What you're referring to even has a name: The Osborne effect. After the computer maker Osborne that went bankrupt in the early 80's because people stopped buying their tremendously successful "portable" Osborne 1 machine because the Osborne 2 was supposed to be so incredibly much better... The Osborne 2 never got to market.
You're falling into the trap of assuming that
Slashdot has a homogenous userbase, and that
a representative sample of the userbase read every article. That's simply not how things work. People read (and comment on) what interests them.
Which means you will see a lot of conflicting views to news items. Deal with it - that's how a diverse user community works. We don't all share the same ideals or views.
Your interest rate argument is fatally flawed. Sure, the VCs don't borrow from banks. But VCs do borrow from investors. And when interest rates increase, it becomes more and more important for banks to get access to funds elsewhere, which means interest rates for deposits go up, and more lucrative bonds gets offered by the banks.
When that happens, the least risk taking segment
of the investors pull out of the stock market and put their money in bonds and as deposits instead.
Which means that liquidity in the stock market is reduced. Which means some medium risk investors will pull out of VC funds, and find safer investments, such as heavy established stocks or (if they're spooked enough) bank deposits and bonds.
So while I do agree with you that hysteria (for good reasons) was the major factor, interest rate hikes certainly do affect the investment climate
too.
Yes, but of the SI units, only the kilogram is
defined in terms of a physical artifact instead
of a textual description of a measurable natural
phenomenon.
The reason you can get weights that are correct
to a high degree is that they've been adjusted
to match the a weight that's been adjusted to
match a weight that's been adjusted to match the
standard kilogram... Or something like that - I don't remember how many intermediaries would be
between the standard kilogram and a commercial
weight.
Anyway, the "standard techniques" for determining
the mass involves the standard kilogram at some
point (even for the US, as the US officially defines all it's measurements using SI units).
Has it occured to you that most of your clients,
if they are buying Lotus Notes apps, hardly is
a representative sample?
A company running primarily Linux, or having Linux
in the mix, is less likely to run Notes and require your services than a pure Windows shop is.
I'm not claiming Linux is dominating the small/medium sized companies, because it isn't (at least not yet:), but claiming that Linux isn't making any inroads with that category of companies because a group of Lotus Notes using companies don't have a clue about Linux isn't exactly very credible.
First of all, as others have noted, this is compulsory licensing, which is legal in Brazil.
Second, it can't be theft, because patents are not legally property, but a government granted protection, in the same way that a copyrighted work isn't property, allthough it is still to some extent protected.
The distinction is important. Patents and copyrights are time limited grants of monopoly or licensing rights done by a government to give an incentive for innovation for the public good. If the government decide that a certain license isn't for the public good, there is a very strong argument for revoking that license - and in Brasil the government indeed have a law to back up that right.
This is the essentially the same concept that is behind copyright and patents, and even corporate
charters in the US. Actually, many states in the US
still have chartering laws on the books (some even in their constitutions) granting the legislature the right to detail control the charter granted to any business, and to revoke it if the corporation acts against the public good.
You disregard the very basis of patents and copyright: Neither are property at all. They are limited monopolies granted to encourage innovation for the benefit of the public. If the public doesn't see a benefit (the drugs are too expensive, for instance), there is no reason for
the government to keep honouring the patents.
Read a book on the history of intellectual "property".
The issue is perception. When you watch a television commercial, do you expect the celebrities have been paid to endorse a product?
Sure you do.
When you receive a form letter, do you expect
the sender to have spent lots of time crafting
a letter conveying their emotions, and that
they deeply care about the subject? Of course
not.
When you receive a letter with personal letterheads, and a seemingly unique content, do
you expect that letter to be a letter not written
by the sender, but by a lobbying organization?
Normally not.
It's the last case that upset people about this
campaign.
The difference lies in how it's done. Normally
a political group would either present you with
a form letter, and ask you to send it in yourself,
in which case the recipient sees that you didn't
write the letter in the first place (since they'll
receive lots of copies), and treats it accordinly (it's easy to get people to sign form letters, so
there has to be more of them to mean anything),
or alternatively they ask you to write a letter
yourself, perhaps giving you some ideas to convey.
In this case they appear to have prepared letters, but tried their best to make it appear as if ordinary citizens wrote them, and then used sleaze phone "interviews" to get people to agree to sign them.
The issue here isn't with pro-Microsoft groups getting people to sign letters supporting Microsoft, but with these groups trying to make it
look like this is something people are doing of
their own device because they are angry with how
the case is handled.
There is no shortage of food on a world basis.
Spending money on growing soy would be a waste.
The real problem with hunger in the world stems
from a lack of distribution, and from poverty, not
from a lack of food.
As it is now, the industrialized countries throw
away more than enough food each year due to "overproduction" to be able to cover for the food
shortages in the rest of the world.
So why doesn't it happen? There's no profit in it,
and they're conserned that it might find it's way
back to their respective countries and bring prices down.
I recently let my girlfriend loose on my Linux
box at home. Setting up an account for her was
all I had to do - she's playing around with it
and has had no complaints so far. And she doesn't
even like computers (but goes out with a geek like me - go figure:). Yes, sometimes she does ask
where she kind find some application or other, but
comparing that to the problems the people at work
have getting their Windows boxes to behave, I
was surprised at how quickly she adapted to Linux.
Everytime they give access to a new service, they
should at least inform about rules for accessing
it. And when you're joining the company they
should inform you about any existing policies.
When where you last informed about excessive
monitoring and restrictions on privacy during a
job interview?
Now, we can most of the time update
code without having to update content and vice
versa. If your content model is well designed
you'll have great flexibility.
And yes, the templates indicate where to plug
in contents, but contains no code. That is what
code-content separation is about. Nobody is
claiming that there's no interaction - that would
be useless.
Provided you've done a thorough design job, throwing out the entire design, or replacing a
major portion of the code without changing basic
functionality, you will get away without breaking
things.
Your example is silly, because it has nothing to
do with what code-content separation is intended
to achieve: The ability to do major changes in
a large application without major effects on
content, and the other way around. Noone sensible
will claim that you always get away with no updates in the other, but the better your content
model the fewer changes. If an example as yours
posed problems, we might look for a solution, or
we might simply document it as an exception.
For the webmail system I mentioned, however,
such changes have been exceedingly rare.
When I co-founded my current company, the first
thing we did when we designed our webmail system
was code-content separation, and to some extent
code-content-design separation.
Yes, of course the content, design and code must be used together to build the finished pages.
The point of separation, however, is that you can update one without the other:
- All behaviour in our webmail system was/is isolated in CGIs.
- All design was separated in XML based templates.
If the design changed, the CGIs didn't need to be changed as long as the overall functionality stays the same.
- Most of the content were separated from the design and stored separately.
We didn't do an entirely clean content-design separation, but we certainly did separate code and content. And it irritated me afterwards that we didn't make the content-design separation cleaner
as well, as it would certainly have simplified
updates to the system.
Your example is silly. No separation is needed when the functionality is small. Separation does however become an issue in a system supporting tons of languages, with several web designers working on updating and improving the looks of the site in parallel with a big team of developers
working on adding new functionality while trying
not to screw up the design.
Adding a new design to a system with tens of thousand lines of code if the code and HTML is intermingled is a nightmare. Doing it to a system
where all HTML is separated out into templates may be a few minutes job.
The additional separation of content and design also helps here: Just as intermingling code and design elements is a bad idea, so is intermingling design elements and content (like a news article on a news site, for instance), again because updates to one shouldn't affect the other.
Sure, you have to stay within certain parameters, but a well designed content model will give you great flexibility in changing content and design without disrupting code and the other way around.
Actually, a company I usd to do consulting for
did prepare "hitlists" listing the top ten visited sites that where "bad" according to company policy. It quickly improved things. Perhaps because people got embarrassed even if they weren't identified to anyone.
It's worth trying, and certainly less invasive than monitoring traffic from individual machines.
Why would you not consider being able to access
some non-work related sites as part of the payment?
I wouldn't have a problem with being monitored if
it was spelt out in my employment contract that I would be. If I cared about that, I could then insist on higher pay, or not take the job, if I didn't want to be monitored.
The problem occurs if employers can just start monitoring without informing employees, and without creating the expectation that they will
follow through, so that employees can decide for
themselves whether to accept it or not.
The question isn't whether or not you have the "right" to use the computer that way, but whether or not you are expected to have any privacy at work.
Whether or not society allow corporations to take
away privacy is entirely up to society.
The right to incorporate is no inherent natural
right - it is a privilege granted by the people
via their governments. It originated as royal
orders to create a "virtual person" with rights
defined in a charter. Even today most corporations
are governed by a charter, but the charter granted
is usually much wider.
In most states in the US laws have been or is on the book to allow the legislatures
in the respective states to dictate the contents
of the charter on a case by case basis, to limit
the time the charter is granted for (a charter used to be time limited), and even to dissolve
the corporation if it is decided that it does not
serve the public.
Revoking charters used to be common if a corporation was seen as abusing the powers granted
to it by the people, and restricting privacy for
its employers could quite well have been considered as abusing its powers.
A corporation in the US can still in many states be argued to have no "rights" other than what is
granted to it by the legislature. Some may say
its unfortunate that its now uncommon for the
legislatures to write, rewrite or revoke charters
on case by case basis...
Uhm. Obviously "optimum compression" would be
an optimal program for calculating Pi, which
I'd hope would be a fair bit smaller than the
size you mentioned:-)
Rephrase it "optimum compression without using
knowledge about specific inputs", and I'd be with you...
It has the slight little weakness that you need
to transmit the key in some form. Which means that the attacker can reasonably assume that the
starting point won't be too high. If the attacker
knows *anything* about the content being sent,
the mechanism is also vulnerable to the attacker
doing guesses about what to expect, and use that
to verify whether or not he's achieving success.
For instance, if I were looking for JPEG or Word
files, I could look for common headers by:
1) Making an assumption about the maxium size of
the key (the starting point in the Pi sequence),
and 2) Take X bytes of known header information from each of a set of file formats, 3) Find every
starting point in the Pi sequence that when XOR'd with the characters in the known headers produced the same data as in the start of your encrypted file, 4) proceed to see if I get valid data from continuing onwards from the starting points identified.
Even if it's an unstructured file format like plain text, I still believe that you'd be able to do "interesting" analysis of the file based on
statistical facts about plain text - the occurence of characters have clear statistical properties and patterns in different languages.
Also the
range of characters used is often very limited -
simply decoding a few characters with every sequence, and see whether it would give you any characters outside the expected set would let you throw away a huge number of starting points very quickly.
And since XOR'ing a non-random sequence with a "random" sequence of digits gives the result non-random properties. XOR'ing it again with another random sequence may obscure that (but it also poses a certain chance of actually totally reversing your initial operation), but not totally, and it may also weaken your initial operation for whole or parts of the file.
Actually in that respect, you could have taken any
good pseudo random number generator, and done almost the same thing as you suggest: The "starting point" is essentially just a seed.
I'm not a cryptographer (not even on hobby basis) nor a mathematician (haven't touched a math book since my first term in University, thank you). When there's weaknesses obvious enough that I can see it like that, then I'm sure that someone who actually knows anything about cryptography can find a lot more problems with it.
It's "random" in the sense that if you don't
resort to formulas or knowledge of Pi in particular you'd have the same likelyhood of
guessing or calculating the next digit of Pi
based purely on the previous digits as you would
in guessing or calculating the next digit in a
sequence of dice throws with a 10-sided dice.
In effect, you can't predict it without knowledge
beyond the previous digits themselves (such external knowledge can in Pi's
case for instance be it's relationship to the
geometric properties of circles).
Slashdot Mirror
There is even find a paper about the problem on the web.
I think you'll find that Steve Jobs knows how to market things to avoid that problem...
You're falling into the trap of assuming that Slashdot has a homogenous userbase, and that a representative sample of the userbase read every article. That's simply not how things work. People read (and comment on) what interests them. Which means you will see a lot of conflicting views to news items. Deal with it - that's how a diverse user community works. We don't all share the same ideals or views.
When that happens, the least risk taking segment of the investors pull out of the stock market and put their money in bonds and as deposits instead.
Which means that liquidity in the stock market is reduced. Which means some medium risk investors will pull out of VC funds, and find safer investments, such as heavy established stocks or (if they're spooked enough) bank deposits and bonds.
So while I do agree with you that hysteria (for good reasons) was the major factor, interest rate hikes certainly do affect the investment climate too.
Yes, but of the SI units, only the kilogram is defined in terms of a physical artifact instead of a textual description of a measurable natural phenomenon.
"The metre is the length of the path travelled by light in vacuum during a time interval of 1/299 792 458 of a second" (17th CGPM, 1983, Resolution 1)
While the kilogram is still defined as:
"The kilogram is the unit of mass: it is equal to the mass of the international prototype of the kilogram" 3rd CGPM (1901)
to a high degree is that they've been adjusted
to match the a weight that's been adjusted to
match a weight that's been adjusted to match the
standard kilogram... Or something like that - I don't remember how many intermediaries would be
between the standard kilogram and a commercial
weight.
Anyway, the "standard techniques" for determining
the mass involves the standard kilogram at some
point (even for the US, as the US officially defines all it's measurements using SI units).
A company running primarily Linux, or having Linux in the mix, is less likely to run Notes and require your services than a pure Windows shop is.
I'm not claiming Linux is dominating the small/medium sized companies, because it isn't (at least not yet :), but claiming that Linux isn't making any inroads with that category of companies because a group of Lotus Notes using companies don't have a clue about Linux isn't exactly very credible.
That is, you can't find a way to prove that any program P will halt, but you can find infinite P for which you can decide whether it will halt or not.
The distinction is important. Patents and copyrights are time limited grants of monopoly or licensing rights done by a government to give an incentive for innovation for the public good. If the government decide that a certain license isn't for the public good, there is a very strong argument for revoking that license - and in Brasil the government indeed have a law to back up that right.
This is the essentially the same concept that is behind copyright and patents, and even corporate charters in the US. Actually, many states in the US still have chartering laws on the books (some even in their constitutions) granting the legislature the right to detail control the charter granted to any business, and to revoke it if the corporation acts against the public good.
Too bad none of them use those rights these days.
the government to keep honouring the patents.
Read a book on the history of intellectual "property".
When you receive a form letter, do you expect the sender to have spent lots of time crafting a letter conveying their emotions, and that they deeply care about the subject? Of course not.
When you receive a letter with personal letterheads, and a seemingly unique content, do you expect that letter to be a letter not written by the sender, but by a lobbying organization? Normally not.
It's the last case that upset people about this campaign.
a political group would either present you with
a form letter, and ask you to send it in yourself,
in which case the recipient sees that you didn't
write the letter in the first place (since they'll
receive lots of copies), and treats it accordinly (it's easy to get people to sign form letters, so
there has to be more of them to mean anything),
or alternatively they ask you to write a letter
yourself, perhaps giving you some ideas to convey.
In this case they appear to have prepared letters, but tried their best to make it appear as if ordinary citizens wrote them, and then used sleaze phone "interviews" to get people to agree to sign them.
The issue here isn't with pro-Microsoft groups getting people to sign letters supporting Microsoft, but with these groups trying to make it
look like this is something people are doing of
their own device because they are angry with how
the case is handled.
That is deceptive at best...
those sentences are two of the sentences duplicated in a lot of letters.
In other words: It's a joke, dammit.
As it is now, the industrialized countries throw away more than enough food each year due to "overproduction" to be able to cover for the food shortages in the rest of the world.
So why doesn't it happen? There's no profit in it, and they're conserned that it might find it's way back to their respective countries and bring prices down.
I recently let my girlfriend loose on my Linux box at home. Setting up an account for her was all I had to do - she's playing around with it and has had no complaints so far. And she doesn't even like computers (but goes out with a geek like me - go figure :). Yes, sometimes she does ask
where she kind find some application or other, but
comparing that to the problems the people at work
have getting their Windows boxes to behave, I
was surprised at how quickly she adapted to Linux.
Everytime they give access to a new service, they should at least inform about rules for accessing it. And when you're joining the company they should inform you about any existing policies. When where you last informed about excessive monitoring and restrictions on privacy during a job interview?
And yes, the templates indicate where to plug in contents, but contains no code. That is what code-content separation is about. Nobody is claiming that there's no interaction - that would be useless.
Provided you've done a thorough design job, throwing out the entire design, or replacing a major portion of the code without changing basic functionality, you will get away without breaking things.
Your example is silly, because it has nothing to do with what code-content separation is intended to achieve: The ability to do major changes in a large application without major effects on content, and the other way around. Noone sensible will claim that you always get away with no updates in the other, but the better your content model the fewer changes. If an example as yours posed problems, we might look for a solution, or we might simply document it as an exception.
For the webmail system I mentioned, however, such changes have been exceedingly rare.
When I co-founded my current company, the first thing we did when we designed our webmail system was code-content separation, and to some extent code-content-design separation.
Yes, of course the content, design and code must be used together to build the finished pages.
The point of separation, however, is that you can update one without the other:
- All behaviour in our webmail system was/is isolated in CGIs.
- All design was separated in XML based templates. If the design changed, the CGIs didn't need to be changed as long as the overall functionality stays the same.
- Most of the content were separated from the design and stored separately.
We didn't do an entirely clean content-design separation, but we certainly did separate code and content. And it irritated me afterwards that we didn't make the content-design separation cleaner as well, as it would certainly have simplified updates to the system.
Your example is silly. No separation is needed when the functionality is small. Separation does however become an issue in a system supporting tons of languages, with several web designers working on updating and improving the looks of the site in parallel with a big team of developers working on adding new functionality while trying not to screw up the design.
Adding a new design to a system with tens of thousand lines of code if the code and HTML is intermingled is a nightmare. Doing it to a system where all HTML is separated out into templates may be a few minutes job.
The additional separation of content and design also helps here: Just as intermingling code and design elements is a bad idea, so is intermingling design elements and content (like a news article on a news site, for instance), again because updates to one shouldn't affect the other.
Sure, you have to stay within certain parameters, but a well designed content model will give you great flexibility in changing content and design without disrupting code and the other way around.
It's worth trying, and certainly less invasive than monitoring traffic from individual machines.
I wouldn't have a problem with being monitored if it was spelt out in my employment contract that I would be. If I cared about that, I could then insist on higher pay, or not take the job, if I didn't want to be monitored.
The problem occurs if employers can just start monitoring without informing employees, and without creating the expectation that they will follow through, so that employees can decide for themselves whether to accept it or not.
Whether or not society allow corporations to take away privacy is entirely up to society.
The right to incorporate is no inherent natural right - it is a privilege granted by the people via their governments. It originated as royal orders to create a "virtual person" with rights defined in a charter. Even today most corporations are governed by a charter, but the charter granted is usually much wider.
In most states in the US laws have been or is on the book to allow the legislatures in the respective states to dictate the contents of the charter on a case by case basis, to limit the time the charter is granted for (a charter used to be time limited), and even to dissolve the corporation if it is decided that it does not serve the public.
Revoking charters used to be common if a corporation was seen as abusing the powers granted to it by the people, and restricting privacy for its employers could quite well have been considered as abusing its powers.
A corporation in the US can still in many states be argued to have no "rights" other than what is granted to it by the legislature. Some may say its unfortunate that its now uncommon for the legislatures to write, rewrite or revoke charters on case by case basis...
(Information about charter revocations in the US can be found here)
I'm using that approach even in most non-web applications I write these days.
Rephrase it "optimum compression without using knowledge about specific inputs", and I'd be with you...
--
Remove Trash+ to reach my actual inbox
For instance, if I were looking for JPEG or Word files, I could look for common headers by:
1) Making an assumption about the maxium size of the key (the starting point in the Pi sequence), and 2) Take X bytes of known header information from each of a set of file formats, 3) Find every starting point in the Pi sequence that when XOR'd with the characters in the known headers produced the same data as in the start of your encrypted file, 4) proceed to see if I get valid data from continuing onwards from the starting points identified.
Even if it's an unstructured file format like plain text, I still believe that you'd be able to do "interesting" analysis of the file based on statistical facts about plain text - the occurence of characters have clear statistical properties and patterns in different languages.
Also the range of characters used is often very limited - simply decoding a few characters with every sequence, and see whether it would give you any characters outside the expected set would let you throw away a huge number of starting points very quickly.
And since XOR'ing a non-random sequence with a "random" sequence of digits gives the result non-random properties. XOR'ing it again with another random sequence may obscure that (but it also poses a certain chance of actually totally reversing your initial operation), but not totally, and it may also weaken your initial operation for whole or parts of the file.
Actually in that respect, you could have taken any good pseudo random number generator, and done almost the same thing as you suggest: The "starting point" is essentially just a seed.
I'm not a cryptographer (not even on hobby basis) nor a mathematician (haven't touched a math book since my first term in University, thank you). When there's weaknesses obvious enough that I can see it like that, then I'm sure that someone who actually knows anything about cryptography can find a lot more problems with it.
--
Remove Trash+ to reach my actual inbox
In effect, you can't predict it without knowledge beyond the previous digits themselves (such external knowledge can in Pi's case for instance be it's relationship to the geometric properties of circles).
--
Remove Trash+ to reach my actual inbox