Re:Green Software + Hardware
on
Building a Green PC
·
· Score: 2, Informative
Switched-mode power supplies (as found in most 'big' computing equipment like PCs, monitors, printers etc) are pretty energy efficient, and can use almost no power in 'standby' mode.
Having one big DC supply with lots of outlets in your home is likely to be less energy efficient (because of the large voltage loss over long cable runs at high currents) than having local ones.
That's not to say that having one SMPS per PC, with DC outlets on the back of the PC for monitor, printer etc wouldn't be a good idea - in fact I'd have thought that would be an excellent idea. (As would having DC outlets on the back of your TV for the DVD, TiVo, XBox etc).
The problem might be that there's a patent on this (http://www.freepatentsonline.com/5768600.html):(
That's the only way it could work. How else do you block illegal downloads of music/films etc, without blocking ALL downloads of music/films etc. How do you know that what someone is downloading is music/film, especially if it's done using P2P when it's sent in blocks? So, you have to block anything which could be P2P. But, P2P is used for legal stuff as well. People might be downloading stuff illegally from https:/// websites, so you obviously have to block all those as well.
It just can't work.
The ISPs ought to do organise a 'trial' of the system between themselves, using MPs' accounts and data from.gov.uk websites as test data. So, every packet needs to/from those sites needs to be inspected (possibly manually), and anything which might possibly be copyrighted and where there's no evidence of a licence, should be blocked (ie pretty much everything except for SYN and ACK packets..)
So, every Internet user would have to have a certificate (when you go to a https:/// website, YOUR web browser is encrypting the data). You wouldn't be able to access secure web sites/VPNs/etc from outside the country etc.
If you need more bandwidth, what's wrong with the ISPs expecting you to pay more than someone who uses less?
If you ran a taxi service and someone wanted 20 people to be taken from A to B, you wouldn't just charge them for 1 car's worth of people, you'd charge then for however many cars it took.
I was mainly talking about in the UK (since that's most affected by the BBC iPlayer). Other countries may not publish the contention ratios, but in the UK it's pretty standard.
BTW, my experience of SAPO.PT is that it's pretty poor compared to all the UK ISP's I've been involved with. Unfortunately, the other Portuguese ISPs seem to be pretty poor too.
An average 8Mbps ADSL line in the UK is 8Mbps with a 50:1 contention ratio.
Really, that's a 160kbps line with 8Mbps burst capability. Unfortunately, most people don't understand that. You will usually get very long 'bursts' of 8Mbps, and may never notice it dropping, but as long as your available bandwidth doesn't go below 160kbps, you're getting what you've paid for. Maybe they're mis-sold, or maybe they're just sold in a way that the average user has a slight chance of understanding.
If you want 8Mbps all the time, talk to your ISP about getting an E2 leased line... Then when you see the cost, realise you're actually getting quite a good deal!
With all these things, as long as you can find out beforehand what the real deal is, then everything is OK. A 5Mbps connection with a 2GB 20Mbps burst allowance sounds quite good to me..
ISTM that the BBC should cohost servers at the big ISPs' data centres. Users would stream data from the servers at their ISP. Network traffic from a server in the data centre to the same ISP's customers should be really cheap (free?) as there's no 'Internet' traffic, so it wouldn't be an issue. The only network traffic outside the ISP would be for the cohosted servers to get the data from the BBC.
AFAICS, this would work, at least for the big ISPs, as the BBC wouldn't have to buy many more servers (most of their load is going to be coming from Tiscali, BT and Thus customers anyway, so hosting servers at those ISPs would mean that fewer servers would be needed at the BBC or the one ISP where they currently cohost them) and would need to pay for less Internet bandwidth themselves. They would have to pay for cohosting, but if the servers were ONLY for use by that ISP's customers, I would have thought they could work out a deal where that cost would be minimal.
The even BETTER solution would be to have streaming servers at the exchanges which can have the data trickle fed to them using unused bandwidth, before making the videos available to the users, but that would require some coordinated effort by all the different streaming providers. IMV, if we're going to have true video on demand in the future, that's what will be needed.
All ISPs give DSL lines with a contention ratio. That means that several users share an certain amount of bandwidth. So, a typical 'home user' contention ratio may be 50:1. So, there are 50 users sharing a 'block' of bandwidth. This is well known, and typically described when you are buying a DSL connection.
Usually that's not an issue, most of the users won't be using the internet at the same time, and most of the ones who are will be doing low bandwidth things like browsing websites or downloading email. The issue comes when large numbers start doing high bandwidth things such as streaming or file sharing.
If you aren't happy with that, you could buy a business DSL connection, which may have a 20:1 contention ratio (that's what I do). It costs about 3 times more than a home user connection, but I potentially get 3 times more bandwidth (as well as no 'fair use' limits)
I suppose if people wanted it, ISPs could provide a 1:1 contention ratio, but how many people would be willing to pay 50 times more than a home user connection for that? You could get a T1 leased line for that sort of price.
So, you do get what you pay for, and as long as ISPs honestly publish their contention ratios, it's your own fault if you don't get what you wanted.
Contention ratios are generally a good idea to give people fast internet connections at a lower cost. If ISPs do traffic prioritisation fairly (eg the first 1/50th of the bandwidth block that you use has a higher priority than any excess) then I don't see that anyone has any right to complain (ISPs or customers). This doesn't need to be against net neutrality, you don't prioritise based on the type of data, just the amount.
Usage limits are another issue, and I *do* get annoyed when you get adverts like 'Unlimited Internet for £15 a month' and then in tiny writing "(fair use limits of 20 GB per month apply)". These should be cracked down on IMHO, but contention ratios are sensible and fair.
You're right - 'recompile the kernel' is at least one reason I could never recommend Linux to anyone who's not a true geek. I'm pretty technical, but when I was trying to install a TV card on a Linux PC (for MythTV) and saw 'recompile the kernel' in the instructions, I thought, "WTF? In this day and age, I have to recompile the kernel? How backwards is that?" My 6809-OS9 kernel back in 1983 didn't need recompiling for any device drivers, why the f should Linux "the superest & bestest OS ever" need to have its kernel recompiled by anyone other than someone who's maintaining it?
OS-9 was written 12 years before Linux and they had true dynamically loaded drivers. Windows manages it reasonably well as well, but Linux can't do something as 'simple' as that. No driver should be harder to install than something like 'copy these files there, and change that configuration file like this'. Having to recompile the kernel just to install a device driver makes the whole thing look like it's still in alpha testing stage!
Maybe there's a good reason for it (but I'm not sure I can think what it could be), but that still doesn't make it friendly for 99% of potential users.
That's just what we were saying here. For all I know, a cable might be being cut every couple of months on average, but they repair it within a week, and use the redundant cables in the meantime. No one would notice. But, if several cables go at the same time, it becomes a big problem. If cable faults were that common, then 3 or 4 (or maybe even 5) could well just be coincidence.
Wait until they repair them and see if they say anything, and what. It should be easy to tell whether they were caught on an anchor, bombed or cut (but whether they were caught on an anchor accidentally or deliberately would be harder to tell).
If the chance of 4 cables being taken offline is 1 in 100,000,000 a week, that means that once in the next 100,000,000 weeks, the probability is that 4 cables will be offline simultaneously. Yes?
There's nothing saying that it has to be at the end of the 100,000,000 weeks. It could just be that this is that week.
Yes, it's unlikely, but coincidences do happen. (And as everyone knows "one in a million chances happen all the time").
We'll just have to wait until divers or submersibles have been down to investigate further. Conjecture at the moment is just that.
OR, we need to get used to having times when there's no electricity available. I suspect this will happen not too far in the future.
Nuclear, hydro & geothermal are the only clean predictable forms of energy generation we have at the moment (Solar is OK in some parts of the world, but not for most people). Hydro & geothermal are (currently) very location sensitive, so the only one which is available to anyone is nuclear - which some people don't like (usually due to misunderstanding how dangerous the waste really is (it's *locally* dangerous, but globally pretty much inert, unlike CO2 which is locally safe, but globally potentially catastrophic)).
So, most places will have to use either unpredictable solar & wind and handle the unpredictable power supply that results.
Has your company done that? I don't know of any ISPs in the UK which have a 'satellite backup plan' you can sign into (you can get your own satellite link, but no one but big companies would even consider that). I doubt any in Europe or the USA would either. Maybe ISPs on islands would, but you just don't expect 2/3rds of a cables for a large chunk of the world to be cut at almost the same time.
Oh come on. How many companies anywhere in the world have a satellite link 'just in case'? Never mind the problem of being able to afford to actually use it.
Most companies will have one DSL connection. Possibly they'll have an ISDN or second DSL available as backup - but that wouldn't help in this case. All that WOULD help would be a satellite link.
The businesses could well be paying more for their Internet link than individuals anyway - we pay about 6 times more for our DSL than a 'home user' account costs. That gives us a lower contention ratio, plus a basic SLA.
Even in the UK, if two of our transatlantic links were severed at the same time, things would slow to a crawl as data gets routed through Germany etc instead. I remember one failing not too long ago and it was very noticeable.
Two out of three failing at the same time is an exceptional event so you can't really expect a developing country to have more than one redundant link available for their two normal ones. How would your region handle the case where **all** their 'normal' Internet links out of the region were severed and they had to fall back to their redundant links???
They're not asking individuals to stop using the Internet at all, just to cut back on all the movie downloads. One movie download is a few hundred thousand emails after all (most of which will be spam..). Also, using the Internet within the country itself would not be a problem
It's not true that 'there are only so many times you can apply patches and fixes'.
Like the Ship of Theseus you can fix things indefinitely, and while it's just an improvement of the old thing it is also a new thing.
Linux has been around longer than Windows 9x or NT, and how many 'patches' have there been for that - loads (even just for the kernel, never mind all the 'extra's, which aren't generally counted as 'part of linux' but their MS equivalents *are* counted as 'part of Windows'), but I don't see anyone saying how it needs to be scrapped and something else started from scratch.
On something as long established as the NT codebase, patching and refactoring is generally a safer & wiser approach than rewriting.
If Vista WAS a complete rewrite, then that is probably at least half the problem. XP was 'good enough' for most people. Really, all that Microsoft should have done is add the new GUI and security features and some other stuff and leave most of it the same. (I'd *hope* that's what they actually did, but I wouldn't be surprised if it was a total rewrite)
Money will always give an advantage - with Google, I can buy for ads to be put at the top/side of the search results page. So, money gives me an advantage.
All my 'ideal' would give would be a general explanation of why a search ranking is low, NOT an improvement in the search ranking.
Page ranking might be "neutral" (not directly affected by anyone at Google), but it's not that simple. If you repeat words, your ranking will drop. If you have duplicated sites, your ranking will drop, if you have lots of links to places with unrelated content, your ranking will drop.
I'm not suggesting that would change. I definitely wouldn't want the exact algorithms to be described by Google. But, being able to have someone give *general* feedback on why a page's ranking has dropped would give a bit more transparency. There would have to be a charge, or people would do it trivially, and Google would get snowed under with requests. There would have to be safeguards to try to stop people doing lots of requests with minor changes to try to reverse engineer Google's algorithms.
SEOs are very good at getting pages to the top of Google's listings for a few days, then to the bottom of their listings for a few years.
FWIW, we do buy ads, but that's not the point. Many people ignore the paid ads on something like Google and just look at the 'proper' listings.
It is quite common for companies to have a working business one day, and then the next, for some unknown reason their site is now listed on page 972 of Google's results, so their business goes down the chute.
This happens because Google change some arbitrary parameter - eg they may now say that they'll treat two duplicated words in the title of any page on a site as spam. So, when Slashdot publishes an article 'Scientific American says that red squirrels prefer red meat', suddenly all of Slashdot disappears from their listings.
It probably wouldn't happen to Slashdot, but that's *probably*. If it did, no one would have any way of finding out why, and fixing it.
>Do Hollywood studios have the right to question Ebert on his methodology? No!- if you get a crappy review, too bad. All these is protected First Amendment free speech.
I agree
BUT! A movie reviewer will usually say WHY they thought the movie was crap (possibly in the vain hope it would improve things in the future).
All google say is '1/10' that's it. They don't say '1/10 - poor storyline' or whatever, just '1/10'
Whilst I agree mostly with Google wanting to keep their search engine ranking algorithms secret, I think it's too easy for people to get dumped at the bottom of the rankings without a clue why.
IMV, it'd be a good thing for someone to be able to pay a (non-trivial) amount of money (say $500) to Google and have them give general feedback on why your page is ranked low - eg 'too many repeated words' or 'irrelevant words' or 'too many crosslinks to bad sites' or whatever, rather than the current scenario of you just being left in the dark with a company which can't make any sales because Google's arbitrary ranking system has taken a dislike to your site.
Being able to pay to have Google re-evaluate your site earlier would be a good thing as well. (Not to be able to increase your ranking, but if you found a problem which you have since fixed).
No it couldn't just be a cookie - it 'notifies an error processing server'
It *could* be Javascript in the ErrorDocument page though, which uses AJAX to talk to the server to get alternate pages to display.
But, if you read the patent it looks like it's actually got to be a plugin or part of the browser itself which notices a 404 error coming from anywhere and does something about it:
"wherein the client component is responsive to detection of the unavailability of a target web page requested by the browser program by sending a request to the error processing server"
How many people really make a mistake? If you buy something from your local shop and then decide you didn't want it after all, the shop has no obligation to give you your money back - especially if they suspect you have used it (eg if it's clothes, a camera etc)
A domain costs virtually nothing to register, and they're not vital for people to live. So, if you screw up and register the wrong domain, tough, it's your fault, not the registrar's, not the rest of the world's. You should have to pay for it.
If GoDaddy are helping spammers by giving them 51 million free domains to use in spam, then I have no sympathy with them!
If anyone else has read the patent, they'll realise that it's a CLIENT-SIDE component that's involved.
Most (all?) 'automatic redirect' systems I've seen are server based - the server runs a script which says 'That page couldn't be found, did you mean any of these...'
I can't imagine who'd put this on the client with client-server communication going on. It sounds like a vastly over-engineered and 'Enterprisey' solution to me. It DOES have the advantage that it can look back in the browser history, but I'm not sure I see how that could benefit the user (the component COULD tell the server what's in the history though, so it could benefit Amazon!)
Criminal cases usually require a more onerous 'burden of proof' than civil cases.
That's why sometimes, (in the UK at least) where people have not been convicted of crimes, their victims will pursue a civil case against them, because it's much easier to get a civil judgement against someone than a criminal conviction (even though the penalties can just be financial not custodial).
If copyright infringement was only allowed as a criminal offence in the USA, I doubt anyone would be convicted of downloading from the Internet. The RIAA wouldn't like that one bit!
Slashdot Mirror
Switched-mode power supplies (as found in most 'big' computing equipment like PCs, monitors, printers etc) are pretty energy efficient, and can use almost no power in 'standby' mode.
:(
Having one big DC supply with lots of outlets in your home is likely to be less energy efficient (because of the large voltage loss over long cable runs at high currents) than having local ones.
That's not to say that having one SMPS per PC, with DC outlets on the back of the PC for monitor, printer etc wouldn't be a good idea - in fact I'd have thought that would be an excellent idea. (As would having DC outlets on the back of your TV for the DVD, TiVo, XBox etc).
The problem might be that there's a patent on this (http://www.freepatentsonline.com/5768600.html)
That's the only way it could work. How else do you block illegal downloads of music/films etc, without blocking ALL downloads of music/films etc. How do you know that what someone is downloading is music/film, especially if it's done using P2P when it's sent in blocks? So, you have to block anything which could be P2P. But, P2P is used for legal stuff as well. People might be downloading stuff illegally from https:/// websites, so you obviously have to block all those as well.
.gov.uk websites as test data. So, every packet needs to/from those sites needs to be inspected (possibly manually), and anything which might possibly be copyrighted and where there's no evidence of a licence, should be blocked (ie pretty much everything except for SYN and ACK packets..)
It just can't work.
The ISPs ought to do organise a 'trial' of the system between themselves, using MPs' accounts and data from
So, every Internet user would have to have a certificate (when you go to a https:/// website, YOUR web browser is encrypting the data). You wouldn't be able to access secure web sites/VPNs/etc from outside the country etc.
Sure, that'd work...
Or maybe not.
If you need more bandwidth, what's wrong with the ISPs expecting you to pay more than someone who uses less?
If you ran a taxi service and someone wanted 20 people to be taken from A to B, you wouldn't just charge them for 1 car's worth of people, you'd charge then for however many cars it took.
I was mainly talking about in the UK (since that's most affected by the BBC iPlayer). Other countries may not publish the contention ratios, but in the UK it's pretty standard.
BTW, my experience of SAPO.PT is that it's pretty poor compared to all the UK ISP's I've been involved with. Unfortunately, the other Portuguese ISPs seem to be pretty poor too.
Exactly.
An average 8Mbps ADSL line in the UK is 8Mbps with a 50:1 contention ratio.
Really, that's a 160kbps line with 8Mbps burst capability. Unfortunately, most people don't understand that. You will usually get very long 'bursts' of 8Mbps, and may never notice it dropping, but as long as your available bandwidth doesn't go below 160kbps, you're getting what you've paid for. Maybe they're mis-sold, or maybe they're just sold in a way that the average user has a slight chance of understanding.
If you want 8Mbps all the time, talk to your ISP about getting an E2 leased line... Then when you see the cost, realise you're actually getting quite a good deal!
With all these things, as long as you can find out beforehand what the real deal is, then everything is OK. A 5Mbps connection with a 2GB 20Mbps burst allowance sounds quite good to me..
ISTM that the BBC should cohost servers at the big ISPs' data centres. Users would stream data from the servers at their ISP. Network traffic from a server in the data centre to the same ISP's customers should be really cheap (free?) as there's no 'Internet' traffic, so it wouldn't be an issue. The only network traffic outside the ISP would be for the cohosted servers to get the data from the BBC.
AFAICS, this would work, at least for the big ISPs, as the BBC wouldn't have to buy many more servers (most of their load is going to be coming from Tiscali, BT and Thus customers anyway, so hosting servers at those ISPs would mean that fewer servers would be needed at the BBC or the one ISP where they currently cohost them) and would need to pay for less Internet bandwidth themselves. They would have to pay for cohosting, but if the servers were ONLY for use by that ISP's customers, I would have thought they could work out a deal where that cost would be minimal.
The even BETTER solution would be to have streaming servers at the exchanges which can have the data trickle fed to them using unused bandwidth, before making the videos available to the users, but that would require some coordinated effort by all the different streaming providers. IMV, if we're going to have true video on demand in the future, that's what will be needed.
All ISPs give DSL lines with a contention ratio. That means that several users share an certain amount of bandwidth. So, a typical 'home user' contention ratio may be 50:1. So, there are 50 users sharing a 'block' of bandwidth. This is well known, and typically described when you are buying a DSL connection.
Usually that's not an issue, most of the users won't be using the internet at the same time, and most of the ones who are will be doing low bandwidth things like browsing websites or downloading email. The issue comes when large numbers start doing high bandwidth things such as streaming or file sharing.
If you aren't happy with that, you could buy a business DSL connection, which may have a 20:1 contention ratio (that's what I do). It costs about 3 times more than a home user connection, but I potentially get 3 times more bandwidth (as well as no 'fair use' limits)
I suppose if people wanted it, ISPs could provide a 1:1 contention ratio, but how many people would be willing to pay 50 times more than a home user connection for that? You could get a T1 leased line for that sort of price.
So, you do get what you pay for, and as long as ISPs honestly publish their contention ratios, it's your own fault if you don't get what you wanted.
Contention ratios are generally a good idea to give people fast internet connections at a lower cost. If ISPs do traffic prioritisation fairly (eg the first 1/50th of the bandwidth block that you use has a higher priority than any excess) then I don't see that anyone has any right to complain (ISPs or customers). This doesn't need to be against net neutrality, you don't prioritise based on the type of data, just the amount.
Usage limits are another issue, and I *do* get annoyed when you get adverts like 'Unlimited Internet for £15 a month' and then in tiny writing "(fair use limits of 20 GB per month apply)". These should be cracked down on IMHO, but contention ratios are sensible and fair.
You're right - 'recompile the kernel' is at least one reason I could never recommend Linux to anyone who's not a true geek. I'm pretty technical, but when I was trying to install a TV card on a Linux PC (for MythTV) and saw 'recompile the kernel' in the instructions, I thought, "WTF? In this day and age, I have to recompile the kernel? How backwards is that?" My 6809-OS9 kernel back in 1983 didn't need recompiling for any device drivers, why the f should Linux "the superest & bestest OS ever" need to have its kernel recompiled by anyone other than someone who's maintaining it?
OS-9 was written 12 years before Linux and they had true dynamically loaded drivers. Windows manages it reasonably well as well, but Linux can't do something as 'simple' as that. No driver should be harder to install than something like 'copy these files there, and change that configuration file like this'. Having to recompile the kernel just to install a device driver makes the whole thing look like it's still in alpha testing stage!
Maybe there's a good reason for it (but I'm not sure I can think what it could be), but that still doesn't make it friendly for 99% of potential users.
That's just what we were saying here. For all I know, a cable might be being cut every couple of months on average, but they repair it within a week, and use the redundant cables in the meantime. No one would notice. But, if several cables go at the same time, it becomes a big problem. If cable faults were that common, then 3 or 4 (or maybe even 5) could well just be coincidence.
Wait until they repair them and see if they say anything, and what. It should be easy to tell whether they were caught on an anchor, bombed or cut (but whether they were caught on an anchor accidentally or deliberately would be harder to tell).
That's a flawed argument.
If the chance of 4 cables being taken offline is 1 in 100,000,000 a week, that means that once in the next 100,000,000 weeks, the probability is that 4 cables will be offline simultaneously. Yes?
There's nothing saying that it has to be at the end of the 100,000,000 weeks. It could just be that this is that week.
Yes, it's unlikely, but coincidences do happen. (And as everyone knows "one in a million chances happen all the time").
We'll just have to wait until divers or submersibles have been down to investigate further. Conjecture at the moment is just that.
OR, we need to get used to having times when there's no electricity available. I suspect this will happen not too far in the future.
Nuclear, hydro & geothermal are the only clean predictable forms of energy generation we have at the moment (Solar is OK in some parts of the world, but not for most people). Hydro & geothermal are (currently) very location sensitive, so the only one which is available to anyone is nuclear - which some people don't like (usually due to misunderstanding how dangerous the waste really is (it's *locally* dangerous, but globally pretty much inert, unlike CO2 which is locally safe, but globally potentially catastrophic)).
So, most places will have to use either unpredictable solar & wind and handle the unpredictable power supply that results.
Has your company done that? I don't know of any ISPs in the UK which have a 'satellite backup plan' you can sign into (you can get your own satellite link, but no one but big companies would even consider that). I doubt any in Europe or the USA would either. Maybe ISPs on islands would, but you just don't expect 2/3rds of a cables for a large chunk of the world to be cut at almost the same time.
Oh come on. How many companies anywhere in the world have a satellite link 'just in case'? Never mind the problem of being able to afford to actually use it.
Most companies will have one DSL connection. Possibly they'll have an ISDN or second DSL available as backup - but that wouldn't help in this case. All that WOULD help would be a satellite link.
The businesses could well be paying more for their Internet link than individuals anyway - we pay about 6 times more for our DSL than a 'home user' account costs. That gives us a lower contention ratio, plus a basic SLA.
Even in the UK, if two of our transatlantic links were severed at the same time, things would slow to a crawl as data gets routed through Germany etc instead. I remember one failing not too long ago and it was very noticeable.
Two out of three failing at the same time is an exceptional event so you can't really expect a developing country to have more than one redundant link available for their two normal ones. How would your region handle the case where **all** their 'normal' Internet links out of the region were severed and they had to fall back to their redundant links???
They're not asking individuals to stop using the Internet at all, just to cut back on all the movie downloads. One movie download is a few hundred thousand emails after all (most of which will be spam..). Also, using the Internet within the country itself would not be a problem
It's not true that 'there are only so many times you can apply patches and fixes'.
Like the Ship of Theseus you can fix things indefinitely, and while it's just an improvement of the old thing it is also a new thing.
Linux has been around longer than Windows 9x or NT, and how many 'patches' have there been for that - loads (even just for the kernel, never mind all the 'extra's, which aren't generally counted as 'part of linux' but their MS equivalents *are* counted as 'part of Windows'), but I don't see anyone saying how it needs to be scrapped and something else started from scratch.
On something as long established as the NT codebase, patching and refactoring is generally a safer & wiser approach than rewriting.
If Vista WAS a complete rewrite, then that is probably at least half the problem. XP was 'good enough' for most people. Really, all that Microsoft should have done is add the new GUI and security features and some other stuff and leave most of it the same. (I'd *hope* that's what they actually did, but I wouldn't be surprised if it was a total rewrite)
I've seen some cars where it's not user-configurable, but the dealership can still change it for you if you want.
But the advantage would be minimal.
Money will always give an advantage - with Google, I can buy for ads to be put at the top/side of the search results page. So, money gives me an advantage.
All my 'ideal' would give would be a general explanation of why a search ranking is low, NOT an improvement in the search ranking.
Page ranking might be "neutral" (not directly affected by anyone at Google), but it's not that simple. If you repeat words, your ranking will drop. If you have duplicated sites, your ranking will drop, if you have lots of links to places with unrelated content, your ranking will drop.
I'm not suggesting that would change. I definitely wouldn't want the exact algorithms to be described by Google. But, being able to have someone give *general* feedback on why a page's ranking has dropped would give a bit more transparency. There would have to be a charge, or people would do it trivially, and Google would get snowed under with requests. There would have to be safeguards to try to stop people doing lots of requests with minor changes to try to reverse engineer Google's algorithms.
SEOs are very good at getting pages to the top of Google's listings for a few days, then to the bottom of their listings for a few years.
FWIW, we do buy ads, but that's not the point. Many people ignore the paid ads on something like Google and just look at the 'proper' listings.
It is quite common for companies to have a working business one day, and then the next, for some unknown reason their site is now listed on page 972 of Google's results, so their business goes down the chute.
This happens because Google change some arbitrary parameter - eg they may now say that they'll treat two duplicated words in the title of any page on a site as spam. So, when Slashdot publishes an article 'Scientific American says that red squirrels prefer red meat', suddenly all of Slashdot disappears from their listings.
It probably wouldn't happen to Slashdot, but that's *probably*. If it did, no one would have any way of finding out why, and fixing it.
>Do Hollywood studios have the right to question Ebert on his methodology? No!- if you get a crappy review, too bad. All these is protected First Amendment free speech.
I agree
BUT! A movie reviewer will usually say WHY they thought the movie was crap (possibly in the vain hope it would improve things in the future).
All google say is '1/10' that's it. They don't say '1/10 - poor storyline' or whatever, just '1/10'
Whilst I agree mostly with Google wanting to keep their search engine ranking algorithms secret, I think it's too easy for people to get dumped at the bottom of the rankings without a clue why.
IMV, it'd be a good thing for someone to be able to pay a (non-trivial) amount of money (say $500) to Google and have them give general feedback on why your page is ranked low - eg 'too many repeated words' or 'irrelevant words' or 'too many crosslinks to bad sites' or whatever, rather than the current scenario of you just being left in the dark with a company which can't make any sales because Google's arbitrary ranking system has taken a dislike to your site.
Being able to pay to have Google re-evaluate your site earlier would be a good thing as well. (Not to be able to increase your ranking, but if you found a problem which you have since fixed).
JM2P
No it couldn't just be a cookie - it 'notifies an error processing server'
It *could* be Javascript in the ErrorDocument page though, which uses AJAX to talk to the server to get alternate pages to display.
But, if you read the patent it looks like it's actually got to be a plugin or part of the browser itself which notices a 404 error coming from anywhere and does something about it:
"wherein the client component is responsive to detection of the unavailability of a target web page requested by the browser program by sending a request to the error processing server"
True - but I'm not sure even that's enterprisey enough to conflict with Amazon's patent.
Why have domain tasting at all?
How many people really make a mistake? If you buy something from your local shop and then decide you didn't want it after all, the shop has no obligation to give you your money back - especially if they suspect you have used it (eg if it's clothes, a camera etc)
A domain costs virtually nothing to register, and they're not vital for people to live. So, if you screw up and register the wrong domain, tough, it's your fault, not the registrar's, not the rest of the world's. You should have to pay for it.
If GoDaddy are helping spammers by giving them 51 million free domains to use in spam, then I have no sympathy with them!
If anyone else has read the patent, they'll realise that it's a CLIENT-SIDE component that's involved.
Most (all?) 'automatic redirect' systems I've seen are server based - the server runs a script which says 'That page couldn't be found, did you mean any of these...'
I can't imagine who'd put this on the client with client-server communication going on. It sounds like a vastly over-engineered and 'Enterprisey' solution to me. It DOES have the advantage that it can look back in the browser history, but I'm not sure I see how that could benefit the user (the component COULD tell the server what's in the history though, so it could benefit Amazon!)
Criminal cases usually require a more onerous 'burden of proof' than civil cases.
That's why sometimes, (in the UK at least) where people have not been convicted of crimes, their victims will pursue a civil case against them, because it's much easier to get a civil judgement against someone than a criminal conviction (even though the penalties can just be financial not custodial).
If copyright infringement was only allowed as a criminal offence in the USA, I doubt anyone would be convicted of downloading from the Internet. The RIAA wouldn't like that one bit!