I use Skype a fair amount, and I find it rather flaky.
Why don't you switch to an open protocol which might not be so flakey?
If anyone has had good experiences with alternatives to Skype, that are multi-platform and support voice conferencing of 4-8 people, please let me know!
Set up a CallWeaver server. I use CallWeaver as my server and Ekiga as my softphone and it works fine (also a UTStarCom F1000G as a WiFi phone, but I have all sorts of problems with that owing to UTStarCom's flakey firmware which they won't fix). At my old job we found that SJPhone and X-Lite were reasonable alternatives to Ekiga for the Windows users (although there is a Windows version of Ekiga but my experience is that it's not entirely stable).
You can also use one of the many SIP/PSTN gateways, such as VoIPUser, to gateway calls in from the PSTN if not everyone is able to use VoIP.
My guess is they just got overrun because they were not prepared for the worst case: ALL clients trying to connect AT THE SAME TIME to their master.
This is a pretty good example of why centralised network topologies such as Skype, MSN, etc. are a really Bad Idea. It doesn't take much to take down the entire network.
SIP, XMPP, SMTP, etc are all examples of distributed topologies - there is centralised service required(*) for these networks - if one service provider's network falls over it only affects a small number of users rather than taking out *all* the users using that protocol.
(* Yes, they all require the root name servers, but these days the root name server architecture is pretty resillient through the use of technologies such as anycase. Certainly a lot more resillient than any one organisation could hope to achieve for their own propriatory protocols).
They should have been prepared for the case, that whenever their network would be down for whatever reason all clients would try to connect concurrently!
This is not really a question of preparation - it's a question of a sensible network design. The Skype network (and most other propriatory services) is a flawed design _because_ they want to have control of every aspect of the network. Open protocols are generally designed to allow interoperation of independent autonomous networks so an outage of this magnetude is pretty much impossible.
Really. From the Comcast AUP prohibited use section:
Thats crazy - go change ISP.
I've been running plenty of public services over my residential internet connection on several servers (My ISP gives me a/29 IPv4 network at no cost) for years. This is completely allowed by my T&Cs and is pretty much the same across most ISPs around here.
if you want a non-throttled pipe with no restrictions you should expect to pay for it.
Well yeah, if you're going to hammer your connection 24/7... But if your connection is idle most of the time and you use BitTorrent to download an ISO once a year then they probably shouldn't be throttling you.
ISPs in the UK are starting to moan about having to carry traffic too, even going so far as to suggest the BBC should pay them.
The big ISPs are certainly complaining ("oops, we underpriced our product and are now making a loss - we'll demand that some random 3rd party bail us out of our mess"). Notably many of the smallerISPs are now very explicit about their limits rather than selling everything as "unlimited". The smaller ISPs are showing that if you charge people appropriately and make it clear what they are paying for, even the high bandwidth users can be profitable customers.
Hopefully the end result will be that all ISPs will stop misleading their customers by selling limited accounts as "unlimited".
I don't think Europe is immune to profiteering by reducing the service standards so you can get by on a lesser investment.
Indeed not - underprovisioning the network and deprioritising bittorrent seems to be a reasonably common bad practice. However, by provisioning the network correctly and setting the pricing model appropriately, the high bandwidth users can pay for their own bandwidth rather than being subsidised by everyone else.
These adds all carry disclaimers at the bottom that refer you to the terms of service. The TOS on every major ISp forbid servers, which is what BT is.
Really? The last time I saw an ISP ban their users from running services was a few years ago. That was NTL and they stopped doing that several years ago AFAIK.
Also, I'm not entirely sure you can claim bittorrent is a service since it uses a peer-to-peer model rather than a client/server model. Certainly from the end user perspective, I use my bittorrent client in the same way as I use my FTP client.
What if BitTorrent consumes 40% of Comcast's bandwidth, but BitTorrent is used by only 5% of Comcast's customers (i.e., accounts for only 5% of Comcast's ISP revenue). Wouldn't the loss of that 5% of their ISP customer revenue be worthwhile from an accounting viewpoint, if it indeed frees up an extra 40% of bandwidth
Not necessarilly. It only makes sense if the cost price of that 40% of their bandwidth is higher than that 5% revenue.
You can always find a bunch of people using a disproportionate amount of bandwidth - once you've got rid of the bittorrent traffic the "average bandwidth used" across the network is going to be reduced. Then you might find that people who play games use a greater than the average amount of bandwidth used. So you kick off all the gamers, the average will drop again and then you can select a new group who's using more than average - eventually you're going to be left with an ISP that refuses to cator for anyone who is doing anything more than SSH.
The solution, which a lot of UK ISPs are now moving towards (especially the smaller ones, who aren't goverened by moronic marketting people quite as much as the big ones) is to stop selling "unlimited" accounts - sell several different grades of account with clearly defined bandwidth limits. The people who use a lot of bandwidth therefore have to pay more than those who are just doing a bit of light web surfing and so the high bandwidth users cover their own costs rather than being subsidised by everyone else.
Take a look at UKFSN's prices for example. If you're a low usage user, you get their basic package which allows 3GB a month in the "peak" hours and 30GB a month at all other times. If you want to run BitTorrent 24/7 you can pay more and get 150GB peak, 420GB off-peak (and various packages in between). The limits are clearly defined and what happens when you exceed them is also explicitly defined (you get bumped up to the next price band for a month and get a 5ukp charge on top).
Sadly a lot of ISPs seem to consider BitTorrent to be an easy target for cost savings. I'm a big fan of traffic shaping - it keeps my SSH and VoIP stuff usable while someone's hammering away with BitTorrent. However, IMHO traffic shaping should only come into effect in exceptionaly situations - if I've been sold an 8Mbps connection with a 30GB/month cap, I damned well should be able to get close to 8MBps most of the time if I haven't gone anywhere near my usage cap no matter what protocol I use. If my traffic has to be throttled by the ISP most of the time then they have underprovisioned their network and that needs fixing.
The ISPs have brought this problem on themselves to a large extent by not being open about the limits. If an ISP says "8Mbps unlimited broadband" should they be at all surprised when people want to be able to do 8Mbps 24/7? They should be making it clear what you're getting much like datacentres do. e.g. Burst speed of 8Mbps, average speed 2Mbps, monthly cap 30GB.
The bottom line is that Comcast needs to make room for more HD TV channels on their cable, and the easiest, most cost-effective way to do that is to kill off the unprofitable BitTorrent traffic.
Or just charge the high bandwidth users more which would pay for the network capacity to be increased. There's no need to single out specific protocols here - if someone's using a large amount of bandwidth for *anything* make them pay more.
They sold internet connections at lower than cost of the bandwidth, betting on the customers not using anywhere near their bandwidth entitlement. Then the BBC produced iPlayer, which is encouraging people to use up more of their bandwidth and thus causing the ISPs to make a loss. So the ISPs are demanding that the BBC pay them to cover the shortfall.
To cut a long story short: the ISPs underpriced their connections and advertised them as "unlimited", were caught out when people actually tried to use what they had paid for and are now demanding that a third party bail them out of their mess. I certainly hope the BBC tell them to go screw themselves - I'm not going to be happy if part of my licence fee goes to propping up idiot ISPs who can't deliver on their commitments.
I think you are confusing the method used to manipulated the DOM with when the script is run.
The two are quite strictly linked - document.write() is required to be executed at load time and inserts the code after the script tag where it is called, whereas updating the DOM tree itself can be done after the page has finished loading.
You know you can just put the ad code at the end of the html. Right before you run the ad code is run, move your element where ever your want.
Moving elements after they have been parsed into the DOM tree constitutes "altering the DOM tree" so my point still stands - the correct method of implementing ads is to load them at the end of the document and insert them into the DOM tree.
I don't understand why developers have to be idiots and put the ad code in the middle of the content.
Because that's what the ad agencies tell them to do. If you sign up to something like AdSense you get handed some code and told "paste this where you want the ad to appear" (or words to that effect). Ignoring what the advertiser tells you to do, sticking the code at the end and hacking up some code to mangle the DOM tree to move the ads to where you want them takes clue, and clue seems to be one thing a good proportion of web developers are missing.
I don't understand why the ad agencies can't just ask you to insert an <object> tag pointing at their ad server - they can work out all they need to know by looking at the GET request and the referrer header, no need for javascript in the main page to generate all the necessary elements.
I still really wish people would stop using document.write() though - it makes it impossible to validate the compliance of the whole page and is completely incompatible with XHTML. I end up having to call a HTML 4.01 page inside an <object> tag in my XHTML 1.1 pages in order to get AdSense to work.
I guess that depends on what you think is most important. I'd imagine advertising agencies consider thier adverts one of the most important things on the page.
I'm sure they do. However, the ad agencies also think that it's important to have exceptionally annoying and distracting adverts and this behaviour is leading to people blocking the ads. If the advertising agencies want people to actually see their ads they will have to go some way to respecting the end-user's preferences as well as their own.
Last time I checked, modifying the dom tree is typically slower than document.write or innerHTML
Quite probably. But it is still many orders of magnitude faster than waiting for the script to load from a slow server. There are 2 methods which could be used:
document.write(): 1. Start parsing/rendering the page 2. Reach the script tag 3. Suspend parsing/rendering 4. Request the script 5. wait 6. Execute the script, which uses document.write() to insert some HTML at the current position in the document 7. Resume parsing/rendering 8. Done
Modifying the DOM tree: 1. Start parsing/rendering the page 2. Finish parsing/rendering 3. Request the script 4. wait 5. Execute the script, which mangles the DOM tree 6. Re-render the page 7. Done
In the document.write() case, most of the page isn't rendered until step (7) has completed - so you had to wait for the script to be received and executed before you could see the content. In the second case, everything except the ads has been displayed by step (2) and you can sit there browsing the page while it's trying to download the script.
The problem with ads is generally not the javascript execution time, but the ad server response time.
The problem is not the JavaScript execution time - the problem is the order of the critical path. If you use document.write(), much of the document can't be parsed and rendered until the script has been fetched and executed, whereas if your script modifies the DOM tree, it can be fetched at the end and therefore the page can be rendered before the script has been fetched.
As with most critical path planning, it is beneficial to do the important stuff first (getting the page into a readable state) and leave the less important stuff until the end (loading the ads and other relatively unimportant scripts), even if it might take marginally longer.
there is no effective gravity between the objects in the rubber sheet experiment.
Correct.
It does indeed illustrate the effect of distortion of the space in which the bodies exist, as opposed to the effects gravity *between* the objects. Yes, the distortion is manifest through external gravity, but so what?
The fact that the distortion is created by gravity seems moot.
You could distort the rubber without needing external gravity - my point is that the external gravity is required to cause the objects to fall into the distortions - the mere existance of the distortions without external gravity won't cause the "gravity-like" behaviour on the rubber sheet.
I would expect, as a more internet saavy pool of users than the general population, that Firefox users tend to spend more online.
Yeah, but that spending is probably going to be through informed research rather than clicking on the first flashing music playing Flash ad they see...
But if you're trying to actually use that illustration as _the_ model, it's a bit like saying that a rose can't exist because the picture of a rose is flat and doesn't smell.
No, I'm not saying "gravity can't possibly be caused by a distortion", I'm saying that the rubber sheet model doesn't describe how *just a distortion* can cause the effects of gravity.
So I don't really see the model as beneficial - it doesn't describe how a distortion in space (the rubber sheet) can cause stuff to accelerate.
I'd think that a better ad blocker would be one that just blocked flash and converted animated GIFs into non-animated images and then displayed them.
Why is it my job to fix their intentionally annoying ads so they are nolonger annoying? If they care that much they can serve up static ads themselves, otherwise I'll block them.
so long as advertisers aren't being dumbasses, their ads will still be seen.
I think creating intentionally annoying flashing music playing ads constitutes "being dumbasses".
Or they are people who, Oh don't know - run a useful and popular free-to-use Web resource and need to raise some income to maintain the service. Shocking idea isn't it?
The user of a web site is free to block ads - there is no law preventing that.
Similarly the owner of the site is free to block any users he wishes (except for discriminating against people based on disability, etc).
It's probably also completely legal for the user to work around any such blocks (seems like it might be a bit of a grey area since you're getting into the realm of "unauthorised access". However, I suspect the site owner doesn't have much of a leg to stand on if it is publically accessible and they are just blocking specific groups of people).
However, which ever way you look at it, blocking ads is _not_ theft. Theft involves depriving someone of something they own - this is not something you are doing when you block ads. Making baseless accusations of theft and conspiracy _is_ illegal - it's called slander.
Sure, you're always going to get people who block ads, just the same as you're always going to get people who fast-forward through them on TV. The trick is that less people will block your ads if they aren't blatently annoying and are possibly actually useful (i.e. targetted).
For example, I'm quite happy to have Google ads and static banner ads displayed on my browser and sometimes even click on them because they are useful. But any ad server that serves up annoyingly animated ads or Flash is going to get their whole domain ad-blocked. As your customer you can advertise to me but I won't put up with being outright abused by your adverts and I have every right to block them.
In any case, his own statistics state that 15% of people use FireFox (I'm not going to debate if that's a trustworthy figure, but it's the figure he is presenting). I'd wager that the vast majority of those people don't use adblock, and of the ones that do probably very few block *all* adverts. 15% is actually a reasonably big proportion and anyone who explicitly blocks 15% of their audiance is clearly a moron.
Flash ads have lots of problems - They use CPU time that I want to use for something else - Just leaving a browser open with a flash ad uses more battery power, shortening the battery life of my notebook - They cause the annoying "this site needs a plugin" bar to appear in FireFox on my 64 bit machines (which don't have Flash installed) - They use lots of bandwidth - Often they have sound that automatically starts playing when the site loads - Often they are designed to be expandable and because FireFox doesn't handle the z-index properly when displaying plugins the space the ad can expand into covers up the page I want to see
I used to be quite selective with what ads I blocked, but these days advertisers that serve up flash ads end up getting their entire domain adblocked.
Is it possible for browsers to render everything *else* on a page while awaiting the ads to be served?
Most ad systems seem to work by placing a <script> tag where you want the ad to appear which loads a script from the ad server that does a document.write() to insert the actual code. This is very bad practice (and explicitly disallowed for XHTML) but even Google do it (which sucks since I have to jump through all sorts of hoops to get AdSense to work on my XHTML site).
document.write() works by actually writing out HTML and feeding it into the parser and thus parsing the page must be suspended at that point until it's finished executing, so you can't render the page until the advert has loaded.
The _correct_ way to do this is for the ad-serving Javascript to actually modify the DOM tree. But that requires the ad server developers to not be lazy and have clue, which seems to be asking too much. (or alternatively, don't use Javascript at all).
The usual analogy is that it's like having a horizontal rubber sheet and placing a steel ball upon it. You'll get an indentation in the sheet. The effects on other nearby bodies, or on their movement, is basically just the consequence of that distortion of space.
I've never understood this explanation since it uses gravity to describe gravity.
1. Placing a weight on a rubber sheet causes a distortion in the rubber sheet because gravity is pulling it down. Without gravity the weight would just sit at the position it was placed (assuming zero velocity) with no distortion.
2. A stationary ball placed on the sheet will be pulled into the well by gravity. If there were no gravity the ball will just sit where it is placed even though there is a distortion in the sheet.
So it seems to me that this model does not show that gravity is a simple distortion - it just shows that a distortion combined with some "downward" force can be used to model gravity (with no explanation as to what that downward force might be).
Re:This is relatively benign ...
on
Manhattan 1984
·
· Score: 2, Interesting
In the UK, we already have to pay fuel tax _and_ road tax, which is then supposed to go into maintaining the road system
The sad thing is that the road tax (coupled with insurance costs) is a disincentive to be fuel efficient. If, like me, you need a van for weekends, the cost of taxing and insuring 2 vehicles means you're more likely to just drive that horribly inefficient van around all week rather than getting a more efficient car for your daily commute to work.
And now the government wants to introduce per-mile charging which not only makes travelling more expensive but reduces the disadvantages of having an inefficient vehicle. Why bother getting an efficient car when you're paying the same per mile as the massive 4x4's? Not only that, but it's a government IT project so I fully expect it to be excessively expensive (tracking hardware in every car and the massive infrastructure needed to use it, plus all the brown envelopes going into the pockets of EDS or similar) and an almighty cock-up (can you say "NHS database"?)
Unfortunately, despite the huge amount of money raised through these taxes, many of the roads are in a terrible state of repair and there are an increasing number of roads with excessive "traffic calming" measures on with questionable legality (I would just love to see speed bumps ruled illegal under the disabilities discrimination act).
Here in Manchester, they want to introduce the 'Congestion Charge' scheme in the same way they have in London.
A few years ago, Southampton council were proposing to do the same thing. At the time, I was working in the city centre and frankly the introduction of a congestion charge would've caused us to move the office. i.e. it would not only kill the city centre through businesses moving out, but it would *prevent* the use of public transport since there is no feasable way to service the out of town locations that the offices would move to.
The current state of affairs seems to be that the government thinks they can prevent people from using their cars through raising the cost. The trouble with that thinking is that there is no alternative - public transport just isn't up to the job.
- Public transport to and from the city centres should be improved (the last few times I took a bus from home into the city centre it took 5 times as long as it would've taken in the car and on a number of occasions the bus didn't even turn up). - Trains need to be made cheaper. It's more expensive for me to take the train to London than drive and park there for a day, even if I'm the only person in the car. - Long distance coach services are a joke - 7 or 8 hours to do a journey that it takes me 3.5 hours to do in the car. This is mostly because they have to get off the motorway and go into a city centre to make each stop along the route. This is easilly solved by dropping people off at motorway service stations and using minibuses to get them from the motorway to the city centre. - I don't believe it's possible to provide sensible bus services for people needing to go from one out of town location to another out of town location, so we need to just accept that people will need to use their cars for this. - Promote flexible working conditions in appropriate industries. Grants for companies that embrace flexible working, taxes for those that don't. If people aren't tied to the 0900-1730 office hours there would be a lot less congestion on the roads.
Now you too can look forward to people using fake license plates to avoid charges
I'm especially amused by the people walking around London with car number plates printed across their T-shirts (yes, the cameras do pick them up and charge people.) Unlike displaying a fake plate on your car, wearing a T-shirt with a number plate printed on it is not illegal.
Slashdot Mirror
I use Skype a fair amount, and I find it rather flaky.
Why don't you switch to an open protocol which might not be so flakey?
If anyone has had good experiences with alternatives to Skype, that are multi-platform and support voice conferencing of 4-8 people, please let me know!
Set up a CallWeaver server. I use CallWeaver as my server and Ekiga as my softphone and it works fine (also a UTStarCom F1000G as a WiFi phone, but I have all sorts of problems with that owing to UTStarCom's flakey firmware which they won't fix). At my old job we found that SJPhone and X-Lite were reasonable alternatives to Ekiga for the Windows users (although there is a Windows version of Ekiga but my experience is that it's not entirely stable).
You can also use one of the many SIP/PSTN gateways, such as VoIPUser, to gateway calls in from the PSTN if not everyone is able to use VoIP.
My guess is they just got overrun because they were not prepared for the worst case: ALL clients trying to connect AT THE SAME TIME to their master.
This is a pretty good example of why centralised network topologies such as Skype, MSN, etc. are a really Bad Idea. It doesn't take much to take down the entire network.
SIP, XMPP, SMTP, etc are all examples of distributed topologies - there is centralised service required(*) for these networks - if one service provider's network falls over it only affects a small number of users rather than taking out *all* the users using that protocol.
(* Yes, they all require the root name servers, but these days the root name server architecture is pretty resillient through the use of technologies such as anycase. Certainly a lot more resillient than any one organisation could hope to achieve for their own propriatory protocols).
They should have been prepared for the case, that whenever their network would be down for whatever reason all clients would try to connect concurrently!
This is not really a question of preparation - it's a question of a sensible network design. The Skype network (and most other propriatory services) is a flawed design _because_ they want to have control of every aspect of the network. Open protocols are generally designed to allow interoperation of independent autonomous networks so an outage of this magnetude is pretty much impossible.
But here in the US, most people have exactly one ISP to choose from. If you don't like it, you can move somewhere else where there's a different ISP.
You could always get a satellite downlink (yup, the latency sucks, no, they aren't really that expensive).
Really. From the Comcast AUP prohibited use section:
/29 IPv4 network at no cost) for years. This is completely allowed by my T&Cs and is pretty much the same across most ISPs around here.
Thats crazy - go change ISP.
I've been running plenty of public services over my residential internet connection on several servers (My ISP gives me a
if you want a non-throttled pipe with no restrictions you should expect to pay for it.
Well yeah, if you're going to hammer your connection 24/7... But if your connection is idle most of the time and you use BitTorrent to download an ISO once a year then they probably shouldn't be throttling you.
ISPs in the UK are starting to moan about having to carry traffic too, even going so far as to suggest the BBC should pay them.
The big ISPs are certainly complaining ("oops, we underpriced our product and are now making a loss - we'll demand that some random 3rd party bail us out of our mess"). Notably many of the smaller ISPs are now very explicit about their limits rather than selling everything as "unlimited". The smaller ISPs are showing that if you charge people appropriately and make it clear what they are paying for, even the high bandwidth users can be profitable customers.
Hopefully the end result will be that all ISPs will stop misleading their customers by selling limited accounts as "unlimited".
I don't think Europe is immune to profiteering by reducing the service standards so you can get by on a lesser investment.
Indeed not - underprovisioning the network and deprioritising bittorrent seems to be a reasonably common bad practice. However, by provisioning the network correctly and setting the pricing model appropriately, the high bandwidth users can pay for their own bandwidth rather than being subsidised by everyone else.
These adds all carry disclaimers at the bottom that refer you to the terms of service. The TOS on every major ISp forbid servers, which is what BT is.
Really? The last time I saw an ISP ban their users from running services was a few years ago. That was NTL and they stopped doing that several years ago AFAIK.
Also, I'm not entirely sure you can claim bittorrent is a service since it uses a peer-to-peer model rather than a client/server model. Certainly from the end user perspective, I use my bittorrent client in the same way as I use my FTP client.
What if BitTorrent consumes 40% of Comcast's bandwidth, but BitTorrent is used by only 5% of Comcast's customers (i.e., accounts for only 5% of Comcast's ISP revenue). Wouldn't the loss of that 5% of their ISP customer revenue be worthwhile from an accounting viewpoint, if it indeed frees up an extra 40% of bandwidth
Not necessarilly. It only makes sense if the cost price of that 40% of their bandwidth is higher than that 5% revenue.
You can always find a bunch of people using a disproportionate amount of bandwidth - once you've got rid of the bittorrent traffic the "average bandwidth used" across the network is going to be reduced. Then you might find that people who play games use a greater than the average amount of bandwidth used. So you kick off all the gamers, the average will drop again and then you can select a new group who's using more than average - eventually you're going to be left with an ISP that refuses to cator for anyone who is doing anything more than SSH.
The solution, which a lot of UK ISPs are now moving towards (especially the smaller ones, who aren't goverened by moronic marketting people quite as much as the big ones) is to stop selling "unlimited" accounts - sell several different grades of account with clearly defined bandwidth limits. The people who use a lot of bandwidth therefore have to pay more than those who are just doing a bit of light web surfing and so the high bandwidth users cover their own costs rather than being subsidised by everyone else.
Take a look at UKFSN's prices for example. If you're a low usage user, you get their basic package which allows 3GB a month in the "peak" hours and 30GB a month at all other times. If you want to run BitTorrent 24/7 you can pay more and get 150GB peak, 420GB off-peak (and various packages in between). The limits are clearly defined and what happens when you exceed them is also explicitly defined (you get bumped up to the next price band for a month and get a 5ukp charge on top).
Sadly a lot of ISPs seem to consider BitTorrent to be an easy target for cost savings. I'm a big fan of traffic shaping - it keeps my SSH and VoIP stuff usable while someone's hammering away with BitTorrent. However, IMHO traffic shaping should only come into effect in exceptionaly situations - if I've been sold an 8Mbps connection with a 30GB/month cap, I damned well should be able to get close to 8MBps most of the time if I haven't gone anywhere near my usage cap no matter what protocol I use. If my traffic has to be throttled by the ISP most of the time then they have underprovisioned their network and that needs fixing.
The ISPs have brought this problem on themselves to a large extent by not being open about the limits. If an ISP says "8Mbps unlimited broadband" should they be at all surprised when people want to be able to do 8Mbps 24/7? They should be making it clear what you're getting much like datacentres do. e.g. Burst speed of 8Mbps, average speed 2Mbps, monthly cap 30GB.
The bottom line is that Comcast needs to make room for more HD TV channels on their cable, and the easiest, most cost-effective way to do that is to kill off the unprofitable BitTorrent traffic.
Or just charge the high bandwidth users more which would pay for the network capacity to be increased. There's no need to single out specific protocols here - if someone's using a large amount of bandwidth for *anything* make them pay more.
This draws fairly interesting parallels with Tiscali and TalkTalk complaining about the Beeb's iPlayer here in the UK.
They sold internet connections at lower than cost of the bandwidth, betting on the customers not using anywhere near their bandwidth entitlement. Then the BBC produced iPlayer, which is encouraging people to use up more of their bandwidth and thus causing the ISPs to make a loss. So the ISPs are demanding that the BBC pay them to cover the shortfall.
To cut a long story short: the ISPs underpriced their connections and advertised them as "unlimited", were caught out when people actually tried to use what they had paid for and are now demanding that a third party bail them out of their mess. I certainly hope the BBC tell them to go screw themselves - I'm not going to be happy if part of my licence fee goes to propping up idiot ISPs who can't deliver on their commitments.
I think you are confusing the method used to manipulated the DOM with when the script is run.
The two are quite strictly linked - document.write() is required to be executed at load time and inserts the code after the script tag where it is called, whereas updating the DOM tree itself can be done after the page has finished loading.
You know you can just put the ad code at the end of the html. Right before you run the ad code is run, move your element where ever your want.
Moving elements after they have been parsed into the DOM tree constitutes "altering the DOM tree" so my point still stands - the correct method of implementing ads is to load them at the end of the document and insert them into the DOM tree.
I don't understand why developers have to be idiots and put the ad code in the middle of the content.
Because that's what the ad agencies tell them to do. If you sign up to something like AdSense you get handed some code and told "paste this where you want the ad to appear" (or words to that effect). Ignoring what the advertiser tells you to do, sticking the code at the end and hacking up some code to mangle the DOM tree to move the ads to where you want them takes clue, and clue seems to be one thing a good proportion of web developers are missing.
I don't understand why the ad agencies can't just ask you to insert an <object> tag pointing at their ad server - they can work out all they need to know by looking at the GET request and the referrer header, no need for javascript in the main page to generate all the necessary elements.
I still really wish people would stop using document.write() though - it makes it impossible to validate the compliance of the whole page and is completely incompatible with XHTML. I end up having to call a HTML 4.01 page inside an <object> tag in my XHTML 1.1 pages in order to get AdSense to work.
I guess that depends on what you think is most important. I'd imagine advertising agencies consider thier adverts one of the most important things on the page.
I'm sure they do. However, the ad agencies also think that it's important to have exceptionally annoying and distracting adverts and this behaviour is leading to people blocking the ads. If the advertising agencies want people to actually see their ads they will have to go some way to respecting the end-user's preferences as well as their own.
Last time I checked, modifying the dom tree is typically slower than document.write or innerHTML
Quite probably. But it is still many orders of magnitude faster than waiting for the script to load from a slow server. There are 2 methods which could be used:
document.write():
1. Start parsing/rendering the page
2. Reach the script tag
3. Suspend parsing/rendering
4. Request the script
5. wait
6. Execute the script, which uses document.write() to insert some HTML at the current position in the document
7. Resume parsing/rendering
8. Done
Modifying the DOM tree:
1. Start parsing/rendering the page
2. Finish parsing/rendering
3. Request the script
4. wait
5. Execute the script, which mangles the DOM tree
6. Re-render the page
7. Done
In the document.write() case, most of the page isn't rendered until step (7) has completed - so you had to wait for the script to be received and executed before you could see the content. In the second case, everything except the ads has been displayed by step (2) and you can sit there browsing the page while it's trying to download the script.
The problem with ads is generally not the javascript execution time, but the ad server response time.
The problem is not the JavaScript execution time - the problem is the order of the critical path. If you use document.write(), much of the document can't be parsed and rendered until the script has been fetched and executed, whereas if your script modifies the DOM tree, it can be fetched at the end and therefore the page can be rendered before the script has been fetched.
As with most critical path planning, it is beneficial to do the important stuff first (getting the page into a readable state) and leave the less important stuff until the end (loading the ads and other relatively unimportant scripts), even if it might take marginally longer.
there is no effective gravity between the objects in the rubber sheet experiment.
Correct.
It does indeed illustrate the effect of distortion of the space in which the bodies exist, as opposed to the effects gravity *between* the objects. Yes, the distortion is manifest through external gravity, but so what?
The fact that the distortion is created by gravity seems moot.
You could distort the rubber without needing external gravity - my point is that the external gravity is required to cause the objects to fall into the distortions - the mere existance of the distortions without external gravity won't cause the "gravity-like" behaviour on the rubber sheet.
I would expect, as a more internet saavy pool of users than the general population, that Firefox users tend to spend more online.
Yeah, but that spending is probably going to be through informed research rather than clicking on the first flashing music playing Flash ad they see...
But if you're trying to actually use that illustration as _the_ model, it's a bit like saying that a rose can't exist because the picture of a rose is flat and doesn't smell.
No, I'm not saying "gravity can't possibly be caused by a distortion", I'm saying that the rubber sheet model doesn't describe how *just a distortion* can cause the effects of gravity.
So I don't really see the model as beneficial - it doesn't describe how a distortion in space (the rubber sheet) can cause stuff to accelerate.
I'd think that a better ad blocker would be one that just blocked flash and converted animated GIFs into non-animated images and then displayed them.
Why is it my job to fix their intentionally annoying ads so they are nolonger annoying? If they care that much they can serve up static ads themselves, otherwise I'll block them.
so long as advertisers aren't being dumbasses, their ads will still be seen.
I think creating intentionally annoying flashing music playing ads constitutes "being dumbasses".
Or they are people who, Oh don't know - run a useful and popular free-to-use Web resource and need to raise some income to maintain the service. Shocking idea isn't it?
The user of a web site is free to block ads - there is no law preventing that.
Similarly the owner of the site is free to block any users he wishes (except for discriminating against people based on disability, etc).
It's probably also completely legal for the user to work around any such blocks (seems like it might be a bit of a grey area since you're getting into the realm of "unauthorised access". However, I suspect the site owner doesn't have much of a leg to stand on if it is publically accessible and they are just blocking specific groups of people).
However, which ever way you look at it, blocking ads is _not_ theft. Theft involves depriving someone of something they own - this is not something you are doing when you block ads. Making baseless accusations of theft and conspiracy _is_ illegal - it's called slander.
Sure, you're always going to get people who block ads, just the same as you're always going to get people who fast-forward through them on TV. The trick is that less people will block your ads if they aren't blatently annoying and are possibly actually useful (i.e. targetted).
For example, I'm quite happy to have Google ads and static banner ads displayed on my browser and sometimes even click on them because they are useful. But any ad server that serves up annoyingly animated ads or Flash is going to get their whole domain ad-blocked. As your customer you can advertise to me but I won't put up with being outright abused by your adverts and I have every right to block them.
In any case, his own statistics state that 15% of people use FireFox (I'm not going to debate if that's a trustworthy figure, but it's the figure he is presenting). I'd wager that the vast majority of those people don't use adblock, and of the ones that do probably very few block *all* adverts. 15% is actually a reasonably big proportion and anyone who explicitly blocks 15% of their audiance is clearly a moron.
What pisses me off are badly designed Flash ads.
Flash ads have lots of problems
- They use CPU time that I want to use for something else
- Just leaving a browser open with a flash ad uses more battery power, shortening the battery life of my notebook
- They cause the annoying "this site needs a plugin" bar to appear in FireFox on my 64 bit machines (which don't have Flash installed)
- They use lots of bandwidth
- Often they have sound that automatically starts playing when the site loads
- Often they are designed to be expandable and because FireFox doesn't handle the z-index properly when displaying plugins the space the ad can expand into covers up the page I want to see
I used to be quite selective with what ads I blocked, but these days advertisers that serve up flash ads end up getting their entire domain adblocked.
Ads *are* the bottle neck, especially on vanilla FF. How? Because these calls, made to google analytics from *my* computer, don't get responses.
Minor point, but since when was Google Analytics an advertising service?
Is it possible for browsers to render everything *else* on a page while awaiting the ads to be served?
Most ad systems seem to work by placing a <script> tag where you want the ad to appear which loads a script from the ad server that does a document.write() to insert the actual code. This is very bad practice (and explicitly disallowed for XHTML) but even Google do it (which sucks since I have to jump through all sorts of hoops to get AdSense to work on my XHTML site).
document.write() works by actually writing out HTML and feeding it into the parser and thus parsing the page must be suspended at that point until it's finished executing, so you can't render the page until the advert has loaded.
The _correct_ way to do this is for the ad-serving Javascript to actually modify the DOM tree. But that requires the ad server developers to not be lazy and have clue, which seems to be asking too much. (or alternatively, don't use Javascript at all).
The usual analogy is that it's like having a horizontal rubber sheet and placing a steel ball upon it. You'll get an indentation in the sheet. The effects on other nearby bodies, or on their movement, is basically just the consequence of that distortion of space.
I've never understood this explanation since it uses gravity to describe gravity.
1. Placing a weight on a rubber sheet causes a distortion in the rubber sheet because gravity is pulling it down. Without gravity the weight would just sit at the position it was placed (assuming zero velocity) with no distortion.
2. A stationary ball placed on the sheet will be pulled into the well by gravity. If there were no gravity the ball will just sit where it is placed even though there is a distortion in the sheet.
So it seems to me that this model does not show that gravity is a simple distortion - it just shows that a distortion combined with some "downward" force can be used to model gravity (with no explanation as to what that downward force might be).
In the UK, we already have to pay fuel tax _and_ road tax, which is then supposed to go into maintaining the road system
The sad thing is that the road tax (coupled with insurance costs) is a disincentive to be fuel efficient. If, like me, you need a van for weekends, the cost of taxing and insuring 2 vehicles means you're more likely to just drive that horribly inefficient van around all week rather than getting a more efficient car for your daily commute to work.
And now the government wants to introduce per-mile charging which not only makes travelling more expensive but reduces the disadvantages of having an inefficient vehicle. Why bother getting an efficient car when you're paying the same per mile as the massive 4x4's? Not only that, but it's a government IT project so I fully expect it to be excessively expensive (tracking hardware in every car and the massive infrastructure needed to use it, plus all the brown envelopes going into the pockets of EDS or similar) and an almighty cock-up (can you say "NHS database"?)
Unfortunately, despite the huge amount of money raised through these taxes, many of the roads are in a terrible state of repair and there are an increasing number of roads with excessive "traffic calming" measures on with questionable legality (I would just love to see speed bumps ruled illegal under the disabilities discrimination act).
Here in Manchester, they want to introduce the 'Congestion Charge' scheme in the same way they have in London.
A few years ago, Southampton council were proposing to do the same thing. At the time, I was working in the city centre and frankly the introduction of a congestion charge would've caused us to move the office. i.e. it would not only kill the city centre through businesses moving out, but it would *prevent* the use of public transport since there is no feasable way to service the out of town locations that the offices would move to.
The current state of affairs seems to be that the government thinks they can prevent people from using their cars through raising the cost. The trouble with that thinking is that there is no alternative - public transport just isn't up to the job.
- Public transport to and from the city centres should be improved (the last few times I took a bus from home into the city centre it took 5 times as long as it would've taken in the car and on a number of occasions the bus didn't even turn up).
- Trains need to be made cheaper. It's more expensive for me to take the train to London than drive and park there for a day, even if I'm the only person in the car.
- Long distance coach services are a joke - 7 or 8 hours to do a journey that it takes me 3.5 hours to do in the car. This is mostly because they have to get off the motorway and go into a city centre to make each stop along the route. This is easilly solved by dropping people off at motorway service stations and using minibuses to get them from the motorway to the city centre.
- I don't believe it's possible to provide sensible bus services for people needing to go from one out of town location to another out of town location, so we need to just accept that people will need to use their cars for this.
- Promote flexible working conditions in appropriate industries. Grants for companies that embrace flexible working, taxes for those that don't. If people aren't tied to the 0900-1730 office hours there would be a lot less congestion on the roads.
Now you too can look forward to people using fake license plates to avoid charges
I'm especially amused by the people walking around London with car number plates printed across their T-shirts (yes, the cameras do pick them up and charge people.) Unlike displaying a fake plate on your car, wearing a T-shirt with a number plate printed on it is not illegal.
will give up their privacy for a discount card at the supermarket.
There is a big difference between voluntarilly giving up your privacy and being *required* to do so though.
They also want to have their bases covered when it comes to liability.
Good luck with holding *any* software company liable after you agree to that EULA disclaiming all liability.