You make some good minor points (check out the high-voted RFEs and bugs in the bug parade for me), but some quibbles:
Currently, C# loads with a footprint of 2 meg to 4 meg.
I would not see that as a fair comparison, since MS can bundle the API into their core operating system DLLs - and Mono doesn't even have a MS-compatible API yet.
You cannot load the Java VM once and run multiple processes (note "processes", not "threads") from the same Java VM memory footprint
You can on Mac OS X, thanks to Apple. I've heard that they're considering rolling this feature back in to the Sun version.
You still cannot do basic OS operations in Java without writing your own JNI library.
Yes you can - multiple solutions exist. JavaUNIX, for example. There's one that's more cross platform, as well.
and also allow you to dictate the algorithm for creating hashes (ex: use first 8 characters, last 8, or middle 8)
What's wrong with creating your own MyString and MyStringHashMap classes?
There should also be a non thread safe version as well as a thread safe version.
Use HashMap, which has been around since JDK 1.2!
In general, the core collection classes should be implemented in pure optimized (down to the actual chosen assembly code) native code.
Have you done actual profiling to prove this is the cause of the slowness? I'm not convinced that this is a big issue - it could be just a case of programmers using the wrong tool for the job. For example, just because a hashtable is O(1) doesn't mean we can be sloppy about it and use hashtables where arrays would be better suited.
The character encoding (UTF8, SJIS, etc.) conversion libraries do not let you control how conversion failure is handled. It doesn't let you find out where the failure occurred or how the bad byte sequences should be handled. In general, error reporting on conversions and parsing is light weight and will not let you pinpoint the source of the error.
That sounds like a niche market. It's perfectly possible to write your own code for that, and for performance, only call it when you need to generate an error message.
And of course, everybody's favorite. Swing runs like a bloated dog (the Eclipse project proves that it did not have to be this way).
Actually, Swing is better now. On my Linux machine, Blackdown JDK 1.4.1 appears to be more efficient than Qt! Qt apps can cause an mp3 player to break up through excessive X server usage when scrolling, but Java apps no longer do (although they used to in a prior version). But I agree, it still has some way to go.
Swing and SWT had different design goals. SWT sounds like it was like an "AWT++", focusing on native widgets. But Swing is not inherently slow by design, it just has a lot of crud (*cough*Taligent?*cough*) inside the implementation.
In a language which has both a strong type system and a strong contract system, such as (to blow my own trumpet) the one I'm currently writing (no official name for it yet), the compiler should be able to catch a wide variety of logic bugs thanks to contract checking, which indispensably depends on type checking.
The point of antitrust law is supposed to be to prevent one company getting too much power over a market (however, it's been weakened from the original intent in the US). So it makes sense to have special conduct restrictions on monopolies to prvent them extending their monopolies into other markets.
The same argument cannot be adduced towards open source because open source is inherently antithetical towards monopolies. If the source is understandable by other developers (as it must be non-obfuscated to count as open source), then there is always the potential for competition if a market need is unsatisfied.
So what are these detrimental social effects of open source that you're worried about? Microsoft losing profits?
Which is why Haskell and ML allow you to make operators like && non-strict (i.e., short-cutting).
I'm confused. && is shortcutting in Java, C++ and (by implication) C. The alleged problem here is that someone could redefine && to give it unexpected semantics.
it'll be a while before Java or C++ gets features like this.
What are you talking about? C++ already has non-strict && and operator overloading. Java only has the former, which some consider saner.
A more broad, sweeping fix? Make it illegal for anyone to sell anything at a loss by default, and require an application for an exception to the law to permit "loss leaders". An exception process is needed because some loss-leaders like soup kitchens are deemed to have social benefits, and others are deemed not to harm the market, such as mints on hotel pillows. That means you too, Mr. Nonprofit. There's nothing that says companies have to be monopolies to cause "harm to the consumer" and there's nothing that says markets depressed at the hands of non-profits are beneficial to the consumer either.
Are you nuts? That would have the effect of criminalising "unapproved by the government" donations in kind.
That's really quite breathtaking. I could be jailed for giving a piece of software I had written to a friend.
The US govt. protects the rights of individuals. It shouldn't be promoting a social agenda at the expense of individual rights,
So, you disagree with the US constitution where it says the purpose of copyright should be to promote the sciences and the useful arts, then?
Just because something is "a social agenda" (e.g. enriching the public domain, tackling the causes of crime, or preventing the spread of disease like SARS, just for 3 random examples which spring to mind) doesn't make it automatically a "communist" agenda.
They're not going to hand out the spoils for free to start with, but that doesn't matter. With sufficient competition (think international), the price of most material goods will plummet. Intellectual property, real estate and services will become the mainstays of the economy (such as it is).
But since basic material needs will become so cheap, people won't need to earn so much. Many will no longer need to be "wage slaves" to businesses. This will set the stage for tipping the balance of power away from corporations and towards the public good.
Dismissing claims of a perpetual motion machine is science.
No, that's medievalism. Empiricism trumps theory. If someone creates a perpetual motion machine, we'll just have to revise the law(s) of thermodynamics.
Denouncing anyone who questions the role of emissions in of global warming strikes me as utter medievalism.
I think those who are being denounced are not just "anyone" but those who specifically are being paid to say that.
I don't denounce certain communist sects for saying that climate change isn't human-induced, but I do think they're misguided. But when we're talking about someone being paid by the Global Climate Coalition, that's a clear case of bias.
In other words, corporations - through their appointed representatives (parliament) - are saying "yeah, we find the internet you guys designed to be very useful - except for this freedom to email thing, and this freedom to share files thing, and... and... and..."
The two are not at all comparable. Free speech is not absolute. Arguing for freedom to spam is like arguing for the freedom to accost people in the street and shout disgusting pornographic advertising slogans in their faces. With a megaphone. Several times a day.
This is absurd. Any email can be considered unsolicited.
Fortunately, the BBC article writer is a moron. This is not what the law actually says. It only applies to direct marketing emails that are without a prior business relationship, as many other slashdot posters have pointed out.
Try not to be so pedantically literal. I'm sure he meant a reply address as in "an address that one can actually reply to, which truthfully identifies the sender", etc. etc.
If you can't reply to a "reply address" it's not really a proper reply address, is it?
Do you think fax laws allow you to put "Mickey Mouse, Timbuctoo" as your return address?
No, it's overbroad. For "double-opt-in", as you call it (also known as "confirmation-based subscription") the user, or someone pretending to be them, has already consented to the subscription. You are just confirming that it was really them.
This rule would allow "one-time" spamming with adverts "disguised" as subscription requests. (I'm sure spammers would find creative ways to get around even the "one-time" restriction - but that's not the point - even one spam per company is too much if a lot of different people spam, which they do!)
In my opinion, there should only be a defense of "I had reasonable cause to believe the person had consented to subscribe" (which could be disproven easily if the spammer gets thousands of complaints on a single mailout, unless the spammer can prove he/she was deceived by his/her list supplier). Not a blanket defense of "you are always allowed to send spam to ask permission to send more spam" - that's ludicrous!
But perhaps the original poster is misrepresenting it.
It took a death penalty threat to Tel$tra Bigpong (no typo - the service stinks) to get them to move against spammers.
I'm not sure that Australia and New Zealand have much political leverage over China. I don't think it really achieves anything to block them too - just pointless collateral damage.
E.g. is a bulk email from my university about issue/announcement X considered legitimate or spam? What about an email from an ISP to all it's subscribers?
If it doesn't contain "direct marketing" it's not affected by this new rule. If it does, in the latter case there is a prior business relationship, and in the former case universities can put something in their matriculation agreement, that every student signs, to allow it.
By taking such a clear and unequivocal position against the incorporated of patented inventions, Open Source excludes itself from large areas of technology.
Actually, the situation is not so clear cut. Technically, the GPL allows patents on covered software, as long as the patent is licensed for free use to all. (But of course, if that condition is followed, the patent then becomes just a piece of evidence of prior art, rather than effective intellectual property.) The BSD and Apache licenses don't say anything about patents at all. And most graphical open source web browsers support GIF even though it's patent-encumbered. So you appear to be misrepresenting the position of Open Source vs patents.
However, there is, as you point out, a very strong contingent in the open source movement who want to abolish (or at least bring some sense into) software patents in the legal system.
Keep Open Source open by the design choices you make and not by political lobbying which diminishes the rights others enjoy.
Please give me a rational argument for why Amazon.com should have a monopoly right over their so-called "invention", "one-click shopping".
Do you deny that the patent system as regards software in the US is vastly overbroad, needs reform and should not be copied in the EU?
(Speaking of which, technically the EU lobbying is not diminishing anyone's rights. In the EU no-one has an absolute right to software patents or business method patents, even if it is judged "non-obvious", etc.)
Plus don't forget MUDs, Talkers, MUSHes etc who really arn't too concerned about accounts getting broken into and whose admins
simply don't have the time (or skills?) required to implement the ssh protocol.
It's pointless reimplementing the ssh protocol when openssh does it for you! Just connect via ssh and login locally. Even if the MUD only accepts telnet connections and nothing else, that's fine - just telnet to localhost in the ssh session, and firewall off the telnet port from all interfaces except the loopback interface.
More to the point, "recompiling" against a changed implementation of an unchanged interface doesn't involve rewriting any code... just an rpm --rebuild or something.
No, the intention of the GPL is to *lower programmer's wages*. Stallman freely states this.
No, he doesn't. Cite one place where he's ever said this.
Now, granted, that's not ALL workers, but given he's a communist
He describes himself as a liberal, not a communist. Of course to some right-wing nutjobs that's the same thing.
- yes, the ideals behind the GPL are communism, no matter how much many out there would prefer to deny it.
So, the ideal of not buying cars with their hoods welded shut is "communist" too, is it?
Or the ideal of having techies paid for services - technical support, and custom programming (which is what a large part of the programming workforce does anyway - most professional programmers aren't paid to work on Office or OpenOffice.) - is "communism"?
If you're not trolling, then you're just a whiner who doesn't realise that no-one gave you the right to have a job handed to you on a silver platter. If someone makes a free replacement for MS Office which is so good that everyone switches over to it - they have just put all the MS Office programmers out of work. Tough. They can now do something more productive with their lives. It doesn't make moral or financial sense to pay them to produce a product which no-one will buy. And that is not "communist" either - that is just free market logic. Same thing happens every year with all kinds of product every year - this is not at all specific to the GPL.
(And I stand by all these points, even though I am pretty much a communist!)
Re:SCO did not copy the infringing code in SCO Lin
on
SCO DOS'ed
·
· Score: 1
Even though SCO is distributing it as SCO Linux, the code is still the property and responsibility of the contributer.
What kind of crackheaded nonsense is that? If it's IBM's property they have every right to do whatever they want with it. I thought the whole point of this suit was that it's SCO's property.
However, if there is any SCO property in the SCO Linux kernel, they've still released it to the world - including IBM - because the GPL says that anyone can use it in their own GPLed code.
And - get this - the GPL says that if you cannot distribute GPL code because it would conflict with some other contract, then you must not distribute it at all.
Re:They needed three days to figure this out?
on
Spam Meeting Wrap-up
·
· Score: 1
It would cause horrible loop problems - imagine that the sender is also using this patch.
I'm not familiar with SMTP, but if RCPT TO comes before RCPT FROM, there is no such loop. Think about it logically.
Slashdot Mirror
Currently, C# loads with a footprint of 2 meg to 4 meg.
I would not see that as a fair comparison, since MS can bundle the API into their core operating system DLLs - and Mono doesn't even have a MS-compatible API yet.
You cannot load the Java VM once and run multiple processes (note "processes", not "threads") from the same Java VM memory footprint
You can on Mac OS X, thanks to Apple. I've heard that they're considering rolling this feature back in to the Sun version.
You still cannot do basic OS operations in Java without writing your own JNI library.
Yes you can - multiple solutions exist. JavaUNIX, for example. There's one that's more cross platform, as well.
and also allow you to dictate the algorithm for creating hashes (ex: use first 8 characters, last 8, or middle 8)
What's wrong with creating your own MyString and MyStringHashMap classes?
There should also be a non thread safe version as well as a thread safe version.
Use HashMap, which has been around since JDK 1.2!
In general, the core collection classes should be implemented in pure optimized (down to the actual chosen assembly code) native code.
Have you done actual profiling to prove this is the cause of the slowness? I'm not convinced that this is a big issue - it could be just a case of programmers using the wrong tool for the job. For example, just because a hashtable is O(1) doesn't mean we can be sloppy about it and use hashtables where arrays would be better suited.
The character encoding (UTF8, SJIS, etc.) conversion libraries do not let you control how conversion failure is handled. It doesn't let you find out where the failure occurred or how the bad byte sequences should be handled. In general, error reporting on conversions and parsing is light weight and will not let you pinpoint the source of the error.
That sounds like a niche market. It's perfectly possible to write your own code for that, and for performance, only call it when you need to generate an error message.
And of course, everybody's favorite. Swing runs like a bloated dog (the Eclipse project proves that it did not have to be this way).
Actually, Swing is better now. On my Linux machine, Blackdown JDK 1.4.1 appears to be more efficient than Qt! Qt apps can cause an mp3 player to break up through excessive X server usage when scrolling, but Java apps no longer do (although they used to in a prior version). But I agree, it still has some way to go.
Swing and SWT had different design goals. SWT sounds like it was like an "AWT++", focusing on native widgets. But Swing is not inherently slow by design, it just has a lot of crud (*cough*Taligent?*cough*) inside the implementation.
In a language which has both a strong type system and a strong contract system, such as (to blow my own trumpet) the one I'm currently writing (no official name for it yet), the compiler should be able to catch a wide variety of logic bugs thanks to contract checking, which indispensably depends on type checking.
The same argument cannot be adduced towards open source because open source is inherently antithetical towards monopolies. If the source is understandable by other developers (as it must be non-obfuscated to count as open source), then there is always the potential for competition if a market need is unsatisfied.
So what are these detrimental social effects of open source that you're worried about? Microsoft losing profits?
Boo hoo, cry me a river.
I'm confused. && is shortcutting in Java, C++ and (by implication) C. The alleged problem here is that someone could redefine && to give it unexpected semantics.
it'll be a while before Java or C++ gets features like this.
What are you talking about? C++ already has non-strict && and operator overloading. Java only has the former, which some consider saner.
Are you nuts? That would have the effect of criminalising "unapproved by the government" donations in kind.
That's really quite breathtaking. I could be jailed for giving a piece of software I had written to a friend.
So, you disagree with the US constitution where it says the purpose of copyright should be to promote the sciences and the useful arts, then?
Just because something is "a social agenda" (e.g. enriching the public domain, tackling the causes of crime, or preventing the spread of disease like SARS, just for 3 random examples which spring to mind) doesn't make it automatically a "communist" agenda.
But since basic material needs will become so cheap, people won't need to earn so much. Many will no longer need to be "wage slaves" to businesses. This will set the stage for tipping the balance of power away from corporations and towards the public good.
No, that's medievalism. Empiricism trumps theory. If someone creates a perpetual motion machine, we'll just have to revise the law(s) of thermodynamics.
Denouncing anyone who questions the role of emissions in of global warming strikes me as utter medievalism.
I think those who are being denounced are not just "anyone" but those who specifically are being paid to say that.
I don't denounce certain communist sects for saying that climate change isn't human-induced, but I do think they're misguided. But when we're talking about someone being paid by the Global Climate Coalition, that's a clear case of bias.
Hello???
That's like saying "ISPs should stop offering Internet access outright".
Can I have some of what you're smoking?
The two are not at all comparable. Free speech is not absolute. Arguing for freedom to spam is like arguing for the freedom to accost people in the street and shout disgusting pornographic advertising slogans in their faces. With a megaphone. Several times a day.
Fortunately, the BBC article writer is a moron. This is not what the law actually says. It only applies to direct marketing emails that are without a prior business relationship, as many other slashdot posters have pointed out.
If you can't reply to a "reply address" it's not really a proper reply address, is it?
Do you think fax laws allow you to put "Mickey Mouse, Timbuctoo" as your return address?
No, it's overbroad. For "double-opt-in", as you call it (also known as "confirmation-based subscription") the user, or someone pretending to be them, has already consented to the subscription. You are just confirming that it was really them.
This rule would allow "one-time" spamming with adverts "disguised" as subscription requests. (I'm sure spammers would find creative ways to get around even the "one-time" restriction - but that's not the point - even one spam per company is too much if a lot of different people spam, which they do!)
In my opinion, there should only be a defense of "I had reasonable cause to believe the person had consented to subscribe" (which could be disproven easily if the spammer gets thousands of complaints on a single mailout, unless the spammer can prove he/she was deceived by his/her list supplier). Not a blanket defense of "you are always allowed to send spam to ask permission to send more spam" - that's ludicrous!
But perhaps the original poster is misrepresenting it.
I'm not sure that Australia and New Zealand have much political leverage over China. I don't think it really achieves anything to block them too - just pointless collateral damage.
If it doesn't contain "direct marketing" it's not affected by this new rule. If it does, in the latter case there is a prior business relationship, and in the former case universities can put something in their matriculation agreement, that every student signs, to allow it.
Actually, the situation is not so clear cut. Technically, the GPL allows patents on covered software, as long as the patent is licensed for free use to all. (But of course, if that condition is followed, the patent then becomes just a piece of evidence of prior art, rather than effective intellectual property.) The BSD and Apache licenses don't say anything about patents at all. And most graphical open source web browsers support GIF even though it's patent-encumbered. So you appear to be misrepresenting the position of Open Source vs patents.
However, there is, as you point out, a very strong contingent in the open source movement who want to abolish (or at least bring some sense into) software patents in the legal system.
Keep Open Source open by the design choices you make and not by political lobbying which diminishes the rights others enjoy.
Please give me a rational argument for why Amazon.com should have a monopoly right over their so-called "invention", "one-click shopping".
Do you deny that the patent system as regards software in the US is vastly overbroad, needs reform and should not be copied in the EU?
(Speaking of which, technically the EU lobbying is not diminishing anyone's rights. In the EU no-one has an absolute right to software patents or business method patents, even if it is judged "non-obvious", etc.)
It's pointless reimplementing the ssh protocol when openssh does it for you! Just connect via ssh and login locally. Even if the MUD only accepts telnet connections and nothing else, that's fine - just telnet to localhost in the ssh session, and firewall off the telnet port from all interfaces except the loopback interface.
When will people learn that the size/speed of some random Hello World has no bearing on the size/speed of realistic applications? I despair!!
No, the intention of the GPL is to *lower programmer's wages*. Stallman freely states this.
No, he doesn't. Cite one place where he's ever said this.
Now, granted, that's not ALL workers, but given he's a communist
He describes himself as a liberal, not a communist. Of course to some right-wing nutjobs that's the same thing.
- yes, the ideals behind the GPL are communism, no matter how much many out there would prefer to deny it.
So, the ideal of not buying cars with their hoods welded shut is "communist" too, is it?
Or the ideal of having techies paid for services - technical support, and custom programming (which is what a large part of the programming workforce does anyway - most professional programmers aren't paid to work on Office or OpenOffice.) - is "communism"?
If you're not trolling, then you're just a whiner who doesn't realise that no-one gave you the right to have a job handed to you on a silver platter. If someone makes a free replacement for MS Office which is so good that everyone switches over to it - they have just put all the MS Office programmers out of work. Tough. They can now do something more productive with their lives. It doesn't make moral or financial sense to pay them to produce a product which no-one will buy. And that is not "communist" either - that is just free market logic. Same thing happens every year with all kinds of product every year - this is not at all specific to the GPL.
(And I stand by all these points, even though I am pretty much a communist!)
What kind of crackheaded nonsense is that? If it's IBM's property they have every right to do whatever they want with it. I thought the whole point of this suit was that it's SCO's property.
However, if there is any SCO property in the SCO Linux kernel, they've still released it to the world - including IBM - because the GPL says that anyone can use it in their own GPLed code.
And - get this - the GPL says that if you cannot distribute GPL code because it would conflict with some other contract, then you must not distribute it at all.
I'm not familiar with SMTP, but if RCPT TO comes before RCPT FROM, there is no such loop. Think about it logically.