Actually, this is not going to help any more. These "disturbed kids" don't need psychiatric treatment, parents to listen, or bullies to go away. They need society at large to learn tolerance for people who are different.
From personal experience and plenty of observation while studying Psychology, it has become obvious to me that the greatest pressue on "geeks" comes from administrators and parents, not bullies. There is a (mostly unspoken) demand for conformance to the norms of society. Daddy harping on every day that "people your age don't all wear black" doesn't do a lot of good, neither does the teacher talking down to you because you're not enthusiastic in history class.
Believe it or not, people NEED bullying. It is a damn good preparation for later life, when things aren't all going to go your way. Protect a child all the time and it doesn't learn to protect itself; and personal emotional defense is particularly important in modern society.
The attitude that "kids with problems must be found and helped" is dangerous. It is intrinsically and ironically a problem-causer, not solver. The "kids with problems" label casts immediate suspicion on anyone "different" in society, forcing further the issue of conformance to norms. The more rigid a society's rules, the more severe the nature in which they are broken.
The BSD license is intended to do exactly the opposite of what you want, for good reasons. You do get credit in the source code (assuming the Copyright belongs to you originall), but the license specifically prohibits using your name to promote the product (which sortof restricts giving you credit).
This sucks in one way - getting credit is nice. On the other hand, someone could make a really lousy application based on your code, and there's your name on it. Uh-oh! No credit protects you from this sort of abuse of your reputation.
Maybe the only/best solution would be that to add a clause to the BSD license forcing any user of the code to contact you and ask for permission to credit you - then you can evaluate the product and decide whether to allow your name to be used or not.
Okay, if you want to know what an object is, and why its not a collection of data, place refer to a book on OOP. If you still don't understand, don't bother to continue this conversation - you obviously don't see the benefit in OOP for your purposes or applications, so there's no point in forcing it on you.
Also, once you've read up on OOP, you should understand the difference between returning an error, and throwing an exception. They are entirely different programming constructs.
Object discovery is the idea of using a registration or naming service. You tell a server that "I have an object [here] that conforms to interface [xyz]". Its a combination of CORBA's Interface Repository and directory services.
Yes, you can implement a service service to do this for you. That's exactly the problem - your implementation is not a standard. XML-RPC is that and nothing more - its RPC over XML. It is a textual representation of a binary data transfer to invoke a procedure that is not executed locally. If that's what you want, that's what you get. BUT XML-RPC is not toted as "just another RPC"... its the standard solution to all your distributed communications problems.
Bullshit. It solves ONE of the multitude of problems associated with distributed computing. CORBA has 12 different aspects to it - communication (which is what XML-RPC does, and nothing else) is just one of those 12 aspects.
My point was not to say that XML-RPC doesn't work. It was to say that XML-RPC is shortsighted. It doesn't do what it SHOULD do to facilitate distributed computing.
As for the XML-RPC spec, it specifies that XML-RPC works over HTTP. It specifies expected responses from the HTTP server (as part of the HTTP protocol). It also doesn't include a DTD or any other formal/rigid specification of dialect. The specification is a set of examples that is mostly cohesive, but no substitute for The Real Thing. If you really want it to be a standard, why not release an RFC about it containing a DTD or BNF description of the protocol?
I agree that you should examine requirements and determine the best tool for the job. I believe that neither of these protocols is a good tool. XML-RPC could be so much more (and still remain simple) if it widened its eyes to the problems of modern computing, and didn't sit solving problems that have been solved a thousand times already.
Add proper exception handling and object addressing to XML-RPC (that's one additional return type and one additional request attribute) and you've got most of the power of SOAP. Now specify that XML-RPC is meant to (but doesn't have to) work within a framework that considers object discovery, security, and so on, and define some standard interfaces to support this functionality, and you've got a decent system.
SOAP supports objects. Point to SOAP. SOAP is part of a framework (WSDL) which supports object discovery. Point to SOAP. SOAP is independant of HTTP and can function over any (suitable) transport. Point to SOAP. SOAP is fully specified and well documented. Point to SOAP. SOAP includes error handling similar to exception constructs in modern languages. Point to SOAP.
That's 5-2 to SOAP unless I've miscounted.
Take a look at the XML-RPC spec and the SOAP spec. Oh... if you are a C coder, don't bother, because XML-RPC is better. If you are a programmer or any other sort of developer with experience in technologies introduced since 1980, you should understand.
You've hit this one on the nose: ESR's opinions are important because he does real coding. This nicely sums up the Unix spirit, and the shortsightedness of The Unix Way as it is currently perceived.
Since the early 1990s we've seen no innovation from the *nix camp. Everything that has been done has followed innovations from MS, Sun and Mac. The very OSen/Companies we look down upon. *nix has been ahead on implementation every time, and not in the design and architecture race once. Take a look at GNOME's precious object model. Now take a look at COM. Now laugh. I thought *nix wasn't into making the same (huge) mistake twice?
XML-RPC is another prime examples of this. Don't get me wrong - I like XML-RPC - it does its job, and it has a purpose. But it doesn't solve the communication problems we need to solve. It is a non-binary version of a procedure call protocol. Yay. *nix has not yet caught onto OOP, let alone distributed OOP. MS looked at the situation and came up with SOAP - they had the right idea, but in true MS style added several tons of crap into what would have been a decent spec. Pity.
So what's the difference? XML-RPC has no architecture, no framework, no environment. It says that if I know a URL and an interface, I can make a procedure call. It doesn't give me the ability to register an object (or procedure), find a URL, and add modules to a distributed system. This is left as an exercise to the implementor who must make another (sub)standard that uses XML-RPC for its lowest level communication.
Oh yeah - XML-RPC doesn't let you deal with object references either. That means you can't pass around objects.
So before you flame me, understand that I'm not saying XML-RPC is shit because of this. I'm saying that the community is shit because they don't have the forsight to look beyond their code optimisation and plan for the future by designing a strong architecture that can lead, not follow.
Almost an entirely different league, but IMNSHO the best games ever created were Wasteland and the Ultima IV - VI.
The plot of these games is amazingly deep and multilinear, so that only after you've played the game through several times are you confident that you've found all the side plots. In fact, with Wasteland you hardly know which is the main plot until late in the game!
What gave both of these games an edge was simplicity in interface. This translates to some as limitation - "I can't use SMGs in both hands!" - but it makes the game more playable, allowing immersion and focus on the story, rather than fiddling with the interface.
I know a number of people who prefer Doom over Quake for this very reason - our 2D interfaces don't translate well into controls for a 3D environment. Learning to play Quake well is a long and irritating process. Why bother when I can shoot up beasties in Doom instead?
We need to see more simple interfaced, amusing and captivating games that have non-linear plot and sacrifice flashyness for story rather than vice versa.
The idea of downloading all sorts of items for your home is a little absured - at least for quite some time. As pointed out, fabbers have a very limited range of materials with which they can work, and take a long time to produce a model.
But a really interesting application could be... electronics. openH.org has just been announced - surely fabber technology promises to be suited to their needs?
If a fabber could be made to use a combination of insulators, conductors and semi-conductors as input materials, you could download the design to print a circuit board or chip (much as we have been promised for years that semi-conductive ink is just around the corner, and we'll be able to print out our next PC on a Lexmark).
Using a fabber in this manner would be a 'lec hackers fantasy: design a chip, print it out, boot it and play around.
The economic implication of this are huge: if you could cheaply prototype electronic components at home, a whole new world of consumer micro-electronics would open up, with single developers competing against established corporations, in the same was as the Internet has levelled out the e-commerce market, and home audio recording on PCs is reaching professional quality (gee... I've referenced TWO recent/. stories in one post - do I get a chocolate now?)
The idea of VOIP raises some interesting concerns. Those of immediate importance to me are legal: I live in a country where there is a state sponsored monolopy on communications, and the telecoms company reserves the exclusive right to voice traffic.
The idea of economics is more of a global issue though. Many telecos have positioned themselves such that they make a loss off local calls, but recoup and profit on long distance calls and calls to other carriers. VOIP promises a sudden (well, maybe not too sudden) demise to this situation, reducing long distance traffic to data only.
The effects, as far as I can determine, would be to raise the cost of long distance data traffic in a manner disproportionate to the increase in load. You see, data requires better quality lines than voice, so many older copper lines that run long distance would need to be replaced or upgraded to keep up with the required digital load.
This, finally, is passed on in cost to everyone - not just those who don't make long distance calls, but all Internet users who have nothing to do with the load generated by VOIP.
Is anyone else worried by the prospect of all subsidising all? Or would you like to see a one-rate-calls-anywhere tarrif system on a global scale?
The problem with new solutions over floppy drives is the price. A floppy drive adds $10 to the price of your system ; disks are 15c each. ZIP can't match this in hardware or media cost. Solid state USB storage is hellishly expensive by comparison. CDR(W) media is cheap, but the drives are more expensive than a HDD and removable case/bay.
Any new removable media needs to meet a few criteria to become accepted:
Must work with existing and new systems. This means relying on technologies such as IDE, USB, serial or parallel communication, or existing removable media devices.
The hardware must be cheap. Both home and business users are slow to find cash for new technology that isn't widely used.
Related to the above: ideally no reader hardware should be required ; so the removable "media" should actually be a hardware solution connecting directly to USB/serial/parallel.
The media must be hotswappable, and all OSes must be able to determine that a swap has occurred. Again we're down to USB/serial/parallel.
The media must be relatively large, and fast. Writing a CD is not a quick and easy task, often even more tedious than copying to a floppy. An 8Mb solid state RAM chip is simply too small, unless it is dirt cheap. Any really small data (<2Mb) can more easily be e-mailed or transferred via an Internet drive.
I don't believe there is room for another floppy standard. CDR(W)s are common enough for large data, but a cheap, accessible medium is required for smaller transfers. USB memory seems promising, but needs to become a lot cheaper before it can be viable.
Correct, under the GPL, any changes made to the TCP/IP stack would have to be released to the customer (and available to all developers everywhere).
But wait: just because the CODE is under the GPL, it doesn't mean the protocol is. Ooops. MS simply creates a clean implementation of the protocol, does an embrace and extend, and usurp the protocol.
We'll let you draw your own conclusions about what can happen under the BSD license.
Don't knock multilayer discs too quickly. Constellation 3D announced a fortnight ago that they have made a partnership to produce their flourescent multilayer discs.
According to the article we can expect 100Gb discs and 10Gb credit cards on the market by mid-2002. C3D claims a limit of 1Tb (on a disc) that they will be able to reach in a few years. This is as promising as other holographic media, if not more so.
Stigmatized? Yes... I would be sure to shut up about having a nasty disease if I was going to be "stigmatized"... considering that in this case stigmatism translated to "will be dragged out of your home in the middle of the night and beaten to death".
I think we need to clear up some facts about AIDS, African problems and drug companies before we can attack any of the parties.
First, be aware that there are several strains of AIDS, and the African strain in particular is not responsive to treatments made for the US strain, for which most research has been done.
Second, African countries feel their hands are tied by drug companies - any country attempting to produce treatments locally (under license) or import cheaply produced drugs from Asia has been met with a cool reception by the pharmaceutical conglomerates, who have threatened to withdraw support and research from the country altogether. This equates to disinvestment in countries which are already in dire economic positions.
Third, let it be noted that that the claim that "... wearing a condom... makes them less of a man" is not strictly correct. Studies have shown that African males prefer "dry sex". That's right - forget the lubricantsm try talc power instead. Personally I don't want to know... but it means that condoms aren't exciting, and the risk of blood contact is higher. Furthermore in many African societies your status in the community is linked to the number of children (and wives) you have. And to put the cherry on the top - tribal custom (in many areas) holds that the young will look after their parents when they grow old... hence more children is a good thing, and condoms aren't.
In most of these African countries there is a strong hatred for Western people and culture. This is largely due to the perceived wealth and success of the West, ostensibly at the cost of exploitation of African. The suggestion/rumour that AIDS is made by Western governments is widely believed in some areas of African, as is the belief that a cure exists, but is being withheld by the West. Alternative beliefs include that AIDS is a punnishment from the ancestors (== gods) and that strong, healthy or rich people don't get AIDS.
Of course, there's always the problem that admitting you have AIDS is a quick signature of your death certificate.
So, sure, feel free to blame Africans for their lack of knowledge, lack of doing anything about it, whatever. Until medical science became accepted in Europe disease was treated with fear and resentment, leading to inhuman treatment, killings, and the belief that God was sending punnishments on man for his heresy.
We may laugh at the foolish beliefs of African cultures. In the West you trust you security in old age to the government or investments based on perceived values of companies. That's also pretty stupid if you think about it. And let's not forget the one billion people worldwide that are told to ignore the "use a condom" message because it is "against the word of God" in the eyes of the Catholic church.
Its easy to blame other people. Its not easy to understand why exploiting them is so easy. Its almost as good business as selling crosses to Christians:)
In the case of SSH, this should be an obvious route. Many applications that work over ssh do so by treating it as rsh. This would even get you around licenses such as those covering mySQL (free as long as you don't lock the product to mySQL specifically).
Interesting... I always thought that self censorship meant that I decided what I did not want to see. Not that my company made that decision, or my ISP, or their backbone ISP, or anyone else... but me.
MAPS does not give you the choice of self censorship - it gives service providers the choice of whether to subject you to censorship or not.
Aaah... so here's an idea: we get a couple of ISPs to jump on ship with the idea of DNS filtering, and then we can prevent access to sites that distribute hacking software! Certainly a more troublesome problem than spam.
Oooh! And don't forget that we could easily add porno sites to the shitlist - that would be great! And those pesky warez sites... now that would make the BSA real friendly-like to us. Wow - you could even take/. offline for being politically incorrect.
Yes... let's fuck the ISP because they uphold freedom of information, what a wonderful place the world will be.
I think most people here seem to be missing the real reasons that telecoms companies are keen to curb VOIP. It has nothing to do with interconnection, and while it does have everything to do with control and profit, you're getting the wrong end of the stick.
Almost all telecoms companies make profit exclusively on long distance communication. Local comms are expensive to maintain and low profit - in fact sometimes they are even free and therefore generate no revenue.
VOIP is an effective means of circumventing the long distance revenue stream - cut all long distance traffic down to data only by encapsulating voice and fax in IP, and you bring a telecomms company to its knees.
Now while I like the idea of telecomms companies in a compromising position, I have a serious problem with the idea of having my local phone call costs increased because the company cannot break even when its long distance lines aren't being used. Its either they (long distance) carry us, or we (local call to ISP) carry them. I prefer the formed. So do you.
As for the technical "it can/can't be done" argument - there is no way to prevent such technology (VOIP) from being used. By telecomms companies can limit it. The vast majority of 'people out there' are scared of words like "illegal", and a sufficient awareness campaign will have them cowering and running "VOIP insta-remove" programs on their computers without a second thought. The minority that 'abuse the system' and slip through the cracks aren't going to break the profit margins.
Of course, an OID system still doesn't solve the problem it purports to address. So you have a registry and an object handle in that registry - what happens when the object is removed, or moved to a different place?
If you change service providers, will you still be using your old OIDs? I doubt it... use of the registries is hardly going to be free. So you're back to the 404 problem... only this time you have to remember what looks like a phone number with a name on the end, instead of a nice simple URL.
Oh, and while we're at it... let's throw DNS into a new crisis by negating the value of everyone's domain names... whoops!
Regarding patents, the situation should be (and as far as I understand, it is in the US, with the exception of the NSA) that once a patent is filed it is publically accessible. That knowledge is no longer hidden. A trade secret and a patent are mutually exclusive.
I strongly believe that patents are still a good thing, but that the laws need to change to be relevant to modern times. Maybe a patent should last for 3 to 5 years - certainly enough time to get a competitive advantage. Also, if a patent holder does not use the patent constructively within one year (or show that the patent is part of an ongoing development within that year) then the holder must be forced to license the patent to one or more other parties, at a reasonable price.
Copyright is a more difficult thing. If I am an author, how can I expect to survive if I may sell only one copy of the book? As an author I don't have access to the resources of publishing companies, so the first publishing house to buy my book sells lots of copies to lots of people, and I get nothing. Some authors already work on this system - but the publishing house must take the financial risk of playing the author up front, so the pay is low, and encourages pulp fiction. I'm sure we can do better in this world.
As a software programmer, I can't even get a lump sum payment out of a publishing house. I sell one copy of my software, and it gets posted to a download site, and that's my $20 income for the year. No-one is going to pay me for a public appearance to autograph the software. And I certainly can't go to a concert. So what exactly must I do...?
Some people think that a service industry is the answer. But its not. There is a huge market for utility applications - a market that does not require a lot of support or service, but requires working utilities. If you release a single good bug-free product, that's it - no more income for you. So either you have to loan your software (which you can't do, because there is no copyright) or you have to write buggy software to ensure that you get paid again.
On the music artist issue - how are you going to AFFORD your first concert? You didn't make any money from the music, remember? Also, in your haste to screw over all authors and software programmers, you seem to have also screwed over your loyal fans... because they are the only ones who will pay for your music (you admit it)!
Copyright is still (and most likely always will be) required. But copyright must be carefully balanced against fair use. No-one should be allowed to copy a copyright work in its entirity - they must obtain one from the author. After that, however, the author should lose rights to that copy - it can be pressed, prodded, ripped apart, put back together, resold, whatever.
An ostrich has feathers and wings, but cannot fly. What is the evolutionary advanage in this? By studying it in its habitat we can determine reasons - but we honestly don't know what dinosaurs ate, not can we watch them eating.
Long necks can provide several advantages apart from reaching high food sources. First, long necks and tails are used by mammels today as weapons (a giraffe can kill a lion by using its head as a club).
Second, IIRC during various periods that had long-necked dinosaurs there were few trees and lots of swamp. A long neck would enable the dinosaur to reach foods on marshy ground without having to enter the marshy ground (and potentially become stuck).
A good reason to restrict a DTD or interface would be for compatibility. Standards are useless when they are extended in an ad hoc fashion. MS has made a point of this, hence their "embrace and extend" policy.
A DTD or interface are as much works in their own right as software programs. I'm not sure of the situation in US law, but under International Copyright Law they all derive protection from the same principle: authorial copyright. You cannot, without permission, produce an exact copy or a derived work.
Unfortunately the GPL is incompatible with DTDs or interfaces, as it explicitly allows (even encourages) modification, whereas the sole purpose of an interface or DTD is to provide a fixed and unmodifable standard.
Further changes to the GPL to curtial the activities of ASPs are likely to hurt free software far more than assist it. Although ASPs may be considered "freeloaders" and apparently are not giving back to the community, they contribute in the form of advertising.
Whether they are blatent or not, word gets around that major sites and companies are using free software, which gives it a credability that cannot be obtained through any amount of technical ranting.
Free software can come to prominance only through acceptance by commerce. Commerce still decides the standards for applications and communication - until that mindset changes to accept open source standards, free software will be relegated to being considered a second rate solution that must support other peoples' standards.
If open source software is going to take a lead in the computing world, and stop being a follower, it must start to define its own (open) standards, and prevent them from being maligned through extension. Hence the important of rigid, unmodifiable DTDs and interfaces.
Consider the example of HTML: despite it being a W3C controlled standard, every HTML author must be aware of the various non-standard behaviours of different browsers. Using "valid" HTML (according to a W3C lint) can restrict the aesthetics of the document, which is why so many authors choose to develop for a particular browser. JavaScript (now ECMAScript) is another classic example.
Standards are only useful when the major players in the arena agree to honour them, for mutual benefit. But we are in Unreal Tournament here - few major players want to cooperate, and free software is hardly a major player, and cannot dictate the rules. Ever seen the ref catch it in a boxing match?
Free software needs to gain credability in order to dictate the way ahead. To do so, it must develop standards for others to follow, become accepted by commerce, and shun the modification of open standards. It cannot do this by shutting out business. It cannot do this through in-fighting and ego mania. It cannot do this by closing its standards inside code that commerce cannot use. It cannot do this through the GPL.
A brand new network object model, nothing like MS's terrible COM. Why then are the core interfaces identical (except in name) and the system does not solve any of the problems that COM presents?
More painfully true than duplication of effort, is the OpenSource communities inability to grasp the benefits of reusable components (no, libraries are not OO and therefore not reusable components) let alone create a standard. Before anyone even thinks of saying "GNOME does that", please remember that the biggest bitch about Win32 is that everything is linked to the GUI.
Try Zero Pollution Motors / MDI
on
Air-Powered Cars
·
· Score: 1
Anyone firing up their local search engine should have found Zero Pollution Motors, the manufacturers of this car. Northerlight has a special report article on it in October last year.
According to this site a taxi version was first road tested in Franch in May 1998. I also read somewhere that Mexico City is looking at replacing its entire fleet of 40000 taxis where these vehicles.
Slashdot Mirror
Actually, this is not going to help any more. These "disturbed kids" don't need psychiatric treatment, parents to listen, or bullies to go away. They need society at large to learn tolerance for people who are different.
From personal experience and plenty of observation while studying Psychology, it has become obvious to me that the greatest pressue on "geeks" comes from administrators and parents, not bullies. There is a (mostly unspoken) demand for conformance to the norms of society. Daddy harping on every day that "people your age don't all wear black" doesn't do a lot of good, neither does the teacher talking down to you because you're not enthusiastic in history class.
Believe it or not, people NEED bullying. It is a damn good preparation for later life, when things aren't all going to go your way. Protect a child all the time and it doesn't learn to protect itself; and personal emotional defense is particularly important in modern society.
The attitude that "kids with problems must be found and helped" is dangerous. It is intrinsically and ironically a problem-causer, not solver. The "kids with problems" label casts immediate suspicion on anyone "different" in society, forcing further the issue of conformance to norms. The more rigid a society's rules, the more severe the nature in which they are broken.
The BSD license is intended to do exactly the opposite of what you want, for good reasons. You do get credit in the source code (assuming the Copyright belongs to you originall), but the license specifically prohibits using your name to promote the product (which sortof restricts giving you credit).
This sucks in one way - getting credit is nice. On the other hand, someone could make a really lousy application based on your code, and there's your name on it. Uh-oh! No credit protects you from this sort of abuse of your reputation.
Maybe the only/best solution would be that to add a clause to the BSD license forcing any user of the code to contact you and ask for permission to credit you - then you can evaluate the product and decide whether to allow your name to be used or not.
Way cool! A non-flame answer :)
Okay, if you want to know what an object is, and why its not a collection of data, place refer to a book on OOP. If you still don't understand, don't bother to continue this conversation - you obviously don't see the benefit in OOP for your purposes or applications, so there's no point in forcing it on you.
Also, once you've read up on OOP, you should understand the difference between returning an error, and throwing an exception. They are entirely different programming constructs.
Object discovery is the idea of using a registration or naming service. You tell a server that "I have an object [here] that conforms to interface [xyz]". Its a combination of CORBA's Interface Repository and directory services.
Yes, you can implement a service service to do this for you. That's exactly the problem - your implementation is not a standard. XML-RPC is that and nothing more - its RPC over XML. It is a textual representation of a binary data transfer to invoke a procedure that is not executed locally. If that's what you want, that's what you get. BUT XML-RPC is not toted as "just another RPC" ... its the standard solution to all your distributed communications problems.
Bullshit. It solves ONE of the multitude of problems associated with distributed computing. CORBA has 12 different aspects to it - communication (which is what XML-RPC does, and nothing else) is just one of those 12 aspects.
My point was not to say that XML-RPC doesn't work. It was to say that XML-RPC is shortsighted. It doesn't do what it SHOULD do to facilitate distributed computing.
As for the XML-RPC spec, it specifies that XML-RPC works over HTTP. It specifies expected responses from the HTTP server (as part of the HTTP protocol). It also doesn't include a DTD or any other formal/rigid specification of dialect. The specification is a set of examples that is mostly cohesive, but no substitute for The Real Thing. If you really want it to be a standard, why not release an RFC about it containing a DTD or BNF description of the protocol?
I agree that you should examine requirements and determine the best tool for the job. I believe that neither of these protocols is a good tool. XML-RPC could be so much more (and still remain simple) if it widened its eyes to the problems of modern computing, and didn't sit solving problems that have been solved a thousand times already.
Add proper exception handling and object addressing to XML-RPC (that's one additional return type and one additional request attribute) and you've got most of the power of SOAP. Now specify that XML-RPC is meant to (but doesn't have to) work within a framework that considers object discovery, security, and so on, and define some standard interfaces to support this functionality, and you've got a decent system.
SOAP supports objects. Point to SOAP. SOAP is part of a framework (WSDL) which supports object discovery. Point to SOAP. SOAP is independant of HTTP and can function over any (suitable) transport. Point to SOAP. SOAP is fully specified and well documented. Point to SOAP. SOAP includes error handling similar to exception constructs in modern languages. Point to SOAP.
That's 5-2 to SOAP unless I've miscounted.
Take a look at the XML-RPC spec and the SOAP spec. Oh ... if you are a C coder, don't bother, because XML-RPC is better. If you are a programmer or any other sort of developer with experience in technologies introduced since 1980, you should understand.
You've hit this one on the nose: ESR's opinions are important because he does real coding. This nicely sums up the Unix spirit, and the shortsightedness of The Unix Way as it is currently perceived.
Since the early 1990s we've seen no innovation from the *nix camp. Everything that has been done has followed innovations from MS, Sun and Mac. The very OSen/Companies we look down upon. *nix has been ahead on implementation every time, and not in the design and architecture race once. Take a look at GNOME's precious object model. Now take a look at COM. Now laugh. I thought *nix wasn't into making the same (huge) mistake twice?
XML-RPC is another prime examples of this. Don't get me wrong - I like XML-RPC - it does its job, and it has a purpose. But it doesn't solve the communication problems we need to solve. It is a non-binary version of a procedure call protocol. Yay. *nix has not yet caught onto OOP, let alone distributed OOP. MS looked at the situation and came up with SOAP - they had the right idea, but in true MS style added several tons of crap into what would have been a decent spec. Pity.
So what's the difference? XML-RPC has no architecture, no framework, no environment. It says that if I know a URL and an interface, I can make a procedure call. It doesn't give me the ability to register an object (or procedure), find a URL, and add modules to a distributed system. This is left as an exercise to the implementor who must make another (sub)standard that uses XML-RPC for its lowest level communication.
Oh yeah - XML-RPC doesn't let you deal with object references either. That means you can't pass around objects.
So before you flame me, understand that I'm not saying XML-RPC is shit because of this. I'm saying that the community is shit because they don't have the forsight to look beyond their code optimisation and plan for the future by designing a strong architecture that can lead, not follow.
Almost an entirely different league, but IMNSHO the best games ever created were Wasteland and the Ultima IV - VI.
The plot of these games is amazingly deep and multilinear, so that only after you've played the game through several times are you confident that you've found all the side plots. In fact, with Wasteland you hardly know which is the main plot until late in the game!
What gave both of these games an edge was simplicity in interface. This translates to some as limitation - "I can't use SMGs in both hands!" - but it makes the game more playable, allowing immersion and focus on the story, rather than fiddling with the interface.
I know a number of people who prefer Doom over Quake for this very reason - our 2D interfaces don't translate well into controls for a 3D environment. Learning to play Quake well is a long and irritating process. Why bother when I can shoot up beasties in Doom instead?
We need to see more simple interfaced, amusing and captivating games that have non-linear plot and sacrifice flashyness for story rather than vice versa.
The idea of downloading all sorts of items for your home is a little absured - at least for quite some time. As pointed out, fabbers have a very limited range of materials with which they can work, and take a long time to produce a model.
But a really interesting application could be ... electronics. openH.org has just been announced - surely fabber technology promises to be suited to their needs?
If a fabber could be made to use a combination of insulators, conductors and semi-conductors as input materials, you could download the design to print a circuit board or chip (much as we have been promised for years that semi-conductive ink is just around the corner, and we'll be able to print out our next PC on a Lexmark).
Using a fabber in this manner would be a 'lec hackers fantasy: design a chip, print it out, boot it and play around.
The economic implication of this are huge: if you could cheaply prototype electronic components at home, a whole new world of consumer micro-electronics would open up, with single developers competing against established corporations, in the same was as the Internet has levelled out the e-commerce market, and home audio recording on PCs is reaching professional quality (gee ... I've referenced TWO recent /. stories in one post - do I get a chocolate now?)
The idea of VOIP raises some interesting concerns. Those of immediate importance to me are legal: I live in a country where there is a state sponsored monolopy on communications, and the telecoms company reserves the exclusive right to voice traffic.
The idea of economics is more of a global issue though. Many telecos have positioned themselves such that they make a loss off local calls, but recoup and profit on long distance calls and calls to other carriers. VOIP promises a sudden (well, maybe not too sudden) demise to this situation, reducing long distance traffic to data only.
The effects, as far as I can determine, would be to raise the cost of long distance data traffic in a manner disproportionate to the increase in load. You see, data requires better quality lines than voice, so many older copper lines that run long distance would need to be replaced or upgraded to keep up with the required digital load.
This, finally, is passed on in cost to everyone - not just those who don't make long distance calls, but all Internet users who have nothing to do with the load generated by VOIP.
Is anyone else worried by the prospect of all subsidising all? Or would you like to see a one-rate-calls-anywhere tarrif system on a global scale?
The problem with new solutions over floppy drives is the price. A floppy drive adds $10 to the price of your system ; disks are 15c each. ZIP can't match this in hardware or media cost. Solid state USB storage is hellishly expensive by comparison. CDR(W) media is cheap, but the drives are more expensive than a HDD and removable case/bay.
Any new removable media needs to meet a few criteria to become accepted:
I don't believe there is room for another floppy standard. CDR(W)s are common enough for large data, but a cheap, accessible medium is required for smaller transfers. USB memory seems promising, but needs to become a lot cheaper before it can be viable.
Correct, under the GPL, any changes made to the TCP/IP stack would have to be released to the customer (and available to all developers everywhere).
But wait: just because the CODE is under the GPL, it doesn't mean the protocol is. Ooops. MS simply creates a clean implementation of the protocol, does an embrace and extend, and usurp the protocol.
We'll let you draw your own conclusions about what can happen under the BSD license.
Don't knock multilayer discs too quickly. Constellation 3D announced a fortnight ago that they have made a partnership to produce their flourescent multilayer discs.
According to the article we can expect 100Gb discs and 10Gb credit cards on the market by mid-2002. C3D claims a limit of 1Tb (on a disc) that they will be able to reach in a few years. This is as promising as other holographic media, if not more so.
Stigmatized? Yes ... I would be sure to shut up about having a nasty disease if I was going to be "stigmatized" ... considering that in this case stigmatism translated to "will be dragged out of your home in the middle of the night and beaten to death".
The infirm seldom make good martyrs.
I think we need to clear up some facts about AIDS, African problems and drug companies before we can attack any of the parties.
First, be aware that there are several strains of AIDS, and the African strain in particular is not responsive to treatments made for the US strain, for which most research has been done.
Second, African countries feel their hands are tied by drug companies - any country attempting to produce treatments locally (under license) or import cheaply produced drugs from Asia has been met with a cool reception by the pharmaceutical conglomerates, who have threatened to withdraw support and research from the country altogether. This equates to disinvestment in countries which are already in dire economic positions.
Third, let it be noted that that the claim that "... wearing a condom ... makes them less of a man" is not strictly correct. Studies have shown that African males prefer "dry sex". That's right - forget the lubricantsm try talc power instead. Personally I don't want to know ... but it means that condoms aren't exciting, and the risk of blood contact is higher. Furthermore in many African societies your status in the community is linked to the number of children (and wives) you have. And to put the cherry on the top - tribal custom (in many areas) holds that the young will look after their parents when they grow old ... hence more children is a good thing, and condoms aren't.
In most of these African countries there is a strong hatred for Western people and culture. This is largely due to the perceived wealth and success of the West, ostensibly at the cost of exploitation of African. The suggestion/rumour that AIDS is made by Western governments is widely believed in some areas of African, as is the belief that a cure exists, but is being withheld by the West. Alternative beliefs include that AIDS is a punnishment from the ancestors (== gods) and that strong, healthy or rich people don't get AIDS.
Of course, there's always the problem that admitting you have AIDS is a quick signature of your death certificate.
So, sure, feel free to blame Africans for their lack of knowledge, lack of doing anything about it, whatever. Until medical science became accepted in Europe disease was treated with fear and resentment, leading to inhuman treatment, killings, and the belief that God was sending punnishments on man for his heresy.
We may laugh at the foolish beliefs of African cultures. In the West you trust you security in old age to the government or investments based on perceived values of companies. That's also pretty stupid if you think about it. And let's not forget the one billion people worldwide that are told to ignore the "use a condom" message because it is "against the word of God" in the eyes of the Catholic church.
Its easy to blame other people. Its not easy to understand why exploiting them is so easy. Its almost as good business as selling crosses to Christians :)
In the case of SSH, this should be an obvious route. Many applications that work over ssh do so by treating it as rsh. This would even get you around licenses such as those covering mySQL (free as long as you don't lock the product to mySQL specifically).
"self-censorship by proxy"
Interesting ... I always thought that self censorship meant that I decided what I did not want to see. Not that my company made that decision, or my ISP, or their backbone ISP, or anyone else ... but me.
MAPS does not give you the choice of self censorship - it gives service providers the choice of whether to subject you to censorship or not.
Look! He's stabbing her in the middle of a crowded room! Quick, my gattling gun so I can stop him!
Aaah ... so here's an idea: we get a couple of ISPs to jump on ship with the idea of DNS filtering, and then we can prevent access to sites that distribute hacking software! Certainly a more troublesome problem than spam.
Oooh! And don't forget that we could easily add porno sites to the shitlist - that would be great! And those pesky warez sites ... now that would make the BSA real friendly-like to us. Wow - you could even take /. offline for being politically incorrect.
Yes ... let's fuck the ISP because they uphold freedom of information, what a wonderful place the world will be.
I think most people here seem to be missing the real reasons that telecoms companies are keen to curb VOIP. It has nothing to do with interconnection, and while it does have everything to do with control and profit, you're getting the wrong end of the stick.
Almost all telecoms companies make profit exclusively on long distance communication. Local comms are expensive to maintain and low profit - in fact sometimes they are even free and therefore generate no revenue.
VOIP is an effective means of circumventing the long distance revenue stream - cut all long distance traffic down to data only by encapsulating voice and fax in IP, and you bring a telecomms company to its knees.
Now while I like the idea of telecomms companies in a compromising position, I have a serious problem with the idea of having my local phone call costs increased because the company cannot break even when its long distance lines aren't being used. Its either they (long distance) carry us, or we (local call to ISP) carry them. I prefer the formed. So do you.
As for the technical "it can/can't be done" argument - there is no way to prevent such technology (VOIP) from being used. By telecomms companies can limit it. The vast majority of 'people out there' are scared of words like "illegal", and a sufficient awareness campaign will have them cowering and running "VOIP insta-remove" programs on their computers without a second thought. The minority that 'abuse the system' and slip through the cracks aren't going to break the profit margins.
Of course, an OID system still doesn't solve the problem it purports to address. So you have a registry and an object handle in that registry - what happens when the object is removed, or moved to a different place?
If you change service providers, will you still be using your old OIDs? I doubt it ... use of the registries is hardly going to be free. So you're back to the 404 problem ... only this time you have to remember what looks like a phone number with a name on the end, instead of a nice simple URL.
Oh, and while we're at it ... let's throw DNS into a new crisis by negating the value of everyone's domain names ... whoops!
Regarding patents, the situation should be (and as far as I understand, it is in the US, with the exception of the NSA) that once a patent is filed it is publically accessible. That knowledge is no longer hidden. A trade secret and a patent are mutually exclusive.
I strongly believe that patents are still a good thing, but that the laws need to change to be relevant to modern times. Maybe a patent should last for 3 to 5 years - certainly enough time to get a competitive advantage. Also, if a patent holder does not use the patent constructively within one year (or show that the patent is part of an ongoing development within that year) then the holder must be forced to license the patent to one or more other parties, at a reasonable price.
Copyright is a more difficult thing. If I am an author, how can I expect to survive if I may sell only one copy of the book? As an author I don't have access to the resources of publishing companies, so the first publishing house to buy my book sells lots of copies to lots of people, and I get nothing. Some authors already work on this system - but the publishing house must take the financial risk of playing the author up front, so the pay is low, and encourages pulp fiction. I'm sure we can do better in this world.
As a software programmer, I can't even get a lump sum payment out of a publishing house. I sell one copy of my software, and it gets posted to a download site, and that's my $20 income for the year. No-one is going to pay me for a public appearance to autograph the software. And I certainly can't go to a concert. So what exactly must I do ...?
Some people think that a service industry is the answer. But its not. There is a huge market for utility applications - a market that does not require a lot of support or service, but requires working utilities. If you release a single good bug-free product, that's it - no more income for you. So either you have to loan your software (which you can't do, because there is no copyright) or you have to write buggy software to ensure that you get paid again.
On the music artist issue - how are you going to AFFORD your first concert? You didn't make any money from the music, remember? Also, in your haste to screw over all authors and software programmers, you seem to have also screwed over your loyal fans ... because they are the only ones who will pay for your music (you admit it)!
Copyright is still (and most likely always will be) required. But copyright must be carefully balanced against fair use. No-one should be allowed to copy a copyright work in its entirity - they must obtain one from the author. After that, however, the author should lose rights to that copy - it can be pressed, prodded, ripped apart, put back together, resold, whatever.
An ostrich has feathers and wings, but cannot fly. What is the evolutionary advanage in this? By studying it in its habitat we can determine reasons - but we honestly don't know what dinosaurs ate, not can we watch them eating.
Long necks can provide several advantages apart from reaching high food sources. First, long necks and tails are used by mammels today as weapons (a giraffe can kill a lion by using its head as a club).
Second, IIRC during various periods that had long-necked dinosaurs there were few trees and lots of swamp. A long neck would enable the dinosaur to reach foods on marshy ground without having to enter the marshy ground (and potentially become stuck).
A good reason to restrict a DTD or interface would be for compatibility. Standards are useless when they are extended in an ad hoc fashion. MS has made a point of this, hence their "embrace and extend" policy.
A DTD or interface are as much works in their own right as software programs. I'm not sure of the situation in US law, but under International Copyright Law they all derive protection from the same principle: authorial copyright. You cannot, without permission, produce an exact copy or a derived work.
Unfortunately the GPL is incompatible with DTDs or interfaces, as it explicitly allows (even encourages) modification, whereas the sole purpose of an interface or DTD is to provide a fixed and unmodifable standard.
Further changes to the GPL to curtial the activities of ASPs are likely to hurt free software far more than assist it. Although ASPs may be considered "freeloaders" and apparently are not giving back to the community, they contribute in the form of advertising.
Whether they are blatent or not, word gets around that major sites and companies are using free software, which gives it a credability that cannot be obtained through any amount of technical ranting.
Free software can come to prominance only through acceptance by commerce. Commerce still decides the standards for applications and communication - until that mindset changes to accept open source standards, free software will be relegated to being considered a second rate solution that must support other peoples' standards.
If open source software is going to take a lead in the computing world, and stop being a follower, it must start to define its own (open) standards, and prevent them from being maligned through extension. Hence the important of rigid, unmodifiable DTDs and interfaces.
Consider the example of HTML: despite it being a W3C controlled standard, every HTML author must be aware of the various non-standard behaviours of different browsers. Using "valid" HTML (according to a W3C lint) can restrict the aesthetics of the document, which is why so many authors choose to develop for a particular browser. JavaScript (now ECMAScript) is another classic example.
Standards are only useful when the major players in the arena agree to honour them, for mutual benefit. But we are in Unreal Tournament here - few major players want to cooperate, and free software is hardly a major player, and cannot dictate the rules. Ever seen the ref catch it in a boxing match?
Free software needs to gain credability in order to dictate the way ahead. To do so, it must develop standards for others to follow, become accepted by commerce, and shun the modification of open standards. It cannot do this by shutting out business. It cannot do this through in-fighting and ego mania. It cannot do this by closing its standards inside code that commerce cannot use. It cannot do this through the GPL.
A brand new network object model, nothing like MS's terrible COM. Why then are the core interfaces identical (except in name) and the system does not solve any of the problems that COM presents?
More painfully true than duplication of effort, is the OpenSource communities inability to grasp the benefits of reusable components (no, libraries are not OO and therefore not reusable components) let alone create a standard. Before anyone even thinks of saying "GNOME does that", please remember that the biggest bitch about Win32 is that everything is linked to the GUI.
Anyone firing up their local search engine should have found Zero Pollution Motors, the manufacturers of this car. Northerlight has a special report article on it in October last year. According to this site a taxi version was first road tested in Franch in May 1998. I also read somewhere that Mexico City is looking at replacing its entire fleet of 40000 taxis where these vehicles.