He's not worried about how to extend it. He's worried about how to maintain those extensions when rails itself changes - which is a particularly acute problem with dynamic languages like Ruby and the very extension methods you are advocating.
Since there are no statically defined interfaces or type checks you are going to have hundreds of silent / subtle bugs appear as you upgrade rails if you are not careful. (Careful probably means documenting the changes to the extreme and implementing a huge wad of unit tests to exercise every possible code path).
> This is really not a surprise. IE is an inferior product. It always has been. The market share it has received is solely attributable to the bundling with the Microsoft operating systems.
Maybe you only been using computers for 5 years?
Because from about 2000 - 2004 Netscape was the hugest piece of crap imaginable and the Mozilla alternative was horrifically bloated and unusable. IE exceeded them in quality, features, performance, every single metric you can think of. It was an excellent product. As a web developer in 2000 I remember breathing a sigh of relief when clients said not to bother coding for for Netscape because IE was so much easier and more pleasant to work with (we still made things work cross browser, but not having that requirement was a huge blessing).
I don't understand why people insist on rewriting this history.
Strange as it may sound, it may actually be the Chief Security Officer who wants IE most.
Why? Because it's only IE that supports the extensive group policy admin type controls that let him sit at his console and control every aspect of what all the browsers on the legion of locked down machines under his / her supervision can do.
They bought it at the peak of it's hype. I think it was a mistake. MySQL is popular because it is the default database in a standard LAMP stack which is the cheapest hosting you can get in 99% of web hosting shops. If you were on a tight budget you were pretty much forced to use MySQL.
However I believe we are in the sunset of that time now - the world of cheap hosting is moving from shared servers to VPS servers and that means you can choose your own image from thousands that include any database or language you want. LAMP is no longer much cheaper than using any language and database stack you want. And with next-generation frameworks like Rails, Django and Grails PHP is no longer the easiest way to make your website quickly.
Why did Sun buy MySQL anyway? I think it was 2 reasons: 1 to try and get mindshare in the software services market. They need to execute a transformation like what IBM has done but they need to do it more quickly. They therefore bought a pure services company as an entry point. I think it will fail because MySQL is at the zenith of it's popularity right now. The other reason they bought it I think was a hedge against Oracle. The fact is, main reason people buy Sun boxes these days *is to run Oracle*. Therefore Sun is existentially dependent on another company. That other company bought key components of MySQL themselves and made an aggressive push to market themselves on Linux instead of Solaris. I think Sun wants a hedge / weapon against this - effectively, if Oracle continues to push people off Sun boxes by selling them OSS operating systems, Sun will attack Oracle's market share by selling their own boxes as a full solution including OSS databases instead of Oracle.
I heard coverage on radio on the morning of the protest.
It was depressing - they featured people from both sides of the argument but gave much more time to the "pro" filtering side wherein the points were made that something like 70% of Australians support mandatory filtering when polled, and a long opinion from a "technical specialist" who stated that all technical problems had been solved and it was possible to filter with zero impact on performance. [ of course, it turns out that said "expert" runs a company selling filtering software... duh ].
So... this left me fairly depressed about the whole thing.
> Rudd's interest in this is that both the filtering and the national broadband scheme were election promises, and while I admire his integrity in trying to carry through with all of his election promises (unlike the previous mob, who turned election lying into a high art), I really wish he would dump the promises that were clearly stupid.
I want to pick on one incorrect point you made because it is very important history is not rewritten here. In the election campaign the Rudd government promised an *optional* filter. That is, one people could opt out of by notifying their ISP.
In August this year it was leaked that the filter had become mandatory.
Please do not allow Conroy and Rudd to get away with claiming they are just "fulfilling an election promise". They are not - they are breaking an election promise. (it is worth noting that the in same week the filter became mandatory Fielding magically turned around and passed reams of government legislation in the senate - this gives me very little hope that Conroy is going to change - if a back room deal has been done then it's been done).
Here is the catch: the blacklist is URLs. Not ip addresses, *URLS*.
How do you know what URL a packet is going to? Packets are at a different layer in the protocol layer. The only way to filter a URL on a packet basis is to capture the whole stream and statefully decode it, buffering the packets in memory and decoding the HTTP protocol to figure out the URL.
So yes, dropping a packet going to / coming from an ip address is something a filter can do efficiently. Decoding billions of simultaneous requests at the protocol layer is *not*. It requires linear CPU and memory proportional to the number of connections, and if an ISP has hundreds of thousands of active customers it's not unreasonable to think they may need a farm of at *hundreds* of filters to do all this work.
I'm a junkie for a hi res screen, I don't think I can take a step back from the beautiful ipod at 480x320. Not to mention that many many web sites are going to be optimized for this slightly higher res so I think 320 x 240 just isn't going to cut it.
The other main criterion that I am waiting for before I splurge on an "all-in-one" device is stereo bluetooth. So many phones support bluetooth but not the A2DP profile (if I understand right, Android itself doesn't support A2DP at all). Why is this so hard to do?
Actually, in all seriousness, this is what scares me more than anything about this insane plan.
Because such businesses *will* spring up, and pedophiles and criminals will indeed flock to them. And within a very short space of time we will have laws introduced that make it not just impractical to bypass the filter, but *illegal* to circumvent it or to offer technology or services for doing so. From there we are but a short hop from making all but government approved (ie. backdoored) encryption illegal and from there to complete totalitarian control over society.
So this rather innocent sounding filter has the potential to cascade, through it's obvious flaws and the inevitable embarrassment of the government when they become publicized, to something much much more serious.
I don't get all this hand wringing as if Yahoo is going broke.
They are booking a profit of ~ $600m / year. Steady or increasing for years now. They own the #1 visited web site on the internet, zillions of people using email, IM, Answers, buzz...
I wonder if it replies in reverse? If I stupidly pay twice as much for something as I could get it elsewhere, can I go back and claim the deal was "too good to be true" for the store I bought it from? That would be rather nice!
The problem is - a key benefit of open source is the authority that you get from having participated.
I recently attended a job interview and a key interest of theirs was that I was a contributor to the web framework stack they were using (Struts). Fortunately I made some contributions under my real name, so I was able to point them to a quick google search that verified the veracity of my claims. However since that time I became much more paranoid about my online privacy and contributed everything anonymously. I now realizing that this has a big downside - it is *very* hard for me to now benefit from any of that work that I have done. So I'm considering actually reverting back to using my real name.
In fact, I'm considering a hybrid scheme - I will use my real name to project a professional persona, but a pen name to engage in more casual stuff (even on the same project). You just don't really want your professional name to show up involved in flame wars or joking around or any overly passionate statements in arguments - it just doesn't look good to an employer no matter how right or wrong you are in context. Yet that is necessary at times to participate in the open source world. So - two names it is.
Exactly my issue. I make my bread and butter coding Java and I will not tolerate a second rate native environment for doing that. By the time you consider that about 80% of the enterprise coding market is either.NET (ie. windows) or Java, the Mac doesn't look so great any more. Apple should just give up on trying to control java on the mac and throw it's weight into making the OpenJDK work on it so that they can get out of the way.
I've had a lot of friends switch and all I hear is a litany of complaints about how this or that is broken or doesn't work on the Mac (example: one colleague is completely unable to do remote debugging in Eclipse - it inexplicably crashes as soon as it tries to connect).
> Kyoto when the protocol itself is inherently unfair to developed countries?
Well spoken from a developed country point of view. But look at it from the other point of view - developed countries are responsible for 95% of the CO2 emitted to date. Why should developing countries cripple their economies to pay for our pollution? Don't you think developed economies should have some accountability for the damage done? Remember that thing about moral hazard everyone was complaining about with the financial collapse?
In any case the whole point of Kyoto was to get developing economies on board in the first round and for them to actually make cuts in the second round. If the US and Australia had actually signed on to the first round then they might have some moral authority in arguing for China and India to take action in the second round. But having twiddled their thumbs and done nothing for 10 years they have no moral high ground to stand on.
>When the test goes live, I'll opt-in to the kiddy filter and complain when I still see some naughty bits. I'll find the sites that have been accidentally blocked (there is no doubt that there will be some, the government's own tests showed that between 2% and 8% of the internet will be accidentally blocked), and complain when I can't reach them
Personally, I think more than complaining is necessary... I think an alliance needs to form in advance to start a class action law suit to sue the government en masse for the impact of any disruption caused. Once they see a line of lawyers forming ready to start high profile suits for damages they may start to take some notice.
Sure, but the problem is, there are some places where google apps is actually being touted as a viable complete replacement for MSOffice / Open Office.
Anyone who writes any kind of non-trivial document knows this is just stupid, Google Docs is just horrendous for it.
I really don't get where this line comes from. Yahoo is not dead just because you personally moved on. They bring in around $6 billion in revenue, let's write that out: 6,000,000,000, yeah it's more digits than a freakin phone number! Their net profit is ~$600 mil / year. They own the highest traffic site on the internet and dozens in the top 100 and are one of the world's most recognized brand names.
There is simply no sense in which Yahoo is either a dead brand or a dead service.
Oh, come on, don't you think MS would have killed to have a monopolistic app store for their platforms at any time in the last 20 years? Of course they would have.
The only reason they *can* do it now is because first Apple and then Google have paved the way and ensured they won't get shot down on competition grounds. So welcome to the new era where the one thing that Windows really had going for it - a truly open platform where anybody could write and deploy an app without fear or favor - is going to get slowly eaten away. An you can thank Steve Jobs for it.
Just to show it can work the other way as well: I remember watching Fox about 3 months after dubya was elected - it was right in the midst of the dot com crash and there was a minor (like 1 - 2%) dead cat bounce rally in the market. They devoted a whole hour segment to what they coined "the Bush recovery" - he had been in office for 1 month!
> In socialized medicine you get the luck of the draw - and that can mean very good care. In private care, you get a choice (to and extent) - and your care is predicated on how well you chose.
That's not the case in Australia or most other places I know of. There seem to be all these negative myths about socialized healthcare floating around. You can go and see *any doctor you like* under Australia's "universal healthcare". They can refer you to any specialist and any hospital for treatment that either they or you prefer. There are no constraints - whoever you see, the bill goes back to the government (or at least is offset by it).
Sorry, I did misread you re: DNS vs ip. However, even DNS is still tremendously problematic, especially in the age of Web 2.0 - would you take down all of youtube.com because an illegal video appeared there?
> The purpose isn't to make it "impossible" or whatever. It's not even to make it "hard". It's to give the general public a filter to shield them from what the government believes the general public wants to be shielded from.
No. Here you are wrong and in a very important way. For the purpose of shielding people an optional, opt-out filter would be fine. The core issue here is that they are introducing a "non-optional" filter to cover so-called "illegal" content (I'm not sure what "illegal" is meant to mean here - if it's hosted overseas then it is clearly not "illegal" under Australian law for the content to exist). So this is not about shielding anybody from anything. The only rational argument for the compulsory filter is that it is intended to stop people who want to see content from seeing it. I think the optional filter is stupid but I can live with it. The compulsory filter is a dangerous violation of basic principles of free speech.
> Filtering websites with this material is easy. You just force the ISPs to blacklist certain addresses from their DNS, and hire some puritans to maintain the blacklist. No, it isn't perfect, but neither are Customs officers. And it won't even result in much of a slow down.
No, you can't filter based on IP address. Many (even most) web sites are operated on shared hosting where hundreds of sites live at the same ip address. You have to dip into the protocol layer to at least look at the URLs (that means fully decoding and buffering packets and joining them in a stateful way). I'm not sure how it's even possible when the connection is SSL, but it wouldn't surprise me if they have some horrific way to attempt to do that as well.
This is sadly my conclusion as well. If you look at the timeline of when the "optional" filter suddenly changed to "mandatory" you will see that it happened about a week after Fielding had a sudden change of heart and supported most of the government's budget bills such as the changes to the medicare surcharge threshold. The only explanation he gave at the time was that the financial crisis meant people needed an extra tax cut - which made no sense because his whole argument previously was that cutting the threshold would raise health care costs (!). So - my conclusion is that a back room deal happened where Fielding exchanged mandatory filtering (and who knows what else) for letting large swathes of the Rudd agenda through parliament.
This probably means we are screwed in trying to oppose this - they will have to go ahead no matter how bad the trials go.
This is the first time in a long time I have seen anyone dumb enough and ignorant enough to trot out that old line.
Java blows ruby out of the water on the server side any way you like to measure it. In fact, for a lot of uses it competes pretty well with C / C++.
He's not worried about how to extend it. He's worried about how to maintain those extensions when rails itself changes - which is a particularly acute problem with dynamic languages like Ruby and the very extension methods you are advocating.
Since there are no statically defined interfaces or type checks you are going to have hundreds of silent / subtle bugs appear as you upgrade rails if you are not careful. (Careful probably means documenting the changes to the extreme and implementing a huge wad of unit tests to exercise every possible code path).
> This is really not a surprise. IE is an inferior product. It always has been. The market share it has received is solely attributable to the bundling with the Microsoft operating systems.
Maybe you only been using computers for 5 years?
Because from about 2000 - 2004 Netscape was the hugest piece of crap imaginable and the Mozilla alternative was horrifically bloated and unusable. IE exceeded them in quality, features, performance, every single metric you can think of. It was an excellent product. As a web developer in 2000 I remember breathing a sigh of relief when clients said not to bother coding for for Netscape because IE was so much easier and more pleasant to work with (we still made things work cross browser, but not having that requirement was a huge blessing).
I don't understand why people insist on rewriting this history.
Strange as it may sound, it may actually be the Chief Security Officer who wants IE most.
Why? Because it's only IE that supports the extensive group policy admin type controls that let him sit at his console and control every aspect of what all the browsers on the legion of locked down machines under his / her supervision can do.
> but how could Sun justify paying that much?
They bought it at the peak of it's hype. I think it was a mistake. MySQL is popular because it is the default database in a standard LAMP stack which is the cheapest hosting you can get in 99% of web hosting shops. If you were on a tight budget you were pretty much forced to use MySQL.
However I believe we are in the sunset of that time now - the world of cheap hosting is moving from shared servers to VPS servers and that means you can choose your own image from thousands that include any database or language you want. LAMP is no longer much cheaper than using any language and database stack you want. And with next-generation frameworks like Rails, Django and Grails PHP is no longer the easiest way to make your website quickly.
Why did Sun buy MySQL anyway? I think it was 2 reasons: 1 to try and get mindshare in the software services market. They need to execute a transformation like what IBM has done but they need to do it more quickly. They therefore bought a pure services company as an entry point. I think it will fail because MySQL is at the zenith of it's popularity right now. The other reason they bought it I think was a hedge against Oracle. The fact is, main reason people buy Sun boxes these days *is to run Oracle*. Therefore Sun is existentially dependent on another company. That other company bought key components of MySQL themselves and made an aggressive push to market themselves on Linux instead of Solaris. I think Sun wants a hedge / weapon against this - effectively, if Oracle continues to push people off Sun boxes by selling them OSS operating systems, Sun will attack Oracle's market share by selling their own boxes as a full solution including OSS databases instead of Oracle.
In other words, it's a game of "tit for tat.".
I heard coverage on radio on the morning of the protest.
It was depressing - they featured people from both sides of the argument but gave much more time to the "pro" filtering side wherein the points were made that something like 70% of Australians support mandatory filtering when polled, and a long opinion from a "technical specialist" who stated that all technical problems had been solved and it was possible to filter with zero impact on performance. [ of course, it turns out that said "expert" runs a company selling filtering software ... duh ].
So ... this left me fairly depressed about the whole thing.
Why yes it does!
http://www-jpc.physics.ox.ac.uk/DemoLinux.html
[ Warning: above link will in fact boot linux in your browser - I'm serious! ]
> Rudd's interest in this is that both the filtering and the national broadband scheme were election promises, and while I admire his integrity in trying to carry through with all of his election promises (unlike the previous mob, who turned election lying into a high art), I really wish he would dump the promises that were clearly stupid.
I want to pick on one incorrect point you made because it is very important history is not rewritten here. In the election campaign the Rudd government promised an *optional* filter. That is, one people could opt out of by notifying their ISP.
In August this year it was leaked that the filter had become mandatory.
Please do not allow Conroy and Rudd to get away with claiming they are just "fulfilling an election promise". They are not - they are breaking an election promise. (it is worth noting that the in same week the filter became mandatory Fielding magically turned around and passed reams of government legislation in the senate - this gives me very little hope that Conroy is going to change - if a back room deal has been done then it's been done).
> drops packets targeted by the blacklist,
Here is the catch: the blacklist is URLs. Not ip addresses, *URLS*.
How do you know what URL a packet is going to? Packets are at a different layer in the protocol layer. The only way to filter a URL on a packet basis is to capture the whole stream and statefully decode it, buffering the packets in memory and decoding the HTTP protocol to figure out the URL.
So yes, dropping a packet going to / coming from an ip address is something a filter can do efficiently. Decoding billions of simultaneous requests at the protocol layer is *not*. It requires linear CPU and memory proportional to the number of connections, and if an ISP has hundreds of thousands of active customers it's not unreasonable to think they may need a farm of at *hundreds* of filters to do all this work.
I'm a junkie for a hi res screen, I don't think I can take a step back from the beautiful ipod at 480x320. Not to mention that many many web sites are going to be optimized for this slightly higher res so I think 320 x 240 just isn't going to cut it.
The other main criterion that I am waiting for before I splurge on an "all-in-one" device is stereo bluetooth. So many phones support bluetooth but not the A2DP profile (if I understand right, Android itself doesn't support A2DP at all). Why is this so hard to do?
Actually, in all seriousness, this is what scares me more than anything about this insane plan.
Because such businesses *will* spring up, and pedophiles and criminals will indeed flock to them. And within a very short space of time we will have laws introduced that make it not just impractical to bypass the filter, but *illegal* to circumvent it or to offer technology or services for doing so. From there we are but a short hop from making all but government approved (ie. backdoored) encryption illegal and from there to complete totalitarian control over society.
So this rather innocent sounding filter has the potential to cascade, through it's obvious flaws and the inevitable embarrassment of the government when they become publicized, to something much much more serious.
I don't get all this hand wringing as if Yahoo is going broke.
They are booking a profit of ~ $600m / year. Steady or increasing for years now. They own the #1 visited web site on the internet, zillions of people using email, IM, Answers, buzz ...
Why such a gloomy analysis?
I wonder if it replies in reverse? If I stupidly pay twice as much for something as I could get it elsewhere, can I go back and claim the deal was "too good to be true" for the store I bought it from? That would be rather nice!
The problem is - a key benefit of open source is the authority that you get from having participated.
I recently attended a job interview and a key interest of theirs was that I was a contributor to the web framework stack they were using (Struts). Fortunately I made some contributions under my real name, so I was able to point them to a quick google search that verified the veracity of my claims. However since that time I became much more paranoid about my online privacy and contributed everything anonymously. I now realizing that this has a big downside - it is *very* hard for me to now benefit from any of that work that I have done. So I'm considering actually reverting back to using my real name.
In fact, I'm considering a hybrid scheme - I will use my real name to project a professional persona, but a pen name to engage in more casual stuff (even on the same project). You just don't really want your professional name to show up involved in flame wars or joking around or any overly passionate statements in arguments - it just doesn't look good to an employer no matter how right or wrong you are in context. Yet that is necessary at times to participate in the open source world. So - two names it is.
Exactly my issue. I make my bread and butter coding Java and I will not tolerate a second rate native environment for doing that. By the time you consider that about 80% of the enterprise coding market is either .NET (ie. windows) or Java, the Mac doesn't look so great any more. Apple should just give up on trying to control java on the mac and throw it's weight into making the OpenJDK work on it so that they can get out of the way.
I've had a lot of friends switch and all I hear is a litany of complaints about how this or that is broken or doesn't work on the Mac (example: one colleague is completely unable to do remote debugging in Eclipse - it inexplicably crashes as soon as it tries to connect).
> Kyoto when the protocol itself is inherently unfair to developed countries?
Well spoken from a developed country point of view. But look at it from the other point of view - developed countries are responsible for 95% of the CO2 emitted to date. Why should developing countries cripple their economies to pay for our pollution? Don't you think developed economies should have some accountability for the damage done? Remember that thing about moral hazard everyone was complaining about with the financial collapse?
In any case the whole point of Kyoto was to get developing economies on board in the first round and for them to actually make cuts in the second round. If the US and Australia had actually signed on to the first round then they might have some moral authority in arguing for China and India to take action in the second round. But having twiddled their thumbs and done nothing for 10 years they have no moral high ground to stand on.
>When the test goes live, I'll opt-in to the kiddy filter and complain when I still see some naughty bits. I'll find the sites that have been accidentally blocked (there is no doubt that there will be some, the government's own tests showed that between 2% and 8% of the internet will be accidentally blocked), and complain when I can't reach them
Personally, I think more than complaining is necessary ... I think an alliance needs to form in advance to start a class action law suit to sue the government en masse for the impact of any disruption caused. Once they see a line of lawyers forming ready to start high profile suits for damages they may start to take some notice.
Sure, but the problem is, there are some places where google apps is actually being touted as a viable complete replacement for MSOffice / Open Office.
Anyone who writes any kind of non-trivial document knows this is just stupid, Google Docs is just horrendous for it.
> Yahoo is a dead brand and a dead service.
I really don't get where this line comes from. Yahoo is not dead just because you personally moved on. They bring in around $6 billion in revenue, let's write that out: 6,000,000,000, yeah it's more digits than a freakin phone number! Their net profit is ~$600 mil / year. They own the highest traffic site on the internet and dozens in the top 100 and are one of the world's most recognized brand names.
There is simply no sense in which Yahoo is either a dead brand or a dead service.
Oh, come on, don't you think MS would have killed to have a monopolistic app store for their platforms at any time in the last 20 years? Of course they would have.
The only reason they *can* do it now is because first Apple and then Google have paved the way and ensured they won't get shot down on competition grounds. So welcome to the new era where the one thing that Windows really had going for it - a truly open platform where anybody could write and deploy an app without fear or favor - is going to get slowly eaten away. An you can thank Steve Jobs for it.
Just to show it can work the other way as well: I remember watching Fox about 3 months after dubya was elected - it was right in the midst of the dot com crash and there was a minor (like 1 - 2%) dead cat bounce rally in the market. They devoted a whole hour segment to what they coined "the Bush recovery" - he had been in office for 1 month!
> In socialized medicine you get the luck of the draw - and that can mean very good care. In private care, you get a choice (to and extent) - and your care is predicated on how well you chose.
That's not the case in Australia or most other places I know of. There seem to be all these negative myths about socialized healthcare floating around. You can go and see *any doctor you like* under Australia's "universal healthcare". They can refer you to any specialist and any hospital for treatment that either they or you prefer. There are no constraints - whoever you see, the bill goes back to the government (or at least is offset by it).
Cheers,
Simon.
Sorry, I did misread you re: DNS vs ip. However, even DNS is still tremendously problematic, especially in the age of Web 2.0 - would you take down all of youtube.com because an illegal video appeared there?
> The purpose isn't to make it "impossible" or whatever. It's not even to make it "hard". It's to give the general public a filter to shield them from what the government believes the general public wants to be shielded from.
No. Here you are wrong and in a very important way. For the purpose of shielding people an optional, opt-out filter would be fine. The core issue here is that they are introducing a "non-optional" filter to cover so-called "illegal" content (I'm not sure what "illegal" is meant to mean here - if it's hosted overseas then it is clearly not "illegal" under Australian law for the content to exist). So this is not about shielding anybody from anything. The only rational argument for the compulsory filter is that it is intended to stop people who want to see content from seeing it. I think the optional filter is stupid but I can live with it. The compulsory filter is a dangerous violation of basic principles of free speech.
> Filtering websites with this material is easy. You just force the ISPs to blacklist certain addresses from their DNS, and hire some puritans to maintain the blacklist. No, it isn't perfect, but neither are Customs officers. And it won't even result in much of a slow down.
No, you can't filter based on IP address. Many (even most) web sites are operated on shared hosting where hundreds of sites live at the same ip address. You have to dip into the protocol layer to at least look at the URLs (that means fully decoding and buffering packets and joining them in a stateful way). I'm not sure how it's even possible when the connection is SSL, but it wouldn't surprise me if they have some horrific way to attempt to do that as well.
This is sadly my conclusion as well. If you look at the timeline of when the "optional" filter suddenly changed to "mandatory" you will see that it happened about a week after Fielding had a sudden change of heart and supported most of the government's budget bills such as the changes to the medicare surcharge threshold. The only explanation he gave at the time was that the financial crisis meant people needed an extra tax cut - which made no sense because his whole argument previously was that cutting the threshold would raise health care costs (!). So - my conclusion is that a back room deal happened where Fielding exchanged mandatory filtering (and who knows what else) for letting large swathes of the Rudd agenda through parliament.
This probably means we are screwed in trying to oppose this - they will have to go ahead no matter how bad the trials go.