynamically, then it ends up being cleaner to make it all PHP code. Most of my scripts have only one set of in them. However, for any real site or application, it makes sense to use templates for any HTML generation, thus allowing the designers to change the look of the page without ever having to touch the code, along with being able to internationalize the output without any extra work besides the actual translation.
You can't really call PHP an interpreted langauge anymore, and I believe the same is true of Perl. PHP 4 uses the Zend engine, which compiles PHP before executing it, the same way that Java must be compiled before it can be executed. I would assume the same is true with most other scripting languages, such as ColdFusion and ASP. The difference is that the compiling is done at run time, so to update some code it is only necessary to save the file. PHP (with add-ons) and ColdFusion can both cache the intermediate compiled code, so that it doesn't have to be compiled everytime. I wouldn't be surprised if Perl (with mod_perl?) and ASP had this sort of feature.
I agree completely that a real comparison would be useful. Coming from someone who's day job is doing web applications, I'll say that programmer time is much more expensive than hardware, so if it takes less time to develop and maintain an application, having to add an extra server, more ram or a faster cpu is going to end up being cheaper in the long run.
Most (all?) cable contracts specifically state that you cannot resell the service in any capacity. Now, if you wanted to give the service away...
My mistake. I didn't realize that this wasn't included in the TOC protocol.
The problem is that AOL can't add anything to their client as a security feature. If a majority of the existing clients don't already have it, then it won't work. Just mimick the last version of the official client that didn't have the security feature.