The price of Porn on PPV is significantly higher than regular movies. According to some friends in the business, the entire PPV business model is based around porn. There's no way they can make a profit based on the random "Let the kids watch some movie they've probably already seen". They just have to offer non-adult programming to make it acceptable to the community.
Since many people are embarrased to buy Porn, even via mail order, they certainly won't go rent it at the local video store. So they use PPV.
By expiring it, they guarantee a revenue stream, compared to letting the viewer record a few dozen shows and repeatedly viewing them.
You can pretty much adjust your withholding to be "right" if you're W-2 or 1099. There are all the right boxes on both sets of forms to tweak it to be right.
And working on minimizing your payment in April isn't the same as minimizing your yearly tax bill (which is really your total taxes paid for the whole year), but that's another semantic nitpick. And I'll assume that you're trying to do both...
When you go from a tax bracket of 10% to 20% at 15,000 (or whatever it is these days), only the money over 15,000 is taxed at the higher rate. It doesn't apply retroactively to the first 15,000.
So lowering your income to the next lower bracket doesn't really save you anything special.
First, he calls these services ineffective because they only fill a few percent of the jobs nationwide. But that percentage jumped from 2.5% 2 years ago to 5% last year (total, for all job boards.) Given that rate, it'll be at 40% in a few years.
Second, he doesn't discuss what fraction of jobs are even posted on these boards. If only 20% of the jobs are posted on these boards, and 5% of them are filled from resumes on them, that's a pretty good percentage.
Last, recruiters use these boards as well, and they probably aren't included in the 5%. The hiring company wouldn't know where the recruiter found your resume.
Overall, I got quite a few hits from the job boards. Some of them were direct and some through recruiters. Not a bad route, especially for high-tech jobs, in my experience.
For what it's worth, I got one contract through the job boards, and then a full-time job through a referral last time I was looking.
The job before that I got through searching the web (altavista at the time) for my keywords. I found several possible companies that had ads on their site but not on the national job boards via google last time I was hunting as well...
They have a patent on use of a process, and the use of that process without a license is a violation of the patent.
Just because someone sold you a device that allows you to do the process does not give you a license for the process.
Just like Dell isn't responsible if you write or use software that violates someone else's patent on a computer they built. They aren't even responsible if someone has a patent on using RAID for a specific application, and they sell you a RAID machine.
Unfortionately, this rapidly degenerates to making any development nearly impossible as you have to do a patent search for every thing your company does. Or else you just do it, and worry about the licensing later.
And software patents are even worse, because there are hundreds of things you could be violating, some of which might even be created by the interaction of different groups that aren't even aware of the others' actions.
Witness the.gif / LZW / Unisys fiasco. They had a patent on using a compression technique, and everyone who supported that technique had to pay a license fee. And the end-user was the one responsible for verifying that their gif encoder was licensed correctly, not the producer of the program.
GEORGIA INSTITUTE OF TECHNOLOGY, CALLED MEETING OF THE GENERAL FACULTY, Followed immediately by a regular meeting of GENERAL FACULTY ASSEMBLY and of the ACADEMIC SENATE
Tuesday, February 5, 2002
http://www.facultysenate.gatech.edu/gfaasmins5fe b. html
At Tech, in 1985, the Freshman SAT was 1245, the HS GPA 3.5 and the Freshman GPA after a year at Tech was 2.5 (with the average GPA for all undergraduates at Tech being 2.6). By 2001, the Freshman SAT was 1331, the HSGPA 3.7 and the freshman GPA after a year at Tech had increased to 2.9 (with the average GPA for all undergraduates at Tech being 3.0).
[...] The percentage of students graduating with honors has increase sharply in the 1990s. [...]
Traditionally a grade of A was seen as a performance better than 90% of the rest of the class. We are now seeing a move towards an A representing mastery of the material.
Sorry, but it's happening at ma tech as well.
Keith G, Gatech '90, ICS. (Member of the freshman class of '86. And the SATs have been re-normed to be higher since then as well.)
And if they don't have pricing power, then the increased taxes on the corporation come out of the individuals who own the corporation's pockets, or else out of the worker's pockets.
Actually, there's plenty of evidence that certain environments increase productivity. You can check Rapid Development for a list of references to the original sources.
There's also plenty of evidence that interruptions decrease productivity. Steps to reduce interruptions, such as putting phones on Voice Mail, increase productivity.
Because embedded systems frequently aren't written as independent applications.
Most embedded products (your microwave, cell phone, or even your palm) don't have virtual memory, and frequently do not have any mechanism for dynamically loading applications. Every task is started by the functional equivilant of "main()", and runs until the device shuts down (or power is removed).
Everything is statically linked when the image is burned into ROM, and the code space never changes. Frequently, the tasks all run in the same memory address space, much like a multi-threaded application.
In this kind of system, GPL code is not usable unless the company is willing to GPL ALL of the code. Which just isn't economically feasible for lots of application domains. In particuliar, many products use third-party, licensed software that isn't available in GPL versions. Therefore, these can't be linked into a GPL real-time kernel.
If you go look at eCos (by Cygnus now Cygnus), you'll see a non-GPL license that addresses these issues by a company that is deeply involved in real-time development. The GPL is a complete failure in deeply embedded environments, and RMS and the FSF doesn't care. (Just like they don't care if you abstract GPL code into another application, communicating via a pipe or shared memory.)
You can get a nice 2500-3000 square foot house on a third of an acre (about 15 years old) for $250k or so out in the burbs. $500k will get you something really nice downtime, or something really huge out in the burbs.
If you're willing to end up 30-40 miles from downtown, you can even buy lots of land and/or a really big newer house for that.
And Atlanta doesn't even have particuliarly cheap real estate.
There's a really good comment up above about how the capabilities of a tool can influence the character of a development process. But as to why you need a tool at all...
In my development paradigm, you need a new branch every time you want to have a different criteria for checking in files. For example, an old code release that only gets security bug fixes would be one branch. A set of code sent to QA might be another branch (and it wouldn't change until QA was ready to test the next candidate.) A developers work area might be a third branch. And the latest and greated development baseline (which is always guaranteed to compile and run regression tests) would be another.
Inevitably, you end up having to merge changes between branches of different policies. Sooner or later you have to push the baseline into QA. And after QA, you have to push a single security fix from the latest and greatest code into that really old release. Managing all of this without tools would be absurd.
As far as how software process interacts with all of this, it defines the policy for each branch. Your process will define if you have a development branch for each feature, or if you just have a main baseline everyone changes. Your process will define when a code change is allowed to be checked into the main codeline. (i.e. when it compiles or after it passes regression?) You process will define how shared artifacts (libraries, etc.) get updated and released into actual products.
And then you have to pick a tool that can handle your software process requirements now and in the future.
P.S. One of the biggest problems with SourceSafe is that it isn't very safe. File corruption is fairly widely reported and it is not immediately detectable. That's just utterly unacceptable for something as critical as the version control system containing hundreds of thousands of dollars worth of effort.
Actually, Microsoft probably uses Perforce, at least on the Windows2000/XP project. (What they actually use is "SourceDepot", but there are no links anywhere on the web to a product by that name, except in some old Perforce Documentation.) See http://www.usenix.org/events/usenix-win2000/invite dtalks/lucovsky.ppt or the google cache for "SourceDepot" for more details.
Although I expect it's difficult to say what they use one thing, since it's a large company and I'm sure parts use Visual Source Safe, and parts might well use ClearCase or CVS.
What it sounds like you're missing is referred to as a 3-way merge.
When merging 2 different code branches, you need the parent from which the two branches started, and the final version in both branches. Then you take the changes between the parent and one of the new versions, and push it into the final version of the other codeline.
The tool we used at my last job for version control was perforce (www.perforce.com). A company called Araxis makes a great 3-way merge tool for Windows that integrates with it nicely. Perforce prevents you from checking in a change until you resolve the merge with the latest version checked into the system. And it doesn't allow any of the files to be checked in until all of the files are merged. (Atomic Changelists.)
If someone has to re-apply a patch multiple times, something is seriously wrong. The other developers are not correctly applying their patch to the latest version of the code, and are instead overwriting it. This is malpractice on the part of the other developers.
The short answer is that it serializes development.
Let's say that you want to add a new paramter to an existing function. You have to touch 25 different source code files that call this function.
You now have to wait until all 25 files are available to be checked out before you can make your change. If someone checks out a file, and then goes on vacation for two weeks, you're stuck.
With a parallel development system, and decent merging software, you just change the files, and when the guy gets back from vacation, he has to merge his changes into the current baseline.
With even medium numbers of developers (N > 25), parallel development is essential to efficiency. Without parallel development, you end up with a potentially large fraction of your developers waiting on something to get finished so they can make their change.
XP says that you should pick a subset of the requirements that are known at the present time, and design exactly as much as needed to implement those requirements. Then implement that design.
Then you pick some additional requirements, and modify the design and the code to meet the combined set of requirements.
Repeat until management says to stop (or you run out of requirements.)
If you have no requirements, XP says you're finished.
And you update your design as requirements are added and changed, which frequently doesn't happen in a design up front approach. No one wants to go back up the waterfall, and then start over with the code. (And without requirements and design expressed as executable tests, you do have to start over with the code every time you make major changes in the design.)
Without relying on fair use provisions, you can't make a copy of the software on your hard disk without a license. (You're violating the copyright by copying from the CD.)
This is, of course, absurd. But that's the way the law reads.
This is part of why the GPL works. You can't copy the software to your harddisk without a license, so you must obviously agree to the GPL to even run it.
Since GPL'ed software is given away, does the first sale doctrine apply to GPL software as well? Specifically, if I download 200 copies (for free), can I give away or sell 200 copies without being bound by the GPL under first sale?
BTW, this is the best advice in any of the threads..
We hired a lawyer to help us close a deal buying a retail shop, and he kept us out of a great deal of trouble.
As part of the Corp-To-Corp setup, you're going to need to negotiate a contract for the terms. You need a lawyer to make sure you're not signing something you really shouldn't.
They can also advise you on the appropriate insurance you'll need for legal liability reasons.
In any case, you want to talk to an accountant. Most of them will do an initial consultation for free, and point you in the right direction. (They're, of couse, expecting your business filing taxes and doing anything else that is required in the future.) That will give you some good ideas of how practical the options are in your case (and if they think the IRS will come take away your Corporate status, and make you an employee.)
In my experience (over 10 years of developing Real Time code on medium to large projects), it's almost never a win to write a home-brew RTOS (or try to do without an RTOS completely).
Except for very simple applications which can run as a single threaded application with interrupt handlers to perform processing, an RTOS is almost always essential. In fact, an RTOS with a correctly partitioned and architected application can be significantly more efficient than a single threaded application that has to periodically poll hardware or status variables. You don't want your protocol stack running in your interrupt handler, and polling to see if there's more data for the protocol stack is very inefficient compared to a context switch in an RTOS when data is actually available.
VxWorks may be overkill or may not be customizable to do what you want, but there are lighter solutions such as Nucleus, eCos (which is open source), RTEMS (BSD license), etc. These solutions are available for relatively small amounts of money. And this cost is a real win compared to the cost to develop and debug thousands of lines of "homebrew OS". And there's much more to the price of building an RTOS than just the developer time. Slipping a product by 3 or 6 months can cause the project to fail completely.
Even the ever-popular "write a homebrew abstraction layer" over a commercial RTOS is almost always a waste. It just introduces overhead, prevents the application from making full use of the underlying RTOS, and almost never makes porting to a new underlying RTOS much simpler.
In the meantime, it really sounds like your application has some serious architectural issues. If you're spending 15 seconds in your ISR, none of your other tasks (or lower priority interrupts) get to run. That's hardly "real time".
The more typical solution is to grab the data from the hardware and send it to a (high priority?) task to process. I find it very difficult to believe that running a single ISR for 15 seconds is optimum. On most systems I've worked on, that would trigger a hardware watchdog time-out.
Basically, it doesn't sound like you've got an competant, experienced embedded developer working on your project. You might consider finding someone to provide consulting services, or hiring someone with more background in the area.
eCos is a RTOS published by Cygnus (now Redhat) that targets the very systems you are describing. It is licensed under the Red Hat eCos Public License Version 1.1.
I don't know if it's fully "free" (especially since the end-user can't reasonably upgrade many of these products), but it's reasonable for the kind of situation you are running into.
See http://www.redhat.com/embedded/technologies/ecos/e coslicense.html
You could offer it under such a license only if the environment does not support dynamic loading of libraries, etc.
Slashdot Mirror
Actually, I'll bet Porn is driving this.
The price of Porn on PPV is significantly higher than regular movies. According to some friends in the business, the entire PPV business model is based around porn. There's no way they can make a profit based on the random "Let the kids watch some movie they've probably already seen". They just have to offer non-adult programming to make it acceptable to the community.
Since many people are embarrased to buy Porn, even via mail order, they certainly won't go rent it at the local video store. So they use PPV.
By expiring it, they guarantee a revenue stream, compared to letting the viewer record a few dozen shows and repeatedly viewing them.
No mystery here. Move along.
I know this is a dead thread, but...
You can pretty much adjust your withholding to be "right" if you're W-2 or 1099. There are all the right boxes on both sets of forms to tweak it to be right.
And working on minimizing your payment in April isn't the same as minimizing your yearly tax bill (which is really your total taxes paid for the whole year), but that's another semantic nitpick. And I'll assume that you're trying to do both...
Tax brackets don't work that way.
When you go from a tax bracket of 10% to 20% at 15,000 (or whatever it is these days), only the money over 15,000 is taxed at the higher rate. It doesn't apply retroactively to the first 15,000.
So lowering your income to the next lower bracket doesn't really save you anything special.
First, he calls these services ineffective because they only fill a few percent of the jobs nationwide. But that percentage jumped from 2.5% 2 years ago to 5% last year (total, for all job boards.) Given that rate, it'll be at 40% in a few years.
Second, he doesn't discuss what fraction of jobs are even posted on these boards. If only 20% of the jobs are posted on these boards, and 5% of them are filled from resumes on them, that's a pretty good percentage.
Last, recruiters use these boards as well, and they probably aren't included in the 5%. The hiring company wouldn't know where the recruiter found your resume.
Overall, I got quite a few hits from the job boards. Some of them were direct and some through recruiters. Not a bad route, especially for high-tech jobs, in my experience.
For what it's worth, I got one contract through the job boards, and then a full-time job through a referral last time I was looking.
The job before that I got through searching the web (altavista at the time) for my keywords. I found several possible companies that had ads on their site but not on the national job boards via google last time I was hunting as well...
SKG
SCO did not drop a nuke. SCO dropped a large box marked "10 Megaton Bomb!!!".
You get to open the box if you sign a non-disclosure.
-SKG
Because that's how patents work.
.gif / LZW / Unisys fiasco. They had a patent on using a compression technique, and everyone who supported that technique had to pay a license fee. And the end-user was the one responsible for verifying that their gif encoder was licensed correctly, not the producer of the program.
They have a patent on use of a process, and the use of that process without a license is a violation of the patent.
Just because someone sold you a device that allows you to do the process does not give you a license for the process.
Just like Dell isn't responsible if you write or use software that violates someone else's patent on a computer they built. They aren't even responsible if someone has a patent on using RAID for a specific application, and they sell you a RAID machine.
Unfortionately, this rapidly degenerates to making any development nearly impossible as you have to do a patent search for every thing your company does. Or else you just do it, and worry about the licensing later.
And software patents are even worse, because there are hundreds of things you could be violating, some of which might even be created by the interaction of different groups that aren't even aware of the others' actions.
Witness the
SKG
To quote from:
e b. html
GEORGIA INSTITUTE OF TECHNOLOGY, CALLED MEETING OF THE GENERAL FACULTY, Followed immediately by a regular meeting of GENERAL FACULTY ASSEMBLY and of the ACADEMIC SENATE
Tuesday, February 5, 2002
http://www.facultysenate.gatech.edu/gfaasmins5f
At Tech, in 1985, the Freshman SAT was 1245, the HS GPA 3.5 and the Freshman GPA after a year at Tech was 2.5 (with the average GPA for all undergraduates at Tech being 2.6). By 2001, the Freshman SAT was 1331, the HSGPA 3.7 and the freshman GPA after a year at Tech had increased to 2.9 (with the average GPA for all undergraduates at Tech being 3.0).
[...] The percentage of students graduating with honors has increase sharply in the 1990s. [...]
Traditionally a grade of A was seen as a performance better than 90% of the rest of the class. We are now seeing a move towards an A representing mastery of the material.
Sorry, but it's happening at ma tech as well.
Keith G, Gatech '90, ICS. (Member of the freshman class of '86. And the SATs have been re-normed to be higher since then as well.)
This is exactly the information used by drug cartels to assassinate informants, as described in a previous Slashdot article.
If the information is being kept, unauthorized access will occur.
SKG
And if they don't have pricing power, then the increased taxes on the corporation come out of the individuals who own the corporation's pockets, or else out of the worker's pockets.
Both of those, in the end, are individuals.
Actually, there's plenty of evidence that certain environments increase productivity. You can check Rapid Development for a list of references to the original sources.
There's also plenty of evidence that interruptions decrease productivity. Steps to reduce interruptions, such as putting phones on Voice Mail, increase productivity.
How about life with a minimum of 28 years?
Corporations get 28 years.
Because embedded systems frequently aren't written as independent applications.
Most embedded products (your microwave, cell phone, or even your palm) don't have virtual memory, and frequently do not have any mechanism for dynamically loading applications. Every task is started by the functional equivilant of "main()", and runs until the device shuts down (or power is removed).
Everything is statically linked when the image is burned into ROM, and the code space never changes. Frequently, the tasks all run in the same memory address space, much like a multi-threaded application.
In this kind of system, GPL code is not usable unless the company is willing to GPL ALL of the code. Which just isn't economically feasible for lots of application domains. In particuliar, many products use third-party, licensed software that isn't available in GPL versions. Therefore, these can't be linked into a GPL real-time kernel.
If you go look at eCos (by Cygnus now Cygnus), you'll see a non-GPL license that addresses these issues by a company that is deeply involved in real-time development. The GPL is a complete failure in deeply embedded environments, and RMS and the FSF doesn't care. (Just like they don't care if you abstract GPL code into another application, communicating via a pipe or shared memory.)
But in Atlanta, $500k WILL buy you a mansion.
You can get a nice 2500-3000 square foot house on a third of an acre (about 15 years old) for $250k or so out in the burbs. $500k will get you something really nice downtime, or something really huge out in the burbs.
If you're willing to end up 30-40 miles from downtown, you can even buy lots of land and/or a really big newer house for that.
And Atlanta doesn't even have particuliarly cheap real estate.
--SKG
There's a really good comment up above about how the capabilities of a tool can influence the character of a development process. But as to why you need a tool at all...
In my development paradigm, you need a new branch every time you want to have a different criteria for checking in files. For example, an old code release that only gets security bug fixes would be one branch. A set of code sent to QA might be another branch (and it wouldn't change until QA was ready to test the next candidate.) A developers work area might be a third branch. And the latest and greated development baseline (which is always guaranteed to compile and run regression tests) would be another.
Inevitably, you end up having to merge changes between branches of different policies. Sooner or later you have to push the baseline into QA. And after QA, you have to push a single security fix from the latest and greatest code into that really old release. Managing all of this without tools would be absurd.
As far as how software process interacts with all of this, it defines the policy for each branch. Your process will define if you have a development branch for each feature, or if you just have a main baseline everyone changes. Your process will define when a code change is allowed to be checked into the main codeline. (i.e. when it compiles or after it passes regression?) You process will define how shared artifacts (libraries, etc.) get updated and released into actual products.
And then you have to pick a tool that can handle your software process requirements now and in the future.
P.S. One of the biggest problems with SourceSafe is that it isn't very safe. File corruption is fairly widely reported and it is not immediately detectable. That's just utterly unacceptable for something as critical as the version control system containing hundreds of thousands of dollars worth of effort.
Actually, Microsoft probably uses Perforce, at least on the Windows2000/XP project. (What they actually use is "SourceDepot", but there are no links anywhere on the web to a product by that name, except in some old Perforce Documentation.) See http://www.usenix.org/events/usenix-win2000/invite dtalks/lucovsky.ppt or the google cache for "SourceDepot" for more details.
Although I expect it's difficult to say what they use one thing, since it's a large company and I'm sure parts use Visual Source Safe, and parts might well use ClearCase or CVS.
What it sounds like you're missing is referred to as a 3-way merge.
When merging 2 different code branches, you need the parent from which the two branches started, and the final version in both branches. Then you take the changes between the parent and one of the new versions, and push it into the final version of the other codeline.
The tool we used at my last job for version control was perforce (www.perforce.com). A company called Araxis makes a great 3-way merge tool for Windows that integrates with it nicely. Perforce prevents you from checking in a change until you resolve the merge with the latest version checked into the system. And it doesn't allow any of the files to be checked in until all of the files are merged. (Atomic Changelists.)
If someone has to re-apply a patch multiple times, something is seriously wrong. The other developers are not correctly applying their patch to the latest version of the code, and are instead overwriting it. This is malpractice on the part of the other developers.
The short answer is that it serializes development.
Let's say that you want to add a new paramter to an existing function. You have to touch 25 different source code files that call this function.
You now have to wait until all 25 files are available to be checked out before you can make your change. If someone checks out a file, and then goes on vacation for two weeks, you're stuck.
With a parallel development system, and decent merging software, you just change the files, and when the guy gets back from vacation, he has to merge his changes into the current baseline.
With even medium numbers of developers (N > 25), parallel development is essential to efficiency. Without parallel development, you end up with a potentially large fraction of your developers waiting on something to get finished so they can make their change.
XP says that you should pick a subset of the requirements that are known at the present time, and design exactly as much as needed to implement those requirements. Then implement that design.
Then you pick some additional requirements, and modify the design and the code to meet the combined set of requirements.
Repeat until management says to stop (or you run out of requirements.)
If you have no requirements, XP says you're finished.
And you update your design as requirements are added and changed, which frequently doesn't happen in a design up front approach. No one wants to go back up the waterfall, and then start over with the code. (And without requirements and design expressed as executable tests, you do have to start over with the code every time you make major changes in the design.)
You missed the more important part.
Without relying on fair use provisions, you can't make a copy of the software on your hard disk without a license. (You're violating the copyright by copying from the CD.)
This is, of course, absurd. But that's the way the law reads.
This is part of why the GPL works. You can't copy the software to your harddisk without a license, so you must obviously agree to the GPL to even run it.
Since GPL'ed software is given away, does the first sale doctrine apply to GPL software as well? Specifically, if I download 200 copies (for free), can I give away or sell 200 copies without being bound by the GPL under first sale?
Most of the major cellular providers in the US are going to be GSM witin 2 years.
AT&T has started their rollout, and Cingular has announced it, IIRC.
BTW, this is the best advice in any of the threads..
We hired a lawyer to help us close a deal buying a retail shop, and he kept us out of a great deal of trouble.
As part of the Corp-To-Corp setup, you're going to need to negotiate a contract for the terms. You need a lawyer to make sure you're not signing something you really shouldn't.
They can also advise you on the appropriate insurance you'll need for legal liability reasons.
In any case, you want to talk to an accountant. Most of them will do an initial consultation for free, and point you in the right direction. (They're, of couse, expecting your business filing taxes and doing anything else that is required in the future.) That will give you some good ideas of how practical the options are in your case (and if they think the IRS will come take away your Corporate status, and make you an employee.)
You can also use QuickBook's online filing to handle much of the tax related paperwork..
If you go this route, it's probably worth their cut to let them handle the withholding and reporting.
In my experience (over 10 years of developing Real Time code on medium to large projects), it's almost never a win to write a home-brew RTOS (or try to do without an RTOS completely).
Except for very simple applications which can run as a single threaded application with interrupt handlers to perform processing, an RTOS is almost always essential. In fact, an RTOS with a correctly partitioned and architected application can be significantly more efficient than a single threaded application that has to periodically poll hardware or status variables. You don't want your protocol stack running in your interrupt handler, and polling to see if there's more data for the protocol stack is very inefficient compared to a context switch in an RTOS when data is actually available.
VxWorks may be overkill or may not be customizable to do what you want, but there are lighter solutions such as Nucleus, eCos (which is open source), RTEMS (BSD license), etc. These solutions are available for relatively small amounts of money. And this cost is a real win compared to the cost to develop and debug thousands of lines of "homebrew OS". And there's much more to the price of building an RTOS than just the developer time. Slipping a product by 3 or 6 months can cause the project to fail completely.
Even the ever-popular "write a homebrew abstraction layer" over a commercial RTOS is almost always a waste. It just introduces overhead, prevents the application from making full use of the underlying RTOS, and almost never makes porting to a new underlying RTOS much simpler.
In the meantime, it really sounds like your application has some serious architectural issues. If you're spending 15 seconds in your ISR, none of your other tasks (or lower priority interrupts) get to run. That's hardly "real time".
The more typical solution is to grab the data from the hardware and send it to a (high priority?) task to process. I find it very difficult to believe that running a single ISR for 15 seconds is optimum. On most systems I've worked on, that would trigger a hardware watchdog time-out.
Basically, it doesn't sound like you've got an competant, experienced embedded developer working on your project. You might consider finding someone to provide consulting services, or hiring someone with more background in the area.
Or you could buy one of the commercial RTOS's that provide source, such as ThreadX or Nucleus.
Those are small and very clean, in my experience.
eCos is a RTOS published by Cygnus (now Redhat) that targets the very systems you are describing. It is licensed under the Red Hat eCos Public License Version 1.1.
e coslicense.html
I don't know if it's fully "free" (especially since the end-user can't reasonably upgrade many of these products), but it's reasonable for the kind of situation you are running into.
See http://www.redhat.com/embedded/technologies/ecos/
You could offer it under such a license only if the environment does not support dynamic loading of libraries, etc.
SKG