Well, the modern stuff is clean, but the original version of DivX was a repackaged binary of Microsoft's MS MPEG-4 codecs from NetShow, hacked to enable them to work in.AVI file format.
The current stuff is clean, of course. But stealing someone else's code and claiming it as your own work can be hard to live down.
Personally, my biggest problem with DivX today is that they're still using the lousy AVI file format. I wish they'd just use.mp4 or.mov instead, like a real codec.
Actually, ancient software runs just fine under OS X via Classic. I caught my dad playing a circa 1989 build of Shanghai a few months ago on his G4 Cube, running 10.2.
The only odd thing it did was to switch the computer to 8-bit mode while it was running, and not restore it afterwards.
Yes, LAME is much more usable than MPEG4IP. I meant it was similar in that it wasn't meant directly for end users (you have to compile it yourself, or get it from somebody else). I use LAME all the time for my own MP3 encoding.
MPEG4IP requires chaining the output of multiple command line apps together. It is scriptable, I suppose, but you get huge uncompressed intermediate files and other painful things. Anyone with the time and tech savvy to make it work would be better off getting a job and then buying a commercial product;).
They have a ATRAC3 derivative for the high data rates for RealAudio. But the "cook" audio codec which most use is developed in house, as is RealVideo.
The ATRAC3 stuff is mainly for mobile audio devices, an area where Real doesn't seem to be playing anymore. I've used cook for everything I've done in the last few years.
Well, Real certainly views their codecs as core IP, and don't want them out in the wild.
That said, the track record of companies monitizing proprietary codecs is NOT good. They could probably GPL RealVideo 9 tomorrow and it really wouldn't hurt them much, and might help in a lot of ways.
Try hopping on the Helix Community IRC sometime. Lots of the developers hang out there, and I've found it's a very responsive way to ask questions.
As for the networking problem, I suspect what you're doing is quite a bit simpler than the full scope of a client/server streaming media architecture. While the theory is reasonably well documented, this isn't the kind of thing a guy writes in a basement in a few months.
Nothing difficult about making a good GUI? I give you exhibit A: MPlayer. Dozens of different UIs available, all terrible.
Nothing difficult about network code for streaming media? Huh? You've got to deal with client/server communication over a lossy connection using UDP, doing retransmission of packets, buffering, doing scalable switching between streams. Doing this well is at least of the same order of difficulty as a good codec.
While I love MPEG-4 for what it's good at, what it's good at doesn't include real-time streaming over the public internet. Darwin + MPEG-4 doesn't offer any good form of scalability. Thus, if you encode a file at 400 Kbps, and a user's connection is 350 Kbps for a few minutes, they'll get a horrible quality experience.
RealMedia supports SureStream, which lets you put up to eight pairs of video and audio into a single file, and the server and player communicate in real time to determine the optimum data rate for the transmission. It'll even raise and lower data rate as connection speed changes - very useful for cable modem and shared bandwidth from work.
This will come in MPEG-4 eventually, via Fine Grain Scalability (FGS), or some future scalable version of the AVC codec. But that's a couple years away from being in real consumer products I'd guess.
Oh, and I totally don't believe that you really regularly use MPEG4IP for volume compression. I mean, the TOOLS are there, but you have to go through like five different command line steps to make a file. It can produce fine results (it uses Xvid), but MPEG4IP is really like LAME - it's not meant as an end-use tool in and of itself. Well, the player is fine stand alone.
I like a cup of tea or two during the day to get and keep me going, but as you say, it doesn't substitute for sleep. I have to be careful to have my last cup by 4pm or so if I'm going to go to sleep by 11pm, or else I'm up late not getting anything done, and am that much more tired, and hence more in need of tea, the next day.
Back when I was a single guy with my parents paying for everything, it was about all I could do to turn in enough work to keep going reasonably well in school. Friends, Harpoon 1, Tetris, doing layout on the school paper, email, obsessing on girls, and straight up writers block often left me rather paralyzed. For stuff I was excited about, I could do good work. Otherwise, it was mostly late and lousy.
Fortunately, I went to a school (Hampshire College) which was very project based, with written evaluations and no tests. So for those four years, I really had to shape up. No coasting through on multiple choice skills, I had to WRITE. And the written evaluations were profoundly more effective than a grade. I remember Eqbal Amahd's "Waggoner shows insight, occasional brilliance, and a certain sloth" better than I ever would have a "C+"
Skip forward a decade or so, and now I'm a father of a 16 month old and a 3.5 year old, and work out of the home doing freelance writing and consulting. In many ways, this should be a nightmare of distractions, but I'm actually more productive than ever before. In the last two years, I've done a lot of consulting for companies like Microsoft, Adobe, Sorenson Media, Getty Images, The Associated Press, The Criterion Collection.
Some tips:
Don't be paralyzed during the first draft. If you're writing, it doesn't have to be perfect the first draft. Just WRITE for a while, and let it take the shape it's going to have. It's much better to write a first draft that you have to throw out entirely than to sit there without writing anything.
Don't sign on to projects you're not excited about. One reason I do well on the consulting and services side is that I only sign on for projects I'm really enthused about. I get pretty enthusiastic, as anyone who has seen any of my presentations knows. But stuff that bores me, bores me, and won't get my best attention. I never offer to do boring stuff for money (one of the better parts of being freelance).
Quit when you're too tired to work well. This was a hard-earned lesson when my daughter was born before my book was done. I spent hours trying to get stuff done, but too tired to work well, when I should have been asleep so I could work well the next day. All nighters rapidly become self defeating.
Cash the advance check. If it's important you get something done, don't leave a way to back out of it. Make completing it your only option. The specter of hideous failure is a great stimulant.
Do what floats your boat at the moment. Turn having multiple commitments into a strength instead of a weakness. When you get bored on your current task or project, switch to something else you aren't stuck on. It's all about finding SOMETHING you can make forward progress on at any given time.
Use, don't abuse deadlines. I pride myself on hitting all my deadlines. But I rarely have stuff done much before the deadlines. The key is having a good idea how long things take, appropriately padded for worst cases. Always leave enough time to do the job well.
Turn off the email. I get a lot of my best work done on the first class cabin on Alaska Airlines 737's. Why? No email, no wife, no kids. Just me, the headphones, and a PowerBook. The email is probably the biggest part. It's really hard for me to not check it when I get that beep, and it definitely throws off my thinking. So I quit Entourage when I'm working well. This goes for Slashdot too.
Need to do it. In school, it's hard to escape the fact that what you do doesn't really matter all that much. It's not like the teacher's job depends on you figuring out some new insights. So, find a question you feel you need to answer. This is easier in the real world. Mortgages and bills have a profound way of focusing your attention!
Well, the long term goal with developing nations is free trade. But as you suggest, the carrot of free trade is quite useful to get countries to modify their behavior in positive ways. Look at how Eastern European nations are reforming, politically and economically, in order to try to get into the EU.
Since environmental protection isn't directly valued economically in most cases, explicit environmental support included in trade pacts is a good thing. The key is to make sure that its use doesn't turn into a veiled protectionism.
Anyway, the floodgates of trade are wide open with China today, so what you're afraid of has largely ever happened.
Since you want to travel the country, here's a striking place in the middle:
http://www.compustory.com/
It's a cute little place, with pretty much every personal computer you've ever heard of on display (except for NeXT...). It'll only take you an hour or two, but it's entertaining, and no lines.
I don't know about Australia (although I'm hoping to be teaching some video compression classes there this "summer"), but Americans have a huge tendency to start little exhibits like this all over the place. Definitely part of the experience.
Our trade deficit doesn't have that much to do with jobs. It's because US currency and stocks are considered investment grade. Lots of the profits made by companies we outsource to get reinvested in the US.
Obviously capitalism doesn't solve all problems. The environment is a classic example. Clearly, we need strong and enforced global environmental laws. And nations absolutely need safety nets, so the unemployed are kept going until they can reenter the workplace.
But reducing trade doesn't help either of those. In fact, increased "low friction" trade is critical to providing improved safety nets and environment, both by increasing the wealth each requires, and by making it possible do do manufacture in the parts of the world best suited to them. Energy-poor nations shouldn't be in the Aluminum business, for example.
As for upward wage pressures, they certainly do exist for Indian IT workers. That's a very small portion of the population there - the pool of Indian engineers is almost certainly smaller than the pool in the USA. And in fact, skilled wages do trend up in integrated economies. It's only economies that have big other problems that show sustained drops in wages. India's recent growth is largely due to their opening up trade barriers, and they've got a lot of room to go on that account!
What you say can be true for a given industry. But the current aggregate differential between the US and India corresponds to the aggregate differential between the two economies. We're richer than India because we produce more, and we're richer per capita because we produce more per capita. It's our productivity in the first place which makes it possible for us to outsource effectively. Wealth doesn't exist apart from productivity.
We're rich because we produce a lot, and export a lot. If we weren't rich, the wage differential wouldn't exist. A better way of thinking about this is "we're so rich now, US workers can't afford to do menial IT tasks."
Protectionism of any form doesn't help make a nation richer. It helps make some sectors richer, at the expense of all the rest. IBM exporting these jobs will help the economy on the whole more than keeping them here.
Of course, the cost of transport has to be factored in. If you make heavy things that don't take much labor, that's likely to happen close to the area of consumption. Of course, freight costs are always going down...
Check out the CIA World Factbook on India some time. It's a largely rural society with huge illiteracy rates. The proportion of Indians with sufficient secondary education to even begin a US caliber computer science degree is tiny, with the gross number certainly smaller than in the US. An Indian programmer makes many, many times the average income in India. It's a much bigger wage gap than in the US between minimum wage and a typical programmer salary.
Of course, as outsourcing progresses, this should improve over time. This is a good thing - India is the world's largest democracy. We want them to progress rapidly, and be a partner to the US in the way only deep economic integration makes possible.
Oh, man, the "zero sum" argument was shot down CENTURIES ago! If you were right, we'd all be 20% as rich as our great-grandparents, since population is up by 5x. Read your Adam Smith.
Seriously, in the whole history of the world, give one example of "racing to the bottom" ever happening? People always talk about this like it's inevitable, but it has NEVER HAPPENED EVEN ONCE. And pretty much by definition, it's impossible to happen. Individual industries can get hurt, but an integrated economy has never on the whole failed due to competition. Failure is more typically caused by bad monetary policy.
What the heck unemployment crisis are you talking about? It's hard in some sectors right now, but unemployment right now isn't bad at all compared to the last 30 years.
Granted, I think that the current lack of economic savvy in the administration isn't helping.
Anyway, I went LIFO on this, and will leave others to critique other elements of your rather amazing errors-per-word post.
Except that the outsourced products are cheaper to make, and hence will be more affordable.
Over time, Indian salaries will rise and US salaries will drop to the point where the relative productivity per dollar will come close to matching. Assuming we can turn out good programmers, the US wages will remain higher.
You're missing the relatively important concept of price elasticity.
If a "good enough" shoe costs $200, people will buy it, since you've got to have shoes. But if somewhere else can make as good a shoe for half the cost, customers will have the choice of buying the "good enough" shoe for $100, or spending the $200 and getting a better shoe.
Either way, the California shoe is going to lose, unless they can figure out how to make a better shoe. This is a good thing, since customers can buy a cheaper shoe, and who knows, maybe California can innovate a better shoe and gain that business back.
Imagine if we HAD to buy only American made televisions and cars over the last 30 years? Would the nation be a better place? Nope, we'd just spend a lot on crappy TVs and cars. American cars are a heck of a lot better now than they would have been without the competitive pressure. So even if you buy an American car today, it's good mainly because others will willing to buy Japanese.
Slashdot Mirror
I suppose this would be akin to someone launching a new open source development tool called "SCO."
/. lead to SCO these days...
Just to prove that all threads on
Well, the modern stuff is clean, but the original version of DivX was a repackaged binary of Microsoft's MS MPEG-4 codecs from NetShow, hacked to enable them to work in .AVI file format.
.mp4 or .mov instead, like a real codec.
The current stuff is clean, of course. But stealing someone else's code and claiming it as your own work can be hard to live down.
Personally, my biggest problem with DivX today is that they're still using the lousy AVI file format. I wish they'd just use
Actually, ancient software runs just fine under OS X via Classic. I caught my dad playing a circa 1989 build of Shanghai a few months ago on his G4 Cube, running 10.2.
The only odd thing it did was to switch the computer to 8-bit mode while it was running, and not restore it afterwards.
You raise a good point. Is there actually much community participation in Darwin development?
I haven't heard of many features or bugs in Darwin being fixed outside of Apple.
So, anyone have any good stories for how the open source parts of Darwin are being used?
Hey Roger!
;).
Yes, LAME is much more usable than MPEG4IP. I meant it was similar in that it wasn't meant directly for end users (you have to compile it yourself, or get it from somebody else). I use LAME all the time for my own MP3 encoding.
MPEG4IP requires chaining the output of multiple command line apps together. It is scriptable, I suppose, but you get huge uncompressed intermediate files and other painful things. Anyone with the time and tech savvy to make it work would be better off getting a job and then buying a commercial product
They have a ATRAC3 derivative for the high data rates for RealAudio. But the "cook" audio codec which most use is developed in house, as is RealVideo.
The ATRAC3 stuff is mainly for mobile audio devices, an area where Real doesn't seem to be playing anymore. I've used cook for everything I've done in the last few years.
Well, Real certainly views their codecs as core IP, and don't want them out in the wild.
That said, the track record of companies monitizing proprietary codecs is NOT good. They could probably GPL RealVideo 9 tomorrow and it really wouldn't hurt them much, and might help in a lot of ways.
Try hopping on the Helix Community IRC sometime. Lots of the developers hang out there, and I've found it's a very responsive way to ask questions.
As for the networking problem, I suspect what you're doing is quite a bit simpler than the full scope of a client/server streaming media architecture. While the theory is reasonably well documented, this isn't the kind of thing a guy writes in a basement in a few months.
The press release listed these codecs for this player:
SMIL 2.0
RealVideo (RV9, RV8, RV7, RVG2)
RealAudio (RA8, G2 audio)
MP3
Ogg Vorbis
MPEG4 (patent license for MPEG4 must be obtained separately)
H.263
Those who like following links before complaining may have found this:
t ro .html
https://player.helixcommunity.org/2003/draft/in
Wow! Support for all the latest video and audio codecs! Including the proprietary-but-excellent Real stuff, Ogg Vorbis, MPEG-4, and others.
Nothing difficult about making a good GUI? I give you exhibit A: MPlayer. Dozens of different UIs available, all terrible.
Nothing difficult about network code for streaming media? Huh? You've got to deal with client/server communication over a lossy connection using UDP, doing retransmission of packets, buffering, doing scalable switching between streams. Doing this well is at least of the same order of difficulty as a good codec.
While I love MPEG-4 for what it's good at, what it's good at doesn't include real-time streaming over the public internet. Darwin + MPEG-4 doesn't offer any good form of scalability. Thus, if you encode a file at 400 Kbps, and a user's connection is 350 Kbps for a few minutes, they'll get a horrible quality experience.
RealMedia supports SureStream, which lets you put up to eight pairs of video and audio into a single file, and the server and player communicate in real time to determine the optimum data rate for the transmission. It'll even raise and lower data rate as connection speed changes - very useful for cable modem and shared bandwidth from work.
This will come in MPEG-4 eventually, via Fine Grain Scalability (FGS), or some future scalable version of the AVC codec. But that's a couple years away from being in real consumer products I'd guess.
Oh, and I totally don't believe that you really regularly use MPEG4IP for volume compression. I mean, the TOOLS are there, but you have to go through like five different command line steps to make a file. It can produce fine results (it uses Xvid), but MPEG4IP is really like LAME - it's not meant as an end-use tool in and of itself. Well, the player is fine stand alone.
I like a cup of tea or two during the day to get and keep me going, but as you say, it doesn't substitute for sleep. I have to be careful to have my last cup by 4pm or so if I'm going to go to sleep by 11pm, or else I'm up late not getting anything done, and am that much more tired, and hence more in need of tea, the next day.
Amen, brother.
Back when I was a single guy with my parents paying for everything, it was about all I could do to turn in enough work to keep going reasonably well in school. Friends, Harpoon 1, Tetris, doing layout on the school paper, email, obsessing on girls, and straight up writers block often left me rather paralyzed. For stuff I was excited about, I could do good work. Otherwise, it was mostly late and lousy.
Fortunately, I went to a school (Hampshire College) which was very project based, with written evaluations and no tests. So for those four years, I really had to shape up. No coasting through on multiple choice skills, I had to WRITE. And the written evaluations were profoundly more effective than a grade. I remember Eqbal Amahd's "Waggoner shows insight, occasional brilliance, and a certain sloth" better than I ever would have a "C+"
Skip forward a decade or so, and now I'm a father of a 16 month old and a 3.5 year old, and work out of the home doing freelance writing and consulting. In many ways, this should be a nightmare of distractions, but I'm actually more productive than ever before. In the last two years, I've done a lot of consulting for companies like Microsoft, Adobe, Sorenson Media, Getty Images, The Associated Press, The Criterion Collection.
Some tips:
Don't be paralyzed during the first draft. If you're writing, it doesn't have to be perfect the first draft. Just WRITE for a while, and let it take the shape it's going to have. It's much better to write a first draft that you have to throw out entirely than to sit there without writing anything.
Don't sign on to projects you're not excited about. One reason I do well on the consulting and services side is that I only sign on for projects I'm really enthused about. I get pretty enthusiastic, as anyone who has seen any of my presentations knows. But stuff that bores me, bores me, and won't get my best attention. I never offer to do boring stuff for money (one of the better parts of being freelance).
Quit when you're too tired to work well. This was a hard-earned lesson when my daughter was born before my book was done. I spent hours trying to get stuff done, but too tired to work well, when I should have been asleep so I could work well the next day. All nighters rapidly become self defeating.
Cash the advance check. If it's important you get something done, don't leave a way to back out of it. Make completing it your only option. The specter of hideous failure is a great stimulant.
Do what floats your boat at the moment. Turn having multiple commitments into a strength instead of a weakness. When you get bored on your current task or project, switch to something else you aren't stuck on. It's all about finding SOMETHING you can make forward progress on at any given time.
Use, don't abuse deadlines. I pride myself on hitting all my deadlines. But I rarely have stuff done much before the deadlines. The key is having a good idea how long things take, appropriately padded for worst cases. Always leave enough time to do the job well.
Turn off the email. I get a lot of my best work done on the first class cabin on Alaska Airlines 737's. Why? No email, no wife, no kids. Just me, the headphones, and a PowerBook. The email is probably the biggest part. It's really hard for me to not check it when I get that beep, and it definitely throws off my thinking. So I quit Entourage when I'm working well. This goes for Slashdot too.
Need to do it. In school, it's hard to escape the fact that what you do doesn't really matter all that much. It's not like the teacher's job depends on you figuring out some new insights. So, find a question you feel you need to answer. This is easier in the real world. Mortgages and bills have a profound way of focusing your attention!
Well, the long term goal with developing nations is free trade. But as you suggest, the carrot of free trade is quite useful to get countries to modify their behavior in positive ways. Look at how Eastern European nations are reforming, politically and economically, in order to try to get into the EU.
Since environmental protection isn't directly valued economically in most cases, explicit environmental support included in trade pacts is a good thing. The key is to make sure that its use doesn't turn into a veiled protectionism.
Anyway, the floodgates of trade are wide open with China today, so what you're afraid of has largely ever happened.
Since you want to travel the country, here's a striking place in the middle:
http://www.compustory.com/
It's a cute little place, with pretty much every personal computer you've ever heard of on display (except for NeXT...). It'll only take you an hour or two, but it's entertaining, and no lines.
I don't know about Australia (although I'm hoping to be teaching some video compression classes there this "summer"), but Americans have a huge tendency to start little exhibits like this all over the place. Definitely part of the experience.
And that's just the main store. The Technical store, which would have seemed huge if you hadn't seen the main store, is a couple of blocks away.
One of the best moments of my life was seeing a book I wrote on the shelves there...
Our trade deficit doesn't have that much to do with jobs. It's because US currency and stocks are considered investment grade. Lots of the profits made by companies we outsource to get reinvested in the US.
Obviously capitalism doesn't solve all problems. The environment is a classic example. Clearly, we need strong and enforced global environmental laws. And nations absolutely need safety nets, so the unemployed are kept going until they can reenter the workplace.
But reducing trade doesn't help either of those. In fact, increased "low friction" trade is critical to providing improved safety nets and environment, both by increasing the wealth each requires, and by making it possible do do manufacture in the parts of the world best suited to them. Energy-poor nations shouldn't be in the Aluminum business, for example.
As for upward wage pressures, they certainly do exist for Indian IT workers. That's a very small portion of the population there - the pool of Indian engineers is almost certainly smaller than the pool in the USA. And in fact, skilled wages do trend up in integrated economies. It's only economies that have big other problems that show sustained drops in wages. India's recent growth is largely due to their opening up trade barriers, and they've got a lot of room to go on that account!
What you say can be true for a given industry. But the current aggregate differential between the US and India corresponds to the aggregate differential between the two economies. We're richer than India because we produce more, and we're richer per capita because we produce more per capita. It's our productivity in the first place which makes it possible for us to outsource effectively. Wealth doesn't exist apart from productivity.
We're rich because we produce a lot, and export a lot. If we weren't rich, the wage differential wouldn't exist. A better way of thinking about this is "we're so rich now, US workers can't afford to do menial IT tasks."
Protectionism of any form doesn't help make a nation richer. It helps make some sectors richer, at the expense of all the rest. IBM exporting these jobs will help the economy on the whole more than keeping them here.
Of course, the cost of transport has to be factored in. If you make heavy things that don't take much labor, that's likely to happen close to the area of consumption. Of course, freight costs are always going down...
Check out the CIA World Factbook on India some time. It's a largely rural society with huge illiteracy rates. The proportion of Indians with sufficient secondary education to even begin a US caliber computer science degree is tiny, with the gross number certainly smaller than in the US. An Indian programmer makes many, many times the average income in India. It's a much bigger wage gap than in the US between minimum wage and a typical programmer salary.
Of course, as outsourcing progresses, this should improve over time. This is a good thing - India is the world's largest democracy. We want them to progress rapidly, and be a partner to the US in the way only deep economic integration makes possible.
Oh, man, the "zero sum" argument was shot down CENTURIES ago! If you were right, we'd all be 20% as rich as our great-grandparents, since population is up by 5x. Read your Adam Smith.
Seriously, in the whole history of the world, give one example of "racing to the bottom" ever happening? People always talk about this like it's inevitable, but it has NEVER HAPPENED EVEN ONCE. And pretty much by definition, it's impossible to happen. Individual industries can get hurt, but an integrated economy has never on the whole failed due to competition. Failure is more typically caused by bad monetary policy.
What the heck unemployment crisis are you talking about? It's hard in some sectors right now, but unemployment right now isn't bad at all compared to the last 30 years.
Granted, I think that the current lack of economic savvy in the administration isn't helping.
Anyway, I went LIFO on this, and will leave others to critique other elements of your rather amazing errors-per-word post.
Except that the outsourced products are cheaper to make, and hence will be more affordable.
Over time, Indian salaries will rise and US salaries will drop to the point where the relative productivity per dollar will come close to matching. Assuming we can turn out good programmers, the US wages will remain higher.
You're missing the relatively important concept of price elasticity.
If a "good enough" shoe costs $200, people will buy it, since you've got to have shoes. But if somewhere else can make as good a shoe for half the cost, customers will have the choice of buying the "good enough" shoe for $100, or spending the $200 and getting a better shoe.
Either way, the California shoe is going to lose, unless they can figure out how to make a better shoe. This is a good thing, since customers can buy a cheaper shoe, and who knows, maybe California can innovate a better shoe and gain that business back.
Imagine if we HAD to buy only American made televisions and cars over the last 30 years? Would the nation be a better place? Nope, we'd just spend a lot on crappy TVs and cars. American cars are a heck of a lot better now than they would have been without the competitive pressure. So even if you buy an American car today, it's good mainly because others will willing to buy Japanese.