Not sure if that's the particular instance of rtbl'ing that the grandparent is referring to, but (s)he's almost certainly refering to the slashcod $rtbl flag.
class Foo {
long SomeFunction();
void SomeOtherFunction(String arg);
void SomeOtherFunction(long arg);
}
SomeOtherFunction( SomeFunction() );
How does the the compiler know whether you wanted "SomeOtherFunction(String)" or "SomeOtherFunction(long)" ?
It's implicit in your code, due to the return type of "SomeFunction".
That's all implied types do.
long l = SomeFunction()
The "long" is redundant. The compiler doesn't need it - it already know what type SomeFunction returns, as shown from the previous example.
If you want to add it because you think it makes your code more readable, then that's allowed, but the compiler doesn't need it.
We have servers that ship with internal mirrored SCSI, and the only things we have on the drives are the OS and tmp.
I much prefer my vendor not to try and guess what my storage needs are.
How do I know whether Microsoft Word will meet my needs? I can't test it before I buy it
Individuals usually don't (although Macs ship with a demo version of Office X), but companies certainly do. We (the company I work for) spend quite a bit of time testing any significant piece of software before we buy it.
Not if we get rid of copyright first.
You'd also need to consider contract law. If I hire you to write software under contract and then you leak that software, I can sue you.
If you drop the copyright laws then I'll lose the ability to sue the people taking copies, but I will still be able to sue you for breach of contract.
If I'm a software developer I just need to enter into a written contract with each purchaser before handing them a version of the software. If they share it, I sue them.
Given that the model requires to me find people who are willing to pay for a share anyway, adding in a contract isn't that much effort.
The primary difficulty I see with a true "share system" is timing.
Who goes first?
I think what you are proposing is that I come up with an idea for what I want to write, and then go and find people who will buy a share in that, and then I write it, and then they pay me.
That can work, but it's tricky. Will people commit to something before they see it? How do they know whether my word processor will meet their needs? It doesn't exist yet, so they can't "test" it. If we give them the option of backing out of their share, then I as the author have to carry that risk.
When do they pay me? Upfront? What if I don't deliver? It starts to turn into something quite a lot like bespoke development, which fraught with danger. How many companies have been burnt by projects that never completed (or ran over time/over budget).
It's do-able, but I don't think the market will want it.
Alternatively, I write the software upfront and then I don't release the software until all the shares are sold. Now I'm running even more of a risk. I could spend a year writing it, and then never recover my costs. The current model has that risk as well, but it also has the opportunity for significant returns. This model suggests that there is capped return, but a significant risk, something most investors aren't keen on, so who will lend me the money for my year of development?
And, even if that is covered, there's still the issue that people will be being asked to commit to a share of something they can't have yet. They will be signing up to pay their $100 share, but there's no guarantee whether I will ever sell all my shares (or in what time frame).
Then, once it's written, everyone gets to use it.
I simply can't see the average company (or individual) wanting to be in that system. If you come to me with an offer of "pay $100 to help fund the development of this software, and then aftwards I'll give it away free" then my immediate reaction will be, "I'll just let some other sucker pay the $100 and at the end I'll get it for nothing".
Now if you found 100 people to fund your software development, then it's effectively a work for hire. They paid for it, so they effectively own the results. Why would they want to give it away? If they're a company, and they own a piece of software that reduces their operating costs, why would they want to let their competitors get it for nothing?
Now you might not give them that option - you could say that all your software will be released free at the end. The problem is, that is likely to reduce their willingness to buy shares. So then you have to drop your costs. If someone else comes along and is willing to comply with their wishes, then they'll buy their shares in that.
The guts of it is, that even if everyone switched to your model, then eventually someone would come along and they would offer a model more like what we have now, and the market would switch to it. You mightn't like what we have, you might consider it "outdated", but I can't see the market being willing to give it up.
Of course it is.
But it's very rare for a single organisation(/individual) to be willing to pay for the true cost of that design. Should every person who needs an email client pay to have one developed? Should I pay for someone to design a word processor specifically for me? Should that person pay someone to design a set of development tools? Is that really what you're arguing for?
The alternative is that we share that design cost amongst a number of people. Each person pays a share of that cost, and in return receives rights in accordance with that share.
The current software licensing model is just a formalisation of that approach. The companies that want to use mumblestheclown's training materials, pay their share of the cost of developing those materials. They enter into an agreement, whereby they pay a relatively small proportion of the price of the design/production of the product and make use of the product in proportion to that price.
For them to go and put that product on file sharing networks is both a violation of the agreement that they volutarily entered into, and also gos against the spirit of what they entered into - they are supposed to be sharing costs with other users, but they're ripping off those other users by allowing others to free load, thus pushing up the price of indiviual share in the development costs.
The licensing model we have is a evolution/distortion (your pick) of that fairly simple approach. We can tweak it if we're not happy, but the only other approach we have is forcing one person (or company) to pay the full price, which is hardly likely to fly.
Free software proponents seem to think that software companies are tying to play some dirty trick - charging everyone for something that has already been paid for. They're not. The license cost is just your way of paying for a share in the development. That's not a secret - it's just obvious.
it makes sense to pay him for his labor, not the number of copies he can sell.
Well, next time you want to pay the full development cost for a piece of software, then go ahead.
If you can find someone out there that will pay mumbles labour costs so that he can give away his software, then I'm sure he'll be happy to change his business model, but what you're proposing is that he personally absorbs that cost. Somehow I suspect that isn't going to keep food on the table for him or his employees.
If Larry makes $1 million, and only spends $100,000 then he is taxed 23% on the $100,000 he spends, which is 2.3% of his "income". That may initially seem unfair.
But while it is true that he hasn't paid tax on the $900,000, he also hasn't actually derived any benefit from it. The value of money comes when you exchange it for a good or service. So if Larry puts it under his bed for 10 years, and then spends it later, he will pay tax on it then.
From a "fairness" point of view, there's no real issue. The minute you derive benefit from your income you pay tax. If you don't spend it, then you don't pay tax - but you also don't get benefit from your income.
There is the question of effectiveness and efficiency though. i.e. Is it in the nations interest to delay collecting tax from Larry until he spends the money, even though he has the money now.
On that question, my primary answer would be that the nation is deriving benefit from the money, because Larry isn't going to put it under his bed - he's going to invest it, and other people can use it to generate more goods and services. So in that case, the nation derives full benefit from the untaxed income, even though Larry isn't getting anything (except more income [in interest] that he then either needs to spend [and be taxed] or invest [and benefit the economy])
I'm not necessarily a supporter of "fair tax", but I don't think "fairness" requires taxing based on income. Outgoings are a more accurate measure of a persons "wealth" than income is.
One of the snippets in the article refers to a supposedly alarming comment in dtrace.
The author of the comment has explained(*) it in his blog.
(*) explained in a way that probably only makes sense to hardware/assembly geeks. (i.e. Not me)
The difference there was that people could install Windows. OS/2 and Windows supported (pretty much) the same hardware.
That's unlikely to be the case for Mac/x86, so it would require people to buy a completely new computer and completely ditch MacOS.
That's still a real possibility, but it's slightly different to the OS/2 case.
Grrr, my previous reply was an accidental submit.. Why isn't the default action on the form to preview instead?
It's basically a half way point between the GPL and the LGPL.
For most Open Source developers the easiest thing to do is to just use the software under the GPL.
However, if you use the software as a library, and only make use of the specific APIs that Red Hat has listed, then it effectively becomes like the LGPL. You are not obliged to release your code under the GPL.
But unlike the LGPL, the set of allowed APIs is fixed, and defined by Red Hat. In a LGPL program you can open up new APIs and change existing APIs and as long as you release your changes to the library, you don't need to GPL your program. With the Fedora Directory Server, you can change the APIs all you want, but the new APIs you create can only be used under the GPL - i.e. you have to release your program under the GPL too.
It's covered at on the wiki.
As for evil/not evil, I'd say "not evil". It's seemingly designed to allow existing closed source software that used the NS directory to keep doing the same thing with FDS, but they've done a good job of preventing those users from getting a free run. They get to keep using the same APIs that they've always used, and now they have the freedom to fix bugs in the directory, but they don't get all the freedoms that GPL developers will get.
It's basically a half way point between the GPL and the LGPL.
For most Open Source developers the easiest thing to do is to just use the software under the GPL.
However, if you use the software as a library, and only make use of the specific APIs that Red Hat has listed, then it effectively becomes like the LGPL. You are not obliged to release your code under the GPL.
But unlike the LGPL, the set of allowed APIs is fixed, and defined by Red Hat. In a LGPL program you can open up new APIs and change existing APIs and as long as you release your changes to the library, you don't need to GPL your program. With the NS Directory Server, you can change the APIs all you want, but the new APIs you create can only be used under the GPL - i.e. you have to release your program under the GPL too.
You asked a question and got some answers. Most of those answers weren't very helpful for you. That much is agreed.
You seem to be saying that anyone who couldn't give you a perfect answer should not say anything at all. Every answer you received was someone going out of their way to offer you the best information that they had. If it wasn't good enough, you just say "Thanks, but that won't be suitable because...".
You're trying to make people wish they hadn't helped you at all, and that ruins this resource for everyone. All those people that tried to help you and got rude responses from you are now much less likely to help someone on slashdot again. Next time I want to ask a question on Slashdot, all those people with slightly imperfect answers will be tempted to stay quiet.
You asked a very specialised question to a very general audience and then were rude to people who, with the best of intentions, gave you advice you couldn't use.
If you stand on the street and ask for medical advice, then you need to be able to filter the responses. It doesn't really make sense to then complain that their advice wasn't right. Asking slashdot about batch control systems is a bit like that. Most people here won't be experts. They'll try to help, and a lot of their advice will be pretty useless, but just accept it as the cost of taking free advice.
if open source ever wants to be more successful, it will have to respond to what customers want as well
Open source isn't a company. It doesn't want anything - least of all the kind of success you refer to.
If there's a company out there trying to produce an open source batch control system, then they really do need to listen to what you want. But I very much doubt that anyone here is doing that. I doubt there is a single person here who has a goal of writing a successful open source batch control system.
You seem to have this idea that there's someone out there trying to produce an open source version of every piece of software ever written. And that open source is somehow a failure if that doesn't happen.
that does not think there's a better way to do it then just plain cron
I definately think there's a better way than cron. But I'm not going to spend my free time writing it just so you can avoid spending money on autosys.
Or none of us want to write a scheduler just so that you can avoid paying "mega bucks".
A lot of us do know what schedulers do but here's some advice that you won't have to pay "mega bucks" for.
Your ask slashdot article didn't say what you were going to use it for, so all the suggestions to write scripts are perfectly reasonable. Your comments like "Writing scripts and make files to do crap like this is NOT acceptable" don't help your cause. You didn't tell give us enough information up front to help you. Maybe if you said "we want something like autosys but without the price tag" then we wouldn't have wasted our time trying to help you save a few (thousand) bucks.
Everyone here gave their time freely to try and help you. Your responses to them have been childish and totally uncalled for.
You seem to expect that we should all go out and write an open source scheduler just so you don't have to pay for autosys (or control-m, etc). Even more strangely, you seem to think that we should care that you can't get something freely. Here's a tip - we don't. If you want autosys, go buy it.
Apart from the fact that it might not be open source (I can't find an actual license anywhere), I think Torque/Maui is exactly what you want.
Yes it can manage across a cluster, but I'm pretty sure you can run it with a single (local) node in the cluster and all jobs with run locally.
I haven't tried that, but in theory...
The syntax is rather weird, but it seems to have everything you wanted. Except I can't quite work out whether it does dependency management. It seems like it should, but I haven't worked out how.
As for Quartz, you didn't actually specify what language you wanted, or how much programming you were willing to do.
Java can call native processes. Quartz has builtin (but poorly documented) support for calling out to the OS to run commands.
It has the foundations to do everything you asked for. Maybe java isn't an option for you, but if you dig a bit, it probably will work.
I am aware that you're after a batch scheduluer, but the thing is, once you get past the cron style schedulers, the next level up tend to have remote job management (etc) so don't let the fact that Torque markets itself as a cluster manager put you off too much. (of course if it really doesn't handle dependencies, then that's a good reason to ditch it)
What you're really asking for is something like Control-M from BMC software. But it's not open source, and it's not cheap.
Slashdot Mirror
But it was paid for by the US government.
How? (serious question).
Do you keep a blog somewhere?
It doesn't look like Aspyr's status page tracks ongoing progress.
Not sure if that's the particular instance of rtbl'ing that the grandparent is referring to, but (s)he's almost certainly refering to the slashcod $rtbl flag.
class Foo {
;
long SomeFunction();
void SomeOtherFunction(String arg);
void SomeOtherFunction(long arg);
}
SomeOtherFunction( SomeFunction() )
How does the the compiler know whether you wanted "SomeOtherFunction(String)" or "SomeOtherFunction(long)" ?
It's implicit in your code, due to the return type of "SomeFunction".
That's all implied types do.
long l = SomeFunction()
The "long" is redundant. The compiler doesn't need it - it already know what type SomeFunction returns, as shown from the previous example.
If you want to add it because you think it makes your code more readable, then that's allowed, but the compiler doesn't need it.
One word: SAN
We have servers that ship with internal mirrored SCSI, and the only things we have on the drives are the OS and tmp.
I much prefer my vendor not to try and guess what my storage needs are.
Individuals usually don't (although Macs ship with a demo version of Office X), but companies certainly do. We (the company I work for) spend quite a bit of time testing any significant piece of software before we buy it.
Not if we get rid of copyright first.
You'd also need to consider contract law. If I hire you to write software under contract and then you leak that software, I can sue you.
If you drop the copyright laws then I'll lose the ability to sue the people taking copies, but I will still be able to sue you for breach of contract.
If I'm a software developer I just need to enter into a written contract with each purchaser before handing them a version of the software. If they share it, I sue them.
Given that the model requires to me find people who are willing to pay for a share anyway, adding in a contract isn't that much effort.
Who goes first?
I think what you are proposing is that I come up with an idea for what I want to write, and then go and find people who will buy a share in that, and then I write it, and then they pay me.
That can work, but it's tricky. Will people commit to something before they see it? How do they know whether my word processor will meet their needs? It doesn't exist yet, so they can't "test" it. If we give them the option of backing out of their share, then I as the author have to carry that risk.
When do they pay me? Upfront? What if I don't deliver? It starts to turn into something quite a lot like bespoke development, which fraught with danger. How many companies have been burnt by projects that never completed (or ran over time/over budget).
It's do-able, but I don't think the market will want it.
Alternatively, I write the software upfront and then I don't release the software until all the shares are sold. Now I'm running even more of a risk. I could spend a year writing it, and then never recover my costs. The current model has that risk as well, but it also has the opportunity for significant returns. This model suggests that there is capped return, but a significant risk, something most investors aren't keen on, so who will lend me the money for my year of development?
And, even if that is covered, there's still the issue that people will be being asked to commit to a share of something they can't have yet. They will be signing up to pay their $100 share, but there's no guarantee whether I will ever sell all my shares (or in what time frame).
Then, once it's written, everyone gets to use it.
I simply can't see the average company (or individual) wanting to be in that system. If you come to me with an offer of "pay $100 to help fund the development of this software, and then aftwards I'll give it away free" then my immediate reaction will be, "I'll just let some other sucker pay the $100 and at the end I'll get it for nothing".
Now if you found 100 people to fund your software development, then it's effectively a work for hire. They paid for it, so they effectively own the results. Why would they want to give it away? If they're a company, and they own a piece of software that reduces their operating costs, why would they want to let their competitors get it for nothing?
Now you might not give them that option - you could say that all your software will be released free at the end. The problem is, that is likely to reduce their willingness to buy shares. So then you have to drop your costs. If someone else comes along and is willing to comply with their wishes, then they'll buy their shares in that.
The guts of it is, that even if everyone switched to your model, then eventually someone would come along and they would offer a model more like what we have now, and the market would switch to it. You mightn't like what we have, you might consider it "outdated", but I can't see the market being willing to give it up.
Of course it is.
But it's very rare for a single organisation(/individual) to be willing to pay for the true cost of that design. Should every person who needs an email client pay to have one developed? Should I pay for someone to design a word processor specifically for me? Should that person pay someone to design a set of development tools? Is that really what you're arguing for?
The alternative is that we share that design cost amongst a number of people. Each person pays a share of that cost, and in return receives rights in accordance with that share.
The current software licensing model is just a formalisation of that approach. The companies that want to use mumblestheclown's training materials, pay their share of the cost of developing those materials. They enter into an agreement, whereby they pay a relatively small proportion of the price of the design/production of the product and make use of the product in proportion to that price.
For them to go and put that product on file sharing networks is both a violation of the agreement that they volutarily entered into, and also gos against the spirit of what they entered into - they are supposed to be sharing costs with other users, but they're ripping off those other users by allowing others to free load, thus pushing up the price of indiviual share in the development costs.
The licensing model we have is a evolution/distortion (your pick) of that fairly simple approach. We can tweak it if we're not happy, but the only other approach we have is forcing one person (or company) to pay the full price, which is hardly likely to fly.
Free software proponents seem to think that software companies are tying to play some dirty trick - charging everyone for something that has already been paid for. They're not. The license cost is just your way of paying for a share in the development. That's not a secret - it's just obvious.
it makes sense to pay him for his labor, not the number of copies he can sell.
Well, next time you want to pay the full development cost for a piece of software, then go ahead.
If you can find someone out there that will pay mumbles labour costs so that he can give away his software, then I'm sure he'll be happy to change his business model, but what you're proposing is that he personally absorbs that cost. Somehow I suspect that isn't going to keep food on the table for him or his employees.
Unfortunately his browser embeds yellow dots in his posts, so his employer could stick track it back to him.
No, this is slashdot.
We use an alternative punctuation system here.
Supposedly everyone(*) is switching to NetBeans.
(*) NB: Switchers may not be real people.
If Larry makes $1 million, and only spends $100,000 then he is taxed 23% on the $100,000 he spends, which is 2.3% of his "income". That may initially seem unfair.
But while it is true that he hasn't paid tax on the $900,000, he also hasn't actually derived any benefit from it. The value of money comes when you exchange it for a good or service. So if Larry puts it under his bed for 10 years, and then spends it later, he will pay tax on it then.
From a "fairness" point of view, there's no real issue. The minute you derive benefit from your income you pay tax. If you don't spend it, then you don't pay tax - but you also don't get benefit from your income.
There is the question of effectiveness and efficiency though. i.e. Is it in the nations interest to delay collecting tax from Larry until he spends the money, even though he has the money now.
On that question, my primary answer would be that the nation is deriving benefit from the money, because Larry isn't going to put it under his bed - he's going to invest it, and other people can use it to generate more goods and services. So in that case, the nation derives full benefit from the untaxed income, even though Larry isn't getting anything (except more income [in interest] that he then either needs to spend [and be taxed] or invest [and benefit the economy])
I'm not necessarily a supporter of "fair tax", but I don't think "fairness" requires taxing based on income. Outgoings are a more accurate measure of a persons "wealth" than income is.
One of the snippets in the article refers to a supposedly alarming comment in dtrace.
The author of the comment has explained(*) it in his blog.
(*) explained in a way that probably only makes sense to hardware/assembly geeks. (i.e. Not me)
I like it.
How do you generally pronounce it? "jay-whiz"?
General Motors is obsolete already.
However in my world, I generally judge how good something is, based on what I can actually do with it.
"Hey boss, you know that project we need finished by the end of this year? I've found this great framework. It'll be ready sometime next year..."
I switched because waiting 10 minutes for Netscape Gold to start up seemed like a serious waste of time when I was on timed internet.
See Wikipedia : Quark
(which you may have known - it's so hard to tell what's intentional and what's accidental around here)
Randomly release 3% of defendants?
A logic puzzle for you.
- You are a judge.
- You have 100 cases for DUI to try today.
- In all cases the defendant was found to be over the legal limit by use of a breathalizer.
- You know that 3% of breathalizer results are incorrect.
How do you fulfil your obligations of fairness and justice in these cases?The difference there was that people could install Windows. OS/2 and Windows supported (pretty much) the same hardware.
That's unlikely to be the case for Mac/x86, so it would require people to buy a completely new computer and completely ditch MacOS.
That's still a real possibility, but it's slightly different to the OS/2 case.
It's basically a half way point between the GPL and the LGPL.
For most Open Source developers the easiest thing to do is to just use the software under the GPL.
However, if you use the software as a library, and only make use of the specific APIs that Red Hat has listed, then it effectively becomes like the LGPL. You are not obliged to release your code under the GPL.
But unlike the LGPL, the set of allowed APIs is fixed, and defined by Red Hat. In a LGPL program you can open up new APIs and change existing APIs and as long as you release your changes to the library, you don't need to GPL your program. With the Fedora Directory Server, you can change the APIs all you want, but the new APIs you create can only be used under the GPL - i.e. you have to release your program under the GPL too.
It's covered at on the wiki.
As for evil/not evil, I'd say "not evil". It's seemingly designed to allow existing closed source software that used the NS directory to keep doing the same thing with FDS, but they've done a good job of preventing those users from getting a free run. They get to keep using the same APIs that they've always used, and now they have the freedom to fix bugs in the directory, but they don't get all the freedoms that GPL developers will get.
For most Open Source developers the easiest thing to do is to just use the software under the GPL.
However, if you use the software as a library, and only make use of the specific APIs that Red Hat has listed, then it effectively becomes like the LGPL. You are not obliged to release your code under the GPL.
But unlike the LGPL, the set of allowed APIs is fixed, and defined by Red Hat. In a LGPL program you can open up new APIs and change existing APIs and as long as you release your changes to the library, you don't need to GPL your program. With the NS Directory Server, you can change the APIs all you want, but the new APIs you create can only be used under the GPL - i.e. you have to release your program under the GPL too.
You seem to be saying that anyone who couldn't give you a perfect answer should not say anything at all. Every answer you received was someone going out of their way to offer you the best information that they had. If it wasn't good enough, you just say "Thanks, but that won't be suitable because...".
You're trying to make people wish they hadn't helped you at all, and that ruins this resource for everyone. All those people that tried to help you and got rude responses from you are now much less likely to help someone on slashdot again. Next time I want to ask a question on Slashdot, all those people with slightly imperfect answers will be tempted to stay quiet.
You asked a very specialised question to a very general audience and then were rude to people who, with the best of intentions, gave you advice you couldn't use.
If you stand on the street and ask for medical advice, then you need to be able to filter the responses. It doesn't really make sense to then complain that their advice wasn't right. Asking slashdot about batch control systems is a bit like that. Most people here won't be experts. They'll try to help, and a lot of their advice will be pretty useless, but just accept it as the cost of taking free advice.
if open source ever wants to be more successful, it will have to respond to what customers want as well
Open source isn't a company. It doesn't want anything - least of all the kind of success you refer to.
If there's a company out there trying to produce an open source batch control system, then they really do need to listen to what you want. But I very much doubt that anyone here is doing that. I doubt there is a single person here who has a goal of writing a successful open source batch control system.
You seem to have this idea that there's someone out there trying to produce an open source version of every piece of software ever written. And that open source is somehow a failure if that doesn't happen.
that does not think there's a better way to do it then just plain cron
I definately think there's a better way than cron. But I'm not going to spend my free time writing it just so you can avoid spending money on autosys.
A lot of us do know what schedulers do but here's some advice that you won't have to pay "mega bucks" for.
Yes it can manage across a cluster, but I'm pretty sure you can run it with a single (local) node in the cluster and all jobs with run locally. I haven't tried that, but in theory...
Have a look at the msub command
The syntax is rather weird, but it seems to have everything you wanted. Except I can't quite work out whether it does dependency management. It seems like it should, but I haven't worked out how.
As for Quartz, you didn't actually specify what language you wanted, or how much programming you were willing to do.
Java can call native processes. Quartz has builtin (but poorly documented) support for calling out to the OS to run commands.
It has the foundations to do everything you asked for. Maybe java isn't an option for you, but if you dig a bit, it probably will work.
I am aware that you're after a batch scheduluer, but the thing is, once you get past the cron style schedulers, the next level up tend to have remote job management (etc) so don't let the fact that Torque markets itself as a cluster manager put you off too much. (of course if it really doesn't handle dependencies, then that's a good reason to ditch it)
What you're really asking for is something like Control-M from BMC software. But it's not open source, and it's not cheap.