In the final paragraphs of your comment, you seem to disapprove of managing of Senate rules, which obviously carries over to the filibuster. I should point out the filibuster as it stands today is not some wholly virtuous element of the Senate that has been around forever. It was never even used until 1837, and there was no way to overcome it with cloture (the 60 votes needed today). One senator could hold up the Senate's business forever. Cloture didn't exist until 1917, and even then it needed two-thirds of the Senate to invoke. It wasn't until 1949 that it became three-fifths of the Senate.
The value of the filibuster is tenuous, and it certainly should not be held up as a glorified, set-in-stone system.
The USPS is in debt up to its eyeballs because electronic documents are causing a drop in volume. FedEx and USPS have adjusted their rates , fleets and staffing to accommodate this drop.
Searching Google for "usps prepay health benefits" turns up plenty of hits as well. That story doesn't resonate as much with the reading public as much as the stories of how "the internet is changing everything", though.
I would think that another strong effect of taking the website down is that any connections plants had with the honeypot participants would be lessened.
I think you are assuming a situation similar to the US, where people access websites from personally owned computers with static IPs. It is quite possible that the users of these websites were using internet cafes or other throw-away connections.
If we know all their information, can we just follow them to their new site? I'm sure the CIA had operatives planted in the website who befriended some of the regular visitors.
The costs associated with a shake-up can be expensive. Assuming we had plants, the value of their established relationships would reasonably lessen as a result of the transition to a new site, due to suspicion among members.
Taking down this website only made us lose the potential capability to identify and infiltrate other extremist websites that are growing in popularity and membership.
I'm not too sure why you say "only made us lose..."; we have no idea the value of the intelligence gathered that may have been specific to that site.
After doing some research and looking at the manpage for openntpd 3.9, the latest portable release, the manpage does not have the documentation you're referring to, leading me to believe that disciplining is only in later versions that only are applicable to OpenBSD. If having any discussion about the benefits of OpenNTPD it should be made clear that disciplining (a feature I consider crucial) is only available on OpenBSD.
Fair enough that the situation has improved. Do you know if portable OpenNTP has this functionality? I've read in multiple places that only non-portable version has this functionality; only 4.x and above has disciplining, and portable has been at at 3.9 for since 2006.
I still find it disingenuous that OpenNTP uses the NTP name but does not go to any lengths to indicate what they don't support.
Any way one goes about it, I find little reason to look at openntp in contrast to chrony, which is just as simple to setup as OpenNTP, more accurate than NTP, more feature-rich than OpenNTP, and has been around a lot longer.
OpenNTPD does not account for hardware drift, which is what I attempted to describe in my second post. Multiple hits on google for "openntpd hardware drift" support this. Unfortunately the OpenNTPD docs do not say what they don't do with regards to NTPD or chrony, so you don't know what you are missing. Without clock disciplining, all it's really doing is setting the time.
The comment about clock disciplining (compensation for systematic skew or drift) is a fair point, within limits.
From their design goals http://www.openntpd.org/goals.html they are not trying to be as accurate as NTP, which they give as a response to claims of not being as accurate as NTP in the OpenBSD manual (http://www.openbsd.org/faq/faq6.html#OpenNTPDaccurate)
Reach a reasonable accuracy. We are not after the last microseconds.
As someone else has pointed out, given the lack of features that OpenNTPD provides, calling it "NTP" is misleading; it's like calling "telnet" "ssh". It's a step backwards in terms of functionality and accuracy, especially since they don't document up-front what they don't support with regards to what is considered standard.
My mistake on the use of the term slew; what I meant to get across is that it doesn't do any of the clock-slowing stuff that NTP or chrony does. All it did was get a packet and set the time. I'm pretty sure it didn't do any backing off or the like. I can't find any good references at the moment, but it was jaw-dropping inappropriate, especially for a situation where we have to keep all of our servers within a dozen microseconds or so.
After looking into a replacement for NTPD, OpenNTPD was a terrible option. If I recall correctly, all it did was a very simplistic setting of the time from what the server says. No slewing, no safety mechanisms, etc. I remember reading that it was simply designed for simplicity, not features, but it went way overboard.
What, specifically, prevents tabs from (and I quote) "being able to relay your picture of the code to another"? For that matter, what exactly is "the picture of the code"? I've never heard of that term before.
The "picture of the code" includes everything with the textual layout of the code, everything visual. Tabs, being non-printable characters, can be rendered in different ways from the author's intention, creating the potential for loss in transferable meaning. Tabs have semantics of "increase the indent level", while spaces are non-semantic pixels that the author is, in a sense, "painting code". How well the author "paints" adds or reduces value.
Frankly, from a logical perspective, I've never figured out what benefit spaces have over tabs...
While I understand your argument for tab characters, I have moved to the opinion that given our current editor generation, printable characters only (e.g., spaces only) are the best option. Part of coding is the presentation: being able to relay your picture of the code to another has significant value. With control characters in the code, the presentation value can be lost. Using only printable characters eliminates this issue.
The argument for control characters such as tabs really should be extended all the way to going XML-style, where the code is completely re-parsed and presented to the author in a completely different way from its underlying structure. This is a reasonable approach, but our current generation of editors (at least the ones I've been exposed to on linux) are not at that level yet.
One big difference between having a sovereign and a corporation do the task is that a corporation can much more easily fold if there is a problem. In other words, a corporation has much less to lose.
As consumers, we might hate it, but you have to think of it this way... a gift card is an outstanding debt. A business doesn't want to have thousands or tens of thousands of tiny outstanding debts floating around FOREVER. That is the main reason there are "fees" to reduce the value of the card to zero when it isn't used.
I agree. I can't imagine that there would be an organization (let's call it a "knab") that if you deposited money with them, got something in return, they could manage these outstanding liabilities that you could redeem for the product at any time in the future, near or distant. In the meantime, this fictitious knabs would be free to invest your deposit safely until you withdrew it. Knabs would have a terrible time trying to keep track of all these accounts on their books, and couldn't possibly make money, so much so that I can't imagine a world with a knab.
Also, what's so hard about keeping track of all these inactive accounts? It's not like they have many businesses have a hand-written ledger that they have to re-copy all account values around. Since all the accounts are likely similar, automated processing should be able to handle the number, whether it's processing 100 or 10,000.
...your local machine will cache frequently accessed sites anyway...
You need to be more clear about how this caching might actually take place; there is no magical program that would do this...except for a DNS server. On Linux you could be talking about nscd, but this doesn't necessarily abide by the DNS caching protocol correctly.
I can also recommend Slashdotter as a great extension for Firefox. I still use version 1 Slashdot comment style, and the ability to dynamically load sub-threshold comments is very handy.
You can't mandate anything in an open market, and I just find the possible motivation for that statement bizarre.
While I agree with most of your comment, I have to state that there can be good motivations for mandating some things, as there are economic functions which are not always "smooth" or require the equivalent of a force to overcome "static friction", which is greater than "dynamic friction". In other words, the overall well-being of the economic system can sometimes be improved by providing subsidies to overcome initial investment hurdles. Open markets are good at improving situations if the necessary investment required to improve upon the existing is low, but less adept at changes requiring large investments that are not designed to benefit one market participant but the whole market.
The certificate system is an unfortunate aspect of SSL; ssh has enormous popularity without demanding a rigid chain-of-authority system, relying instead on cached host keys instead. It's a "good enough" solution that balances the needs and simplicity well.
It's important to remember that one big advantage that web applications have over traditional desktop applications is that they connect and integrate with the entire network of the web better. Traditional applications tend to behave a lot like flash applications do: they work well on their own, but tend to be closed in terms of integration with their environment. Going forward, integrating with all other applications and information available on the internet will continue to be a key driving force.
Well, to be fair, "punitive" taxes, as you call them, are specifically designed to "distort" the market... in the sense that they *un-distort* the market by forcing the cost of negative externalities back onto the consumer.
Or: Carbon taxes are a *perfect* example of government taxation to *correct* market inefficiencies.
The decision mentions that the "test" for being allowed to identify a poster that some places have applied is usually something close to the plaintiff needing to at last prove that he can survive a summary motion to dismiss. If you read the decision, they talk quite a bit about the approaches several different states have taken.
Slashdot Mirror
Flexible spending accounts are a different beast than HSA accounts. FSA accounts are use-it-or-lose it, HSA's are not.
In the final paragraphs of your comment, you seem to disapprove of managing of Senate rules, which obviously carries over to the filibuster. I should point out the filibuster as it stands today is not some wholly virtuous element of the Senate that has been around forever. It was never even used until 1837, and there was no way to overcome it with cloture (the 60 votes needed today). One senator could hold up the Senate's business forever. Cloture didn't exist until 1917, and even then it needed two-thirds of the Senate to invoke. It wasn't until 1949 that it became three-fifths of the Senate.
The value of the filibuster is tenuous, and it certainly should not be held up as a glorified, set-in-stone system.
Additionally, one of the main USPS's main source of debt is that it has to prepay its health benefits premiums, something no other public or private entity does:
http://www.usps.com/communications/newsroom/testimony/2009/pr09_pmg0128.htm
Searching Google for "usps prepay health benefits" turns up plenty of hits as well. That story doesn't resonate as much with the reading public as much as the stories of how "the internet is changing everything", though.
I would think that another strong effect of taking the website down is that any connections plants had with the honeypot participants would be lessened.
I think you are assuming a situation similar to the US, where people access websites from personally owned computers with static IPs. It is quite possible that the users of these websites were using internet cafes or other throw-away connections.
The costs associated with a shake-up can be expensive. Assuming we had plants, the value of their established relationships would reasonably lessen as a result of the transition to a new site, due to suspicion among members.
I'm not too sure why you say "only made us lose..."; we have no idea the value of the intelligence gathered that may have been specific to that site.
After doing some research and looking at the manpage for openntpd 3.9, the latest portable release, the manpage does not have the documentation you're referring to, leading me to believe that disciplining is only in later versions that only are applicable to OpenBSD. If having any discussion about the benefits of OpenNTPD it should be made clear that disciplining (a feature I consider crucial) is only available on OpenBSD.
Fair enough that the situation has improved. Do you know if portable OpenNTP has this functionality? I've read in multiple places that only non-portable version has this functionality; only 4.x and above has disciplining, and portable has been at at 3.9 for since 2006.
I still find it disingenuous that OpenNTP uses the NTP name but does not go to any lengths to indicate what they don't support.
Any way one goes about it, I find little reason to look at openntp in contrast to chrony, which is just as simple to setup as OpenNTP, more accurate than NTP, more feature-rich than OpenNTP, and has been around a lot longer.
OpenNTPD does not account for hardware drift, which is what I attempted to describe in my second post. Multiple hits on google for "openntpd hardware drift" support this. Unfortunately the OpenNTPD docs do not say what they don't do with regards to NTPD or chrony, so you don't know what you are missing. Without clock disciplining, all it's really doing is setting the time.
From http://www.advogato.org/person/dtucker/diary.html?start=52
From their design goals http://www.openntpd.org/goals.html they are not trying to be as accurate as NTP, which they give as a response to claims of not being as accurate as NTP in the OpenBSD manual (http://www.openbsd.org/faq/faq6.html#OpenNTPDaccurate)
As someone else has pointed out, given the lack of features that OpenNTPD provides, calling it "NTP" is misleading; it's like calling "telnet" "ssh". It's a step backwards in terms of functionality and accuracy, especially since they don't document up-front what they don't support with regards to what is considered standard.
My mistake on the use of the term slew; what I meant to get across is that it doesn't do any of the clock-slowing stuff that NTP or chrony does. All it did was get a packet and set the time. I'm pretty sure it didn't do any backing off or the like. I can't find any good references at the moment, but it was jaw-dropping inappropriate, especially for a situation where we have to keep all of our servers within a dozen microseconds or so.
After looking into a replacement for NTPD, OpenNTPD was a terrible option. If I recall correctly, all it did was a very simplistic setting of the time from what the server says. No slewing, no safety mechanisms, etc. I remember reading that it was simply designed for simplicity, not features, but it went way overboard.
While that is true in many cases, when one drive dies you will get much better read and write performance out of RAID 1 than RAID 0.
The "picture of the code" includes everything with the textual layout of the code, everything visual. Tabs, being non-printable characters, can be rendered in different ways from the author's intention, creating the potential for loss in transferable meaning. Tabs have semantics of "increase the indent level", while spaces are non-semantic pixels that the author is, in a sense, "painting code". How well the author "paints" adds or reduces value.
While I understand your argument for tab characters, I have moved to the opinion that given our current editor generation, printable characters only (e.g., spaces only) are the best option. Part of coding is the presentation: being able to relay your picture of the code to another has significant value. With control characters in the code, the presentation value can be lost. Using only printable characters eliminates this issue.
The argument for control characters such as tabs really should be extended all the way to going XML-style, where the code is completely re-parsed and presented to the author in a completely different way from its underlying structure. This is a reasonable approach, but our current generation of editors (at least the ones I've been exposed to on linux) are not at that level yet.
One big difference between having a sovereign and a corporation do the task is that a corporation can much more easily fold if there is a problem. In other words, a corporation has much less to lose.
I agree. I can't imagine that there would be an organization (let's call it a "knab") that if you deposited money with them, got something in return, they could manage these outstanding liabilities that you could redeem for the product at any time in the future, near or distant. In the meantime, this fictitious knabs would be free to invest your deposit safely until you withdrew it. Knabs would have a terrible time trying to keep track of all these accounts on their books, and couldn't possibly make money, so much so that I can't imagine a world with a knab.
Also, what's so hard about keeping track of all these inactive accounts? It's not like they have many businesses have a hand-written ledger that they have to re-copy all account values around. Since all the accounts are likely similar, automated processing should be able to handle the number, whether it's processing 100 or 10,000.
When it is dark out, I can certainly see it being difficult to determine where the unlit lights are.
You need to be more clear about how this caching might actually take place; there is no magical program that would do this...except for a DNS server. On Linux you could be talking about nscd, but this doesn't necessarily abide by the DNS caching protocol correctly.
Interestingly, if you have flash disabled (e.g., via NoScript), a JPG does show up instead of the flash image.
I can also recommend Slashdotter as a great extension for Firefox. I still use version 1 Slashdot comment style, and the ability to dynamically load sub-threshold comments is very handy.
While I agree with most of your comment, I have to state that there can be good motivations for mandating some things, as there are economic functions which are not always "smooth" or require the equivalent of a force to overcome "static friction", which is greater than "dynamic friction". In other words, the overall well-being of the economic system can sometimes be improved by providing subsidies to overcome initial investment hurdles. Open markets are good at improving situations if the necessary investment required to improve upon the existing is low, but less adept at changes requiring large investments that are not designed to benefit one market participant but the whole market.
The certificate system is an unfortunate aspect of SSL; ssh has enormous popularity without demanding a rigid chain-of-authority system, relying instead on cached host keys instead. It's a "good enough" solution that balances the needs and simplicity well.
It's important to remember that one big advantage that web applications have over traditional desktop applications is that they connect and integrate with the entire network of the web better. Traditional applications tend to behave a lot like flash applications do: they work well on their own, but tend to be closed in terms of integration with their environment. Going forward, integrating with all other applications and information available on the internet will continue to be a key driving force.
You are spot on.
The decision mentions that the "test" for being allowed to identify a poster that some places have applied is usually something close to the plaintiff needing to at last prove that he can survive a summary motion to dismiss. If you read the decision, they talk quite a bit about the approaches several different states have taken.