Does Ada code compiled by GNAT run as fast as comparable C code on 16.8 MHz battery powered devices with 256 KB of RAM?
Speaking as someone who does embedded programming in Ada (using Rational Apex, not GNAT) for a living, I can tell you that the Ada version has a good chance of being faster, if you are writing the C code to be just as safe and not just for speed.
The reason is that the type checking and input validation code is an intrinsic part of the Ada compiler. It is used by every single program it compiles, therefore the compiler writers put a lot of effort into optimizing that code and eliminating it whenever possible. By contrast, whatever additional code you add to a C program for safety purposes is unique to your application and not subject to the same level of scrutiny and optimization.
Of course you can sacrifice safety for speed, but most of the time you don't want to. If you were to look at assembly generated by equivalent C and Ada code, you would have a difficult time telling them apart. You will see safety code generated in Ada where it makes sense, like the first time a user input is converted to an array index, but a good C programmer should include that code manually anyway. It's not like an Ada variable is validated every single time it is used. The compiler knows when it is necessary and when it isn't.
That sounds interesting. I admit to an ignorance of c++ templates and will have to look that up. However, I think unit checking using c++ templates can hardly be called common practice. I once submitted a patch that employed simple textbook polymorphism to an open source c++ project and received oohs and ahhs. In a language like Ada, that kind of type checking is obvious, ubiquitous, and intrinsic to using the language.
I don't know if you or the essay author realize it, but the essay proves the grandparent's point more than yours. The opening statement is, "The more you have to remember to maintain
security the easier it is to forget something." The author then goes on to show a long list of things that must be remembered in a C program in order to maintain security.
Yes, it is possible to write secure code in an unsafe language, but it is a lot more than "a bit" harder. Any talented programmer almost instinctively knows when code has a potential for security problems. Unfortunately, most programming teams consist of a wide range of talent. When a large project with a large team is considered as a whole, the probability of a security problem slipping through becomes much higher when an unsafe language is used.
Have you ever tried to write a buffer overflow in Ada? It's very difficult and requires a strong grasp of the language. Hardly a description of a bad coder. The first time you experience the joy of a compiler error telling you that you tried to add a distance in meters with a distance in feet without a conversion, you'll wonder why you ever chose to program in C.
I'm confused. Are The Gimp and OpenOffice.org merely blatant knockoffs of Photoshop and Microsoft Office? Or are they so different that they will never be accepted by the industry? Somebody had better give me a straight answer soon, or I'm going to stop relying on slashdot to make these kinds of decisions for me.
You have described a textbook example of a paradigm. Another textbook example of a paradigm is Swiss Watches. The Swiss actually invented the quartz watch, but stuck with the industry standard mechanical watch, allowing Seiko to gain a strong foothold in the watch market that was previously monopolized by the Swiss.
Don't get me wrong; I'm not saying that Photoshop isn't a stellar application. I'm merely saying that being an industry standard is not a good enough reason to choose it. Granted I only have a small side business, but I've never had a problem using The Gimp to interact with vendors that only use Photoshop for graphic design work. If you can't think of a task you can accomplish with Photoshop and not with The Gimp, then you are a victim of peer pressure, plain and simple. And I bet you won't be switching to The Gimp no matter what anyone says, even though the interface is now practically identical.
Not that us open source geeks really care what you run or what you say about what we run. Most of us became accustomed before high school to ignoring those who advocated what was popular in favor of what makes sense.
No Quicktime or WMV plugin means a lot of websites like CNN and Yahoo don't really work well.
I think it's important to distinguish between no Quicktime or WMV plugin and no easy to install on Ubuntu Quicktime or WMV plugin. CNN works great for me using mplayer-plugin on my Gentoo systems, thank you very much, and just required a USE flag setting if I remember right. Getting it to work on most other distros I have tried, however, can be quite an experience, but it is still possible. And I believe it works out of the box on commercial, desktop-oriented distros like Linspire.
What's patented is not the actual FS image on the disk, but the method to create, read, or write the filesystem. That code presumably exists on every iPod, whether it is used or not, unless the firmware is flashed too when the disk is reformatted. You can format as many hard drives as you want without paying licensing fees for each individual one, just like you can create as many Word documents as you like without licensing each one. A filesystem is no different than any other document, as far as the hardware is concerned.
Of course, one could argue that it is unfair to require fees to use the best means of a storage device interacting with a monopolistic operating system. Microsoft's refusal to include native support for other file systems also ironically leads to a skewed sense of interoperability among the naive. "I can use my Windows disks on the Macs at school, but I can't use Mac disks on Windows. Windows is therefore the most interoperable system because they obviously make better disks."
Hey, it's what got me interested in programming. It's interactive so it gives instant results. It can teach the basics of algorithmic thinking without burdening the beginner with all the necessary quirks that go with "normal" programming languages. If they find they actually have a talent and interest in programming, a more conventional language can be used for a 102 course.
I've actually thought a lot about this question since my wife took a Java 101 class a few years ago. The following are comments she made while she was doing her homework:
What the heck does "public static void main" mean?
What's all this import stuff about?
What's the point of this try/catch stuff?
Why do I have to recompile the whole thing when I just changed one letter?
These error messages don't make sense and it's hard to tell where they occurred.
I know what all the commands do individually, but I really wanted to learn more about how they work together.
I spend so much time dealing with the compiler and stuff, that I haven't been able to really learn about programming.
What's the point of floats, doubles, and ints? Why don't they just call it a number and leave it at that?
What I got out of her comments was that the programming language was getting in the way of learning programming. While most of her questions seem pretty basic to an experienced programmer, they were still outside the scope of her class. The teacher didn't have time and the students didn't have the background to understand the answers.
In my opinion, the best languages for introduction to programming are loosely-typed scripting languages. Let the students get the basics of programming down first, then concepts like type, storage, classes, packages, error handling, compiling, etc. can be added later.
I feel sorry for your future kids if you think that physical harm is the worst thing that can happen to a child. Broken bones mend in months. I know 50 year-olds who are still suffering from bad choices they made in their youth.
You can't count on an immature person to make a mature decision. Parents need to check up on and discipline their kids. Ironically, it is the best way to preserve a child's freedom to make mistakes. Still young enough to believe that children with less rules are more free? Let me put it to you this way: by enforcing smaller rules parents shield their child from the long term consequences that small actions can lead to, until the child is mature enough to consider the long term consequences for him or herself.
Would you rather have a parent occasionally looking over your shoulder while you use the internet or become a victim of an online predator? (And yes, I won't go into them here, but there are long term consequences to a pornography addiction.)
Would you rather get punished for smoking in high school or become addicted and die from lung cancer when you are 35?
Would you rather get grounded for staying out too late with your boyfriend or become a mother at age 16?
Would you rather get punished for skipping school or wake up in your studio apartment when you are 25 and realize you are at the peak of your earning potential?
I could go on and on, but my point is that by checking up on your kids now, you are giving them more choices when they are mature enough to make them, and therefore more freedom in the long run. By all means teach your children right from wrong and give them room to make mistakes, but let it be small mistakes until they truly understand what they are choosing between. Children don't need to be exposed to all the harshness of the world in order to understand what is and isn't good for them; all they need to know is that they were starting to head down a wrong path and why that was the wrong path. You don't need to get shot to understand that guns are dangerous. The same applies to non-physical harm.
I think sometimes the OSS community does itself a disfavour by trying to upsell the OSS projects out there as comparable to their professional counterparts when they aren't.
Well, often they are comparable for the situation. No one would argue that a Chevy Cavalier is better than a Ferrari, but what if all you need it for is taking the kids to school and buying groceries? Then you are just paying for a status symbol.
I always recommend going with an open source solution until you come across a need that it doesn't provide. Then either implement it yourself (which I have done before), or find a commercial solution that will work better. Once you have already paid for commercial software, there is little sense in evaluating whether an open source package will meet your needs, unless you are facing yearly subscription or upgrade costs, or perhaps want to change operating systems for other reasons. Especially complex software like graphics or NLE, which have a large learning curve.
It is a lot harder to convince a long-time FCP user of the benefits of cinelerra, even if they don't actually use anything that only FCP provides, just because of inertia. Conversely, I would have a difficult time moving to FCP right now, because of the time I have invested in cinelerra.
However, look at a new user like my stepdad, who is not tech savvy, but wants to do something with the hundreds of hours of video he has recorded but never watched. I'm going to recommend something like Kino until he outgrows it, and he probably never will.
You are the first response I have seen that sounds like you actually have done this before.
In addition, I would mention reliability and backup. All cameras and hard drives eventually break. Except in extremely rare circumstances, a camera breaking is not going to ruin anything you have already taped. If the computer capturing the video crashes, you are probably going to lose at least the last few minutes. If the hard drive physically crashes, the whole day is gone.
I don't know anyone who uses a firestore-type device that doesn't also use tape as a backup. They love the time it saves in capturing, but also realize that one day the hard drive is going to crash in the middle of someone's wedding vows or something and they will be glad for the tape.
Does this list meet your standards, then? Or are Pixar, Dreamworks, and ILM just deluding themselves as well?
I won't argue with you about the quality of FCP compared to open source offerings, having never used it myself. From what I've heard I'm sure it is a stellar app. However, the grandparent was pointing out that "all but shafted as far as editing goes" is certainly an unfair assessment of the video editing capabilities on linux, even if high end commercial apps like those in the list above are taken out of the mix.
Kino is polished and easy to use. Although it has a very limited feature set, it is more than sufficient for something as simple as a conference or home video. Cinelerra is a little rough around the edges, but has a nice feature set and is easy to extend with plugins. Most of the difficulty in learning Cinelerra is because it is so full-featured. I can't speak to MainActor, as I have been too happy with the other two to fork out the money for it, but I'm sure it is a capable app. And all 3 of these are continuously getting better.
And do I even need to mention price? For $999, FCP had better be good. With Cinelerra I have made high quality videos with keyframe animation, color correction, nice transitions, excellent audio filtering, 5 channel surround, compositing to remove unwanted elements from a scene, anti-aliased truetype titling, and more. I have even set up a small renderfarm. I'm not saying FCP isn't a good investment, but for me if the end product doesn't look any different with an open source NLE, why should I spend the money? Maybe if it saved a lot of time, but I honestly don't see where any time could be cut from an experienced user of any NLE.
Don't mistake a bad implementation for a bad concept. I'd love to hear from a class that didn't have the bugs yours did, to get a true evaluation of the concept. The article mentions enhancements of integrating questions into a powerpoint presentation to prevent the in-class delays, and a red/green indication on individual devices to make sure the responses were received.
It's easy to agree with the market share theory, but I think you would find it hard to prove that it is easier to support Windows drivers than Linux.
I'll admit to some ignorance about the Windows side of things. All I know is that a certain usb flash storage device wouldn't work until they installed a service pack on my Win2K machine at work, and that I spent most of an afternoon finding and installing new drivers when my in-laws got XP.
Contrast that with my experience using commercial binary drivers on my Linux boxes. I use 4 total: nvidia, win4lin, ndiswrapper for a wireless pcmcia card, and proprietary vpn software for logging in to my work network from home. The only time I really had to download updates was during the 2.4->2.6 switch.
In addition, hardware vendors can take advantage of community support to create Linux device drivers. This tips the balance heavily in Linux's favor. I only know of one example of this working in the Windows world. The manufacturer of my hf ham radio merely released the computer control specs without writing any interface software for any platform. They counted on the ingenuity of hams to fill the need and it worked.
I favour object-oriented (OO) over procedural,... OO code has been verified to be better.
That's probably why Gnome/GTK+ is one of the best examples of object-oriented C code I have seen. Just like procedural code can be written in Java, even though the language requires everything to be part of a class, object-oriented code likewise can be written in most every language. OO is a design methodology, not a programming language feature, although certain languages are unarguably designed to be better suited to OO programming than others.
Personally, I learned the basics of OO programming from the camel book. Larry Wall's witty writing style and the pragmatic way classes are defined and used in Perl helped OO theory to "click" in a way that any C++ text I had read previously failed to do for me.
I'm not trying to say that Perl or C are better than C++ for object-oriented programming. I'm just affirming that OO programming is not limited to C++ and Java.
Interesting parable. I have to admit you may be right. I was basing my assertion on the fact that I have worked in the industry for 10 years and don't remember ever meeting someone who was currently or previously employed at a software company. I do live in the Phoenix area, which has a lot of technical jobs, but a much lower concentration of software companies than areas like Silicon Valley or Seattle.
My definition of a software company may be different than yours as well. I exclude consulting firms who pretty much only sell one copy of their software, writing different software for each contract, or who get paid to heavily customize and install their software for each individual customer. I consider those companies to be selling software engineering services rather than selling software, and that matches with the point of my original post.
I did find this link, which states that 46% of software engineers work in the computer and data processing services industry. 54% is only a small majority, so "vast" is perhaps an overstatement. But it includes the companies that I excluded in my above statement, so there is no way of knowing what the actual figure is from that source.
You are doing a disservice to yourself and the company you work for if you don't think outside of the box you've put yourself in. Scripts are fast to develop, easy to debug, fast to execute, compact, portable, and reusable. Don't confuse a poorly written script with writing a script being a poor choice. Most of the software written on my floor is in C or Ada, but even people who don't know Perl can tell when a Perl script is a better choice.
And implementing a gui interface for a script is not a hack; it is a textbook example of the model-view separation principle. While you're at it, look up coupling and cohesion in a software engineering text, if you want to know why having the same program use and write a configuration file is not necessarily a good idea.
I configure everything on my gentoo boxes via command line due to personal preference. However, I installed, configured, and continue to administer my suse laptop using only their gui tools. Just because a Linux box can be configured via the command line, and most Linux gurus choose to do it that way, doesn't mean it is the only option available.
All other arguments aside, I had to laugh at your choice of a video driver to make your point. If your X server is broken, how is it going to export an API and display a gui so you can fix it? I wish I lived in your universe. I could just drive my car to the repair shop when it breaks down instead of paying for a tow truck.
I use kino for capture (also a great single-track editor), cinelerra for editing and compositing, the gimp for cover art and dvd menus, dvdauthor for creating dvds, and xine for previewing dvds. If I were to pick a commercial solution for Linux, I would have to go with MainActor.
One of these years I would like to make/find an end to end open source solution, specifically optimized for the DV to DVD videographer market. However, in true Unix tradition, each separate tool I currently use does its job very well in my opinion.
Although I have no real basis for comparison to FCP or Premiere as the only Windows video editing software I have used was the bare bones single-track editor that came with my firewire card over 5 years ago. I only made one video before my permanent switch to Linux.
Source? Most of the wedding/corporate videographers I know prefer Vegas on Windows or Final Cut Pro if they are Mac people. They are definitely a cut above the home video market, but not full-time film editors either.
The poor aren't buying computers at any price, and the middle class, drawn to more upscale retailers like Target, hasn't shown the slightest interest in Linux.
I haven't had a windows partition since before XP. The problem is, I don't have a good way to show retailers my interest in Linux and protect my middle-class pocketbook at the same time.
I recently bought 2 laptops. I wasn't in a hurry so I shopped around for ages trying to buy them with Linux or sans OS. I ended up buying Windows laptops from a major retailer and overwriting the hard drives with Linux on the first boot. Why? Because I got better laptops for $300 less each.
I haven't bought a desktop in a long time, but what I have seen in that arena isn't too promising either. One can either buy a $200 bottom-of-the-barrel machine with a Linux distro nobody has heard of that doesn't even include the modem drivers on the disk, or a $2000 machine with RHEL or similar targeted to the server market.
If a major retailer actually makes a decent effort at a consumer-level linux offering, and doesn't give up before the next time I need to replace my machine, I will be the first in line to vote with my dollars. However, if I was to spend $300 extra just to get Linux, I would want the money to go to developers and not be wasted on an OEM that sells too few computers to make volume discounts possible.
First of all, a vast majority of people who actually write software for a living don't work for a software company. Most of those that do, don't get paid per copy. They get paid according to how much time they spend creating the software.
I can't speak for all free software developers, but the main reason I donate time to free software is that I want to contribute to the sustenance of an environment that gives me a lot more back than I put in. I have over 800 packages installed on my system, and have only contributed significantly to one.
If you could let someone stay in your house for a month in exchange for 800 months of rent, would you house him?
If you could cook a meal and get 800 free meals in return, would you cook it?
If you could knit a sock and get 800 articles of clothing in return, would you knit it?
The fact that software is essentially free to duplicate makes contributing one package and receiving 800 possible, what isn't possible with material goods. It even allows plenty of room for freeloaders, as long as there are enough people like me working to sustain the environment.
In summary, I don't expect software to be free, but I know a good investment when I see one.
(Volume of computer with xmms) - (Volume of computer without xmms) = 0 ml
(Mass of computer with xmms) - (Mass of computer without xmms) = 0 g
Songs available with internet connection = essentially infinite
Reference please? The word "combatant" doesn't appear anywhere in the text of the Patriot Act. With apologies to Inigo Montoya, I do not think the Patriot Act says what you think it says. If you insist on disparaging the Patriot Act, at least do it for a valid reason.
I might be inclined to believe your analysis if you got your facts straight. The USA PATRIOT Act has nothing to do with the authority to declare a U.S. citizen an enemy combatant and hold him or her as a prisoner of war. That authority comes from the war powers invoked in Public Law 107-40. Individual cases are subject to oversight both by the Supreme Court and by Congress.
I suppose you also believe that the USA PATRIOT Act allows the FBI to perform a mass library record search without the approval of a judge. Libraries aren't even mentioned in the entire text of the law, and the language people blow out of proportion would only allow the search of a single person's records on approval of a federal judge with respect to a specific ongoing investigation. Please try to verify things you read on Slashdot before passing it on, including what I just said.
Not that I disagree with you that things could quickly get out of hand. I assume because you feel so strongly on the matter that you have contacted your Congressman to support H.R. 1076: Detention of Enemy Combatants Act, which acknowledges the need to detain enemy combatants who are U.S. citizens, but enacts specific requirements on the duration, conditions, and judicial review of such detentions.
Slashdot Mirror
Speaking as someone who does embedded programming in Ada (using Rational Apex, not GNAT) for a living, I can tell you that the Ada version has a good chance of being faster, if you are writing the C code to be just as safe and not just for speed.
The reason is that the type checking and input validation code is an intrinsic part of the Ada compiler. It is used by every single program it compiles, therefore the compiler writers put a lot of effort into optimizing that code and eliminating it whenever possible. By contrast, whatever additional code you add to a C program for safety purposes is unique to your application and not subject to the same level of scrutiny and optimization.
Of course you can sacrifice safety for speed, but most of the time you don't want to. If you were to look at assembly generated by equivalent C and Ada code, you would have a difficult time telling them apart. You will see safety code generated in Ada where it makes sense, like the first time a user input is converted to an array index, but a good C programmer should include that code manually anyway. It's not like an Ada variable is validated every single time it is used. The compiler knows when it is necessary and when it isn't.
That sounds interesting. I admit to an ignorance of c++ templates and will have to look that up. However, I think unit checking using c++ templates can hardly be called common practice. I once submitted a patch that employed simple textbook polymorphism to an open source c++ project and received oohs and ahhs. In a language like Ada, that kind of type checking is obvious, ubiquitous, and intrinsic to using the language.
Yes, it is possible to write secure code in an unsafe language, but it is a lot more than "a bit" harder. Any talented programmer almost instinctively knows when code has a potential for security problems. Unfortunately, most programming teams consist of a wide range of talent. When a large project with a large team is considered as a whole, the probability of a security problem slipping through becomes much higher when an unsafe language is used.
Have you ever tried to write a buffer overflow in Ada? It's very difficult and requires a strong grasp of the language. Hardly a description of a bad coder. The first time you experience the joy of a compiler error telling you that you tried to add a distance in meters with a distance in feet without a conversion, you'll wonder why you ever chose to program in C.
You have described a textbook example of a paradigm. Another textbook example of a paradigm is Swiss Watches. The Swiss actually invented the quartz watch, but stuck with the industry standard mechanical watch, allowing Seiko to gain a strong foothold in the watch market that was previously monopolized by the Swiss.
Don't get me wrong; I'm not saying that Photoshop isn't a stellar application. I'm merely saying that being an industry standard is not a good enough reason to choose it. Granted I only have a small side business, but I've never had a problem using The Gimp to interact with vendors that only use Photoshop for graphic design work. If you can't think of a task you can accomplish with Photoshop and not with The Gimp, then you are a victim of peer pressure, plain and simple. And I bet you won't be switching to The Gimp no matter what anyone says, even though the interface is now practically identical.
Not that us open source geeks really care what you run or what you say about what we run. Most of us became accustomed before high school to ignoring those who advocated what was popular in favor of what makes sense.
Of course, one could argue that it is unfair to require fees to use the best means of a storage device interacting with a monopolistic operating system. Microsoft's refusal to include native support for other file systems also ironically leads to a skewed sense of interoperability among the naive. "I can use my Windows disks on the Macs at school, but I can't use Mac disks on Windows. Windows is therefore the most interoperable system because they obviously make better disks."
Hey, it's what got me interested in programming. It's interactive so it gives instant results. It can teach the basics of algorithmic thinking without burdening the beginner with all the necessary quirks that go with "normal" programming languages. If they find they actually have a talent and interest in programming, a more conventional language can be used for a 102 course.
I've actually thought a lot about this question since my wife took a Java 101 class a few years ago. The following are comments she made while she was doing her homework:
- What the heck does "public static void main" mean?
- What's all this import stuff about?
- What's the point of this try/catch stuff?
- Why do I have to recompile the whole thing when I just changed one letter?
- These error messages don't make sense and it's hard to tell where they occurred.
- I know what all the commands do individually, but I really wanted to learn more about how they work together.
- I spend so much time dealing with the compiler and stuff, that I haven't been able to really learn about programming.
- What's the point of floats, doubles, and ints? Why don't they just call it a number and leave it at that?
What I got out of her comments was that the programming language was getting in the way of learning programming. While most of her questions seem pretty basic to an experienced programmer, they were still outside the scope of her class. The teacher didn't have time and the students didn't have the background to understand the answers.In my opinion, the best languages for introduction to programming are loosely-typed scripting languages. Let the students get the basics of programming down first, then concepts like type, storage, classes, packages, error handling, compiling, etc. can be added later.
You can't count on an immature person to make a mature decision. Parents need to check up on and discipline their kids. Ironically, it is the best way to preserve a child's freedom to make mistakes. Still young enough to believe that children with less rules are more free? Let me put it to you this way: by enforcing smaller rules parents shield their child from the long term consequences that small actions can lead to, until the child is mature enough to consider the long term consequences for him or herself.
Would you rather have a parent occasionally looking over your shoulder while you use the internet or become a victim of an online predator? (And yes, I won't go into them here, but there are long term consequences to a pornography addiction.)
Would you rather get punished for smoking in high school or become addicted and die from lung cancer when you are 35?
Would you rather get grounded for staying out too late with your boyfriend or become a mother at age 16?
Would you rather get punished for skipping school or wake up in your studio apartment when you are 25 and realize you are at the peak of your earning potential?
I could go on and on, but my point is that by checking up on your kids now, you are giving them more choices when they are mature enough to make them, and therefore more freedom in the long run. By all means teach your children right from wrong and give them room to make mistakes, but let it be small mistakes until they truly understand what they are choosing between. Children don't need to be exposed to all the harshness of the world in order to understand what is and isn't good for them; all they need to know is that they were starting to head down a wrong path and why that was the wrong path. You don't need to get shot to understand that guns are dangerous. The same applies to non-physical harm.
I always recommend going with an open source solution until you come across a need that it doesn't provide. Then either implement it yourself (which I have done before), or find a commercial solution that will work better. Once you have already paid for commercial software, there is little sense in evaluating whether an open source package will meet your needs, unless you are facing yearly subscription or upgrade costs, or perhaps want to change operating systems for other reasons. Especially complex software like graphics or NLE, which have a large learning curve.
It is a lot harder to convince a long-time FCP user of the benefits of cinelerra, even if they don't actually use anything that only FCP provides, just because of inertia. Conversely, I would have a difficult time moving to FCP right now, because of the time I have invested in cinelerra.
However, look at a new user like my stepdad, who is not tech savvy, but wants to do something with the hundreds of hours of video he has recorded but never watched. I'm going to recommend something like Kino until he outgrows it, and he probably never will.
In addition, I would mention reliability and backup. All cameras and hard drives eventually break. Except in extremely rare circumstances, a camera breaking is not going to ruin anything you have already taped. If the computer capturing the video crashes, you are probably going to lose at least the last few minutes. If the hard drive physically crashes, the whole day is gone.
I don't know anyone who uses a firestore-type device that doesn't also use tape as a backup. They love the time it saves in capturing, but also realize that one day the hard drive is going to crash in the middle of someone's wedding vows or something and they will be glad for the tape.
I won't argue with you about the quality of FCP compared to open source offerings, having never used it myself. From what I've heard I'm sure it is a stellar app. However, the grandparent was pointing out that "all but shafted as far as editing goes" is certainly an unfair assessment of the video editing capabilities on linux, even if high end commercial apps like those in the list above are taken out of the mix.
Kino is polished and easy to use. Although it has a very limited feature set, it is more than sufficient for something as simple as a conference or home video. Cinelerra is a little rough around the edges, but has a nice feature set and is easy to extend with plugins. Most of the difficulty in learning Cinelerra is because it is so full-featured. I can't speak to MainActor, as I have been too happy with the other two to fork out the money for it, but I'm sure it is a capable app. And all 3 of these are continuously getting better.
And do I even need to mention price? For $999, FCP had better be good. With Cinelerra I have made high quality videos with keyframe animation, color correction, nice transitions, excellent audio filtering, 5 channel surround, compositing to remove unwanted elements from a scene, anti-aliased truetype titling, and more. I have even set up a small renderfarm. I'm not saying FCP isn't a good investment, but for me if the end product doesn't look any different with an open source NLE, why should I spend the money? Maybe if it saved a lot of time, but I honestly don't see where any time could be cut from an experienced user of any NLE.
Don't mistake a bad implementation for a bad concept. I'd love to hear from a class that didn't have the bugs yours did, to get a true evaluation of the concept. The article mentions enhancements of integrating questions into a powerpoint presentation to prevent the in-class delays, and a red/green indication on individual devices to make sure the responses were received.
I'll admit to some ignorance about the Windows side of things. All I know is that a certain usb flash storage device wouldn't work until they installed a service pack on my Win2K machine at work, and that I spent most of an afternoon finding and installing new drivers when my in-laws got XP.
Contrast that with my experience using commercial binary drivers on my Linux boxes. I use 4 total: nvidia, win4lin, ndiswrapper for a wireless pcmcia card, and proprietary vpn software for logging in to my work network from home. The only time I really had to download updates was during the 2.4->2.6 switch.
In addition, hardware vendors can take advantage of community support to create Linux device drivers. This tips the balance heavily in Linux's favor. I only know of one example of this working in the Windows world. The manufacturer of my hf ham radio merely released the computer control specs without writing any interface software for any platform. They counted on the ingenuity of hams to fill the need and it worked.
Personally, I learned the basics of OO programming from the camel book. Larry Wall's witty writing style and the pragmatic way classes are defined and used in Perl helped OO theory to "click" in a way that any C++ text I had read previously failed to do for me.
I'm not trying to say that Perl or C are better than C++ for object-oriented programming. I'm just affirming that OO programming is not limited to C++ and Java.
My definition of a software company may be different than yours as well. I exclude consulting firms who pretty much only sell one copy of their software, writing different software for each contract, or who get paid to heavily customize and install their software for each individual customer. I consider those companies to be selling software engineering services rather than selling software, and that matches with the point of my original post.
I did find this link, which states that 46% of software engineers work in the computer and data processing services industry. 54% is only a small majority, so "vast" is perhaps an overstatement. But it includes the companies that I excluded in my above statement, so there is no way of knowing what the actual figure is from that source.
And implementing a gui interface for a script is not a hack; it is a textbook example of the model-view separation principle. While you're at it, look up coupling and cohesion in a software engineering text, if you want to know why having the same program use and write a configuration file is not necessarily a good idea.
I configure everything on my gentoo boxes via command line due to personal preference. However, I installed, configured, and continue to administer my suse laptop using only their gui tools. Just because a Linux box can be configured via the command line, and most Linux gurus choose to do it that way, doesn't mean it is the only option available.
All other arguments aside, I had to laugh at your choice of a video driver to make your point. If your X server is broken, how is it going to export an API and display a gui so you can fix it? I wish I lived in your universe. I could just drive my car to the repair shop when it breaks down instead of paying for a tow truck.
One of these years I would like to make/find an end to end open source solution, specifically optimized for the DV to DVD videographer market. However, in true Unix tradition, each separate tool I currently use does its job very well in my opinion.
Although I have no real basis for comparison to FCP or Premiere as the only Windows video editing software I have used was the bare bones single-track editor that came with my firewire card over 5 years ago. I only made one video before my permanent switch to Linux.
Source? Most of the wedding/corporate videographers I know prefer Vegas on Windows or Final Cut Pro if they are Mac people. They are definitely a cut above the home video market, but not full-time film editors either.
I recently bought 2 laptops. I wasn't in a hurry so I shopped around for ages trying to buy them with Linux or sans OS. I ended up buying Windows laptops from a major retailer and overwriting the hard drives with Linux on the first boot. Why? Because I got better laptops for $300 less each.
I haven't bought a desktop in a long time, but what I have seen in that arena isn't too promising either. One can either buy a $200 bottom-of-the-barrel machine with a Linux distro nobody has heard of that doesn't even include the modem drivers on the disk, or a $2000 machine with RHEL or similar targeted to the server market.
If a major retailer actually makes a decent effort at a consumer-level linux offering, and doesn't give up before the next time I need to replace my machine, I will be the first in line to vote with my dollars. However, if I was to spend $300 extra just to get Linux, I would want the money to go to developers and not be wasted on an OEM that sells too few computers to make volume discounts possible.
I can't speak for all free software developers, but the main reason I donate time to free software is that I want to contribute to the sustenance of an environment that gives me a lot more back than I put in. I have over 800 packages installed on my system, and have only contributed significantly to one.
If you could let someone stay in your house for a month in exchange for 800 months of rent, would you house him?
If you could cook a meal and get 800 free meals in return, would you cook it?
If you could knit a sock and get 800 articles of clothing in return, would you knit it?
The fact that software is essentially free to duplicate makes contributing one package and receiving 800 possible, what isn't possible with material goods. It even allows plenty of room for freeloaders, as long as there are enough people like me working to sustain the environment.
In summary, I don't expect software to be free, but I know a good investment when I see one.
(Volume of computer with xmms) - (Volume of computer without xmms) = 0 ml
(Mass of computer with xmms) - (Mass of computer without xmms) = 0 g
Songs available with internet connection = essentially infinite
Reference please? The word "combatant" doesn't appear anywhere in the text of the Patriot Act. With apologies to Inigo Montoya, I do not think the Patriot Act says what you think it says. If you insist on disparaging the Patriot Act, at least do it for a valid reason.
I suppose you also believe that the USA PATRIOT Act allows the FBI to perform a mass library record search without the approval of a judge. Libraries aren't even mentioned in the entire text of the law, and the language people blow out of proportion would only allow the search of a single person's records on approval of a federal judge with respect to a specific ongoing investigation. Please try to verify things you read on Slashdot before passing it on, including what I just said.
Not that I disagree with you that things could quickly get out of hand. I assume because you feel so strongly on the matter that you have contacted your Congressman to support H.R. 1076: Detention of Enemy Combatants Act , which acknowledges the need to detain enemy combatants who are U.S. citizens, but enacts specific requirements on the duration, conditions, and judicial review of such detentions.