I have been saying this for years: the post offices of the of the world - not the governments - can totally put an end to spam.
Post offices should be selling digital "stamps" at a nominal price per KB so that we can configure our mailserver not to handle unexpected traffic unless postage has been paid.
A digital stamp is just like a paper stamp, in that it is a hard-to-fake receipt for postage paid.
How will this eliminate spam?
Unsolicited mail without a stamp can be rejected out of hand (there's no reason you couldn't make exemptions for known addresses) which will make zombie farms pointless.
When it is necessary to pay postage to get at the recipient, spam ceases to be an economic proposition.
As has been noted by many others, spammers don't spam because it's fun, they do it for the money. The economic foundation of spam is theft of resource. By ensuring that sender has to pay at least one cent per email we radically alter the cost model, rendering the exercise unprofitable and uneconomic.
Post offices the world over have been suffering an identity crisis ever since email became widespread. This, I believe, is their place in the communications universe.
Before you all pipe up with all the flaws in this plan, I'll save some time by enumerating the issues and the solutions.
Single use, single destination
Stamps have to be single-use for a single destination or spammers will use the same stamp for every target domain.
Single use is easy: the recipient domain should note the serial number of the stamp and allow only one use. Single domain is trickier, and requires the message-id, date and size to be sent to the post-office server so that a custom stamp can be generated on demand.
Everything through one server?! Centralised systems are bottlenecks. Happily stamps will be necessary only for unrecognised senders. Any given organisation will have a fairly static set of correspondant domains and correspondants, and the burden of managing a postage exemption list is far less onerous than the current firehose of spam.
So it isn't everything through one server, it's only unsolicited materials that get bottlenecked. You did want your unsolicited mail slowed down...didn't you?
I think the problem is not the managers themselves but the whole rank culture. Pun intended but serious nonetheless.
There are two issues at play here:
Companies are ultimately about making money, not about playing with tech
CEOs and other managers are administrative flunkies, yet somehow our culture exalts them.
As far as I can see, the problem is that those most qualified to decide are also most qualified to do, and while their attention is focussed on the doing, the foxes declare themselves guardians of the henhouse.
I have a tech startup operation, and it is never going to IPO, because it is never going to fall into the hands of the foxes. I know that I am not good at sales or administration. I will almost certainly hire people to fulfil these roles, but I will also certainly have trouble keeping them in a world where mine is the only company that daily reminds them that they are servants, not rulers. It is not in their interests to put my company's interests ahead of their own. They don't do this even when they are treated as rulers. Neither do IT people, in my experience. We're just quieter in our larceny.
It'll never happen. I'm looking for a new job. I haven't lost the old one yet but I know it will happen because the codebase is an unmaintainable pile of crap due in large part to the principle under discussion.
More than a year ago I warned the (then) new head of development that the product was wonderful as a prototype and garbage as a retail product. I told him that if we didn't treat it as a specification from which to build a real product it would grow exponentially in size and messiness, something like the plant in Little Shop of Horrors.
It has, and only a growing customer base (he's an astoundingly good salesman) has masked the fact that I was right about the cost of customer support becoming unsupportable. Not to mention codebase maintenance taking longer and longer...
Part of the reason that they (PHBs) never learn about this sort of thing is that programmers always say we should start over, and very often attempts at rework fail dismally. Possibly this is due to poor design as described earlier in this thread, but it's nonetheless damaging to the credibility of the idea.
Juan Ramón Jiménez: Spanish poet who won the Nobel Prize for Literature in 1956. He liked to flaunt spelling rules...
No he didn't. He liked to flout spelling rules. To flaunt spelling rules, he would have to adhere to them meticulously and conspicuously, because flaunt means a gratuitous display of something - for example, driving a fancy car in Harlem is flaunting your wealth, or wearing snug yellow spandex over impressive cleavage flaunts another kind of asset.
RFCs are for people who make things. You know, the nerdy people you mock but without whose efforts the cool people would still be living in trees and caves.
The net these days has been fouled by thieves like telecommunication companies and spammers, both of which commit theft of resource without making any real contribution.
Since the defining characteristic of humans is the making of things, it follows that those who do not make things are not human. Therefore, it is ethically acceptable to exterminate them as vermin provided they do not suffer unduly. Given the trouble they've caused, their due suffering gives us quite a bit of latitude.
This is your cue to start your own project or begin contributing to one of the "alpha/beta" ones. Or perhaps you think those of us who write software exist only to provide you with free toys?
"...memorising a set of strategies for coping with the stupidity of other people" sounds to me a lot like a neural net learning social skills. Living memory is very different from simple recording because in neural nets transcoding and recording are inextricably entwined (arguably they are the same operation).
How is this different from "acquiring experience" ?
Joe and Jane Dell can hardly find the power button. If they cannot connect to the net this is a good thing because bandwidth will not be wasted on spamming their friend and relatives with said baby photos. The world is not responsible for protecting them from their cluelessness - that's their own responsibility.
Benefits of cottage industry distribution channels AKA piracy
Do you ever worry that widespread piracy hurts your salary and even your employability?
I do not worry about piracy harming my industry. The impact of software piracy on the livelihoods of professional software developers such as myself is negligible. Less than five percent of global software development budget is for retail software. If anything, piracy represents a net benefit to us due to widespread education via hands-on experience. Better yet this experience is obtained in a context in which the users must think for themselves instead of pestering the helpdesk.
They won't choose to do this unless the alternative is admitting to a crime. A credible case can be made that this lowers the operating costs (printed documentation, media, packaging, warehousing and delivery, and most of all support) of retail software provision, so even in a retail context piracy is quite beneficial.
I should like to point out that with software is like love: the more you give away the more you have. Copying others' software (or music, or book, etc) does not deprive them of the use of it. The "right" to extort monies for the use of ideas is entirely artificial and quite recent, and it is important to remember that when law is used as a weapon it is tyranny, not justice. There is nothing intrinsically right about law. Slaughtering Jews was entirely legal in Nazi Germany.
Closed source software is a recent aberration -- one that can credibly be largely attributed to none other than [drum roll] good ole uncle Bill! This is documented in great and boring detail by Stallman et al in their justifications of the FSF.
In the days before widespread microcomputer availability, it was normal, when a business had software developed or customised, for the source code to be supplied and for that particular incarnation of the software (assuming it to be a derivative work) to become the property of the business that commissioned and paid for it.
So invalidation of copyright law as currently applied to software could reasonably be regarded as a return to normal. It would also be extremely amusing to see this upset the RIAA's grubby little applecart.
Neither you nor the record companies have any moral right to force your way between the music and the people. You are selling a service - distribution - and your complaint is essentially that you want a legally enforced opportunity to charge people your price for a service they can get much cheaper in a free market.
One necessary thing the recording labels do is promotion. However, if they didn't exist then it would be a lot easier for independents to get airplay simply because all would be inpedendent. Oh no, not a genuinely free market!
I've been thinking about the dynamics of music distribution. Most tours make a loss; they are subsidised by the record companies to promote the bands' latest albums. In the absence of record companies, who would organise and promote band tours? One credible possibility is radio stations. Tickets as prizes already happens. Like any other variety of sponsorship, the sponsor profits by reflected glory, and it is cost-effective for radio stations to promote upcoming performances for their localities. It also strikes me that radio stations have studios...
As to the revenue model of a record-company-free industry, it seems to me that the simplest thing to do is encourage the sharing of mediocre quality MP3s and make high quality CD recordings available for online purchase from the band's website, CD arriving by snailmail. This suits the medium very well; the heavy compression necessary for online delivery degrades recording until it cannot compete with a hi-fi recording on CD - never underestimate the bandwidth of a Fedex full of CDs.
Some people will no doubt settle for mediocre recordings rather than pay. This is true but irrelevant: most artists are lucky to see $1 for every CD sold. Most of the loot goes to supply-chain middlemen such as record store owners and record companies. If sales drop by 90% and artists make $5 a CD from online sales then from the artist's point of view profits are up 500%.
Of course the record companies and record store owners won't be too thrilled, but they do not have a right to govern the relationship between a musician and his audience. They are not entitled to my money.
Actually, "sic" is not an acronym for "spelling is correct." This interpretation is merely a mnemonic for people who don't know any Latin.
"Sic" translates to English as "so" or "thus" and in the context in which he uses it is not inappropriate, although it is more usual to annotate reported speech using quotation marks (").
That is a great idea. But I think that with the advent of IPv6 we should, as a community, take the opportunity to take back the network. The concept underpinning the internet is the absence of centralised control. And while we're at it, we need a root certificate issuer that isn't corporate. How about parts of the keys being held in separate countries just to keep it well out of the clutches of any jurisdiction?
>Try running standalones on anything lower than a mhz computer.
I haven't had anything lower than a MHz since the mid eighties.
Quite apart from that, stop blaming your tools, O poor workman. Algorithmic deficiencies contribute far more to program overhead than the overheads of bytecode interpretation.
For example, I once proved this point with a stock implementation of that classic benchmark, the Sieve of Erastosthenes, in C++ and Visual Basic 3.0 (which also runs on a bytecode interpreter).
Naturally, the C++ version was faster for otherwise identical code.
I made a subtle tweak to the algorithm in the VB version, and Lo! it outran the C++ version 12 to 1.
Permit me to bring to your attention that the precepts underpinning the JVM are not original, but have a long history of high performance, high reliability use on mainframes, where the VM is the only way things are done.
The architecture of mainframe CPUs are targeted toward running VMs; x86 and suchlike are not.
To be fair, the overhead incurred by an interpreter is a function of the opcode granularity, which is to say that if you emulate machine code performance will suck, but if each bytecode does quite a lot then overheads will not be significant. For example, PORT (a Cray emulator for PCs) is pretty snappy, because Cray opcodes do things like read 32K block of disk at (track,sector) into this buffer or invert this matrix of floating point numbers and return the determinant.
Java treads an unfortunate middle ground between these extremes. Java opcodes are too fine grained for my taste in interpreters.
Notwithstanding issues of bytecode granularity or the relative merits of VM architecture, Java is a wonderful language for a variety of reasons. The language itself:
Is semantically complete and consistent in semantic level.
Lends itself to layered abstraction.
Lends itself to library development.
Lends itself to library management.
Is well adapted to straightforward implementation of design patterns.
Due to its aptness to library use, there is a proliferation of libraries across the widest range of applications I have ever seen for a single language, and the quality of these libraries is quite astonishing. Unlike (say) C++, where you tend to get dozens of libraries all doing much the same thing equally poorly, with Java you tend to find one or two libraries of very high quality, frequently with source. Moreover, where significant overlap exists it is often due to differences of approach rather than simple redundancy.
These things on their own are enough for me to prefer it as a language.
Again, don't blame your tools. If you don't have any talent, look for another career.
Slashdot Mirror
I have been saying this for years: the post offices of the of the world - not the governments - can totally put an end to spam.
Post offices should be selling digital "stamps" at a nominal price per KB so that we can configure our mailserver not to handle unexpected traffic unless postage has been paid.
A digital stamp is just like a paper stamp, in that it is a hard-to-fake receipt for postage paid.
How will this eliminate spam?
As has been noted by many others, spammers don't spam because it's fun, they do it for the money. The economic foundation of spam is theft of resource. By ensuring that sender has to pay at least one cent per email we radically alter the cost model, rendering the exercise unprofitable and uneconomic.
Post offices the world over have been suffering an identity crisis ever since email became widespread. This, I believe, is their place in the communications universe.
Before you all pipe up with all the flaws in this plan, I'll save some time by enumerating the issues and the solutions.
Single use, single destination
Stamps have to be single-use for a single destination or spammers will use the same stamp for every target domain.
Single use is easy: the recipient domain should note the serial number of the stamp and allow only one use. Single domain is trickier, and requires the message-id, date and size to be sent to the post-office server so that a custom stamp can be generated on demand.
Everything through one server?!
Centralised systems are bottlenecks. Happily stamps will be necessary only for unrecognised senders. Any given organisation will have a fairly static set of correspondant domains and correspondants, and the burden of managing a postage exemption list is far less onerous than the current firehose of spam.
So it isn't everything through one server, it's only unsolicited materials that get bottlenecked. You did want your unsolicited mail slowed down...didn't you?
I think the problem is not the managers themselves but the whole rank culture. Pun intended but serious nonetheless.
There are two issues at play here:
As far as I can see, the problem is that those most qualified to decide are also most qualified to do, and while their attention is focussed on the doing, the foxes declare themselves guardians of the henhouse.
I have a tech startup operation, and it is never going to IPO, because it is never going to fall into the hands of the foxes. I know that I am not good at sales or administration. I will almost certainly hire people to fulfil these roles, but I will also certainly have trouble keeping them in a world where mine is the only company that daily reminds them that they are servants, not rulers. It is not in their interests to put my company's interests ahead of their own. They don't do this even when they are treated as rulers. Neither do IT people, in my experience. We're just quieter in our larceny.
It'll never happen. I'm looking for a new job. I haven't lost the old one yet but I know it will happen because the codebase is an unmaintainable pile of crap due in large part to the principle under discussion.
More than a year ago I warned the (then) new head of development that the product was wonderful as a prototype and garbage as a retail product. I told him that if we didn't treat it as a specification from which to build a real product it would grow exponentially in size and messiness, something like the plant in Little Shop of Horrors.
It has, and only a growing customer base (he's an astoundingly good salesman) has masked the fact that I was right about the cost of customer support becoming unsupportable. Not to mention codebase maintenance taking longer and longer...
Part of the reason that they (PHBs) never learn about this sort of thing is that programmers always say we should start over, and very often attempts at rework fail dismally. Possibly this is due to poor design as described earlier in this thread, but it's nonetheless damaging to the credibility of the idea.
No he didn't. He liked to flout spelling rules. To flaunt spelling rules, he would have to adhere to them meticulously and conspicuously, because flaunt means a gratuitous display of something - for example, driving a fancy car in Harlem is flaunting your wealth, or wearing snug yellow spandex over impressive cleavage flaunts another kind of asset.
RFCs are for people who make things. You know, the nerdy people you mock but without whose efforts the cool people would still be living in trees and caves.
The net these days has been fouled by thieves like telecommunication companies and spammers, both of which commit theft of resource without making any real contribution.
Since the defining characteristic of humans is the making of things, it follows that those who do not make things are not human. Therefore, it is ethically acceptable to exterminate them as vermin provided they do not suffer unduly. Given the trouble they've caused, their due suffering gives us quite a bit of latitude.
This is your cue to start your own project or begin contributing to one of the "alpha/beta" ones. Or perhaps you think those of us who write software exist only to provide you with free toys?
"...memorising a set of strategies for coping with the stupidity of other people" sounds to me a lot like a neural net learning social skills. Living memory is very different from simple recording because in neural nets transcoding and recording are inextricably entwined (arguably they are the same operation).
How is this different from "acquiring experience" ?
Joe and Jane Dell can hardly find the power button. If they cannot connect to the net this is a good thing because bandwidth will not be wasted on spamming their friend and relatives with said baby photos. The world is not responsible for protecting them from their cluelessness - that's their own responsibility.
I do not worry about piracy harming my industry. The impact of software piracy on the livelihoods of professional software developers such as myself is negligible. Less than five percent of global software development budget is for retail software. If anything, piracy represents a net benefit to us due to widespread education via hands-on experience. Better yet this experience is obtained in a context in which the users must think for themselves instead of pestering the helpdesk.
They won't choose to do this unless the alternative is admitting to a crime. A credible case can be made that this lowers the operating costs (printed documentation, media, packaging, warehousing and delivery, and most of all support) of retail software provision, so even in a retail context piracy is quite beneficial.
I should like to point out that with software is like love: the more you give away the more you have. Copying others' software (or music, or book, etc) does not deprive them of the use of it. The "right" to extort monies for the use of ideas is entirely artificial and quite recent, and it is important to remember that when law is used as a weapon it is tyranny, not justice. There is nothing intrinsically right about law. Slaughtering Jews was entirely legal in Nazi Germany.
Closed source software is a recent aberration -- one that can credibly be largely attributed to none other than [drum roll] good ole uncle Bill! This is documented in great and boring detail by Stallman et al in their justifications of the FSF.
In the days before widespread microcomputer availability, it was normal, when a business had software developed or customised, for the source code to be supplied and for that particular incarnation of the software (assuming it to be a derivative work) to become the property of the business that commissioned and paid for it.
So invalidation of copyright law as currently applied to software could reasonably be regarded as a return to normal. It would also be extremely amusing to see this upset the RIAA's grubby little applecart.
Reality check!
I've been thinking about the dynamics of music distribution. Most tours make a loss; they are subsidised by the record companies to promote the bands' latest albums. In the absence of record companies, who would organise and promote band tours? One credible possibility is radio stations. Tickets as prizes already happens. Like any other variety of sponsorship, the sponsor profits by reflected glory, and it is cost-effective for radio stations to promote upcoming performances for their localities. It also strikes me that radio stations have studios...
As to the revenue model of a record-company-free industry, it seems to me that the simplest thing to do is encourage the sharing of mediocre quality MP3s and make high quality CD recordings available for online purchase from the band's website, CD arriving by snailmail. This suits the medium very well; the heavy compression necessary for online delivery degrades recording until it cannot compete with a hi-fi recording on CD - never underestimate the bandwidth of a Fedex full of CDs.
Some people will no doubt settle for mediocre recordings rather than pay. This is true but irrelevant: most artists are lucky to see $1 for every CD sold. Most of the loot goes to supply-chain middlemen such as record store owners and record companies. If sales drop by 90% and artists make $5 a CD from online sales then from the artist's point of view profits are up 500%.
Of course the record companies and record store owners won't be too thrilled, but they do not have a right to govern the relationship between a musician and his audience. They are not entitled to my money.
Actually, "sic" is not an acronym for "spelling is correct." This interpretation is merely a mnemonic for people who don't know any Latin.
"Sic" translates to English as "so" or "thus" and in the context in which he uses it is not inappropriate, although it is more usual to annotate reported speech using quotation marks (").
That is a great idea. But I think that with the advent of IPv6 we should, as a community, take the opportunity to take back the network. The concept underpinning the internet is the absence of centralised control. And while we're at it, we need a root certificate issuer that isn't corporate. How about parts of the keys being held in separate countries just to keep it well out of the clutches of any jurisdiction?
>Try running standalones on anything lower than a mhz computer.
I haven't had anything lower than a MHz since the mid eighties.
Quite apart from that, stop blaming your tools, O poor workman. Algorithmic deficiencies contribute far more to program overhead than the overheads of bytecode interpretation.
For example, I once proved this point with a stock implementation of that classic benchmark, the Sieve of Erastosthenes, in C++ and Visual Basic 3.0 (which also runs on a bytecode interpreter).
Naturally, the C++ version was faster for otherwise identical code.
I made a subtle tweak to the algorithm in the VB version, and Lo! it outran the C++ version 12 to 1.
Permit me to bring to your attention that the precepts underpinning the JVM are not original, but have a long history of high performance, high reliability use on mainframes, where the VM is the only way things are done.
The architecture of mainframe CPUs are targeted toward running VMs; x86 and suchlike are not.
To be fair, the overhead incurred by an interpreter is a function of the opcode granularity, which is to say that if you emulate machine code performance will suck, but if each bytecode does quite a lot then overheads will not be significant. For example, PORT (a Cray emulator for PCs) is pretty snappy, because Cray opcodes do things like read 32K block of disk at (track,sector) into this buffer or invert this matrix of floating point numbers and return the determinant.
Java treads an unfortunate middle ground between these extremes. Java opcodes are too fine grained for my taste in interpreters.
Notwithstanding issues of bytecode granularity or the relative merits of VM architecture, Java is a wonderful language for a variety of reasons. The language itself:
These things on their own are enough for me to prefer it as a language.
Again, don't blame your tools. If you don't have any talent, look for another career.