The summary explicitly states that it is a "presentation".
No, it states that there was a presentation, i.e. somebody standing up in front of a room full of people. I expected a transcript of the presentation, not the slides that accompanied the presentation.
Couldn't you just have some extension to drop the credentials server-side so that the client would have to authenticate again if they clicked on such-and-such a link?
HTTP is stateless. The client has to supply authentication for each request anyway. The only reason you see a login the first time and then you are "logged in" for the remainder of your visit is because your browser hides that implementation detail for you.
Imagine if your browser asked you for your username and password each and every time you clicked on a link when you think you are logged in. That's the way it actually works, except your browser automatically retransmits the authentication using the information you typed in at the beginning of your session.
Basically it makes no sense to "drop the credentials server-side" because it's the client that continues to authenticate, over and over again, and the server has no official way of stopping it as per the spec.
I've never found a link so compelling that I would click on it without knowing something about where I was about to go and what the file type was.
You don't know the file type. The file extension is meaningless. You can follow a link to some resource named foo.html and it might be a PDF, and you can follow a link to some resource named foo.pdf and it might be an HTML document.
Same goes for the location, unless you disable client-side scripting or manually copy & paste the URIs instead of clicking on the links.
Unless you manually perform an HTTP HEAD request, you don't know the type of file you are requesting.
I've never had a download from an FTP server ever fail. I've had *many* fail from HTTP served downloads that I really do try avoid downloading anything over about 3MB on HTTP.
I've downloaded many, many large files (e.g. ISOs) over HTTP with no problem and have done for years. If you are having problems downloading anything over 3MB, then I would guess that you are misconfiguring these computers. Really - you think the rest of the world is just putting up with flaky downloads?
TCP ensures an error-free connection for both FTP and HTTP. Neither FTP nor HTTP handle that part of the work. When you say "fail", what do you mean, exactly? Dropped connections? Corrupted files?
In any case, your personal experience and my personal experience is unimportant. That's what I was asking for stats. You are the one claiming that HTTP is unsuitable for large downloads; the burden of proof is on you to show that.
You keep skipping over torrents.
I think you are confusing me with somebody else.
Really, are you trying to attack one point by ignoring points you can't argue?
Er, what? I'm arguing that HTTP isn't as bad compared with FTP as you make it seem. BitTorrent doesn't factor into that argument whatsoever.
Or will you acknowledge that torrents can be far better than HTTP for downloads of large files?
I'll acknowledge that all three protocols have advantages and disadvantages. BitTorrent is not a silver bullet, the fact that users have to download and install additional software is a showstopper for many people, as is the fact that it's not simply a client downloading from a server (e.g. you have to open up ports and sacrifice upstream bandwidth to get a decent speed).
All three protocols "can be" far better than the other two. It depends on the circumstances. For large files, it depends on what servers are available, the update schedule, the bandwidth available, and so on. It's wrong to simply call one "far better" than another.
FTP has the ability to resume from a set point in a file
So does HTTP. Plus it has caching and virtual hosts.
My justification isn't its abbreviation, its the fact that when you're downloading files of larger than 2G, you don't want to risk anything going wrong with a HTTP download (the major risk of being unable to resume in most cases).
What? Only really ancient servers don't support byte range requests. Apache certainly does. Unless you are talking about dynamically generated files, in which case, how many 2GB dynamically generated files have you come across lately?
Do you have any stats to back up your claim that most servers don't support byte range requests, or are you just going on memories from years ago?
The basic auth logout: yes, people have been asking for it for years, but it's HTTP itself that doesn't provide a mechanism for logging out users, it's not Apache's fault.
The lax syntax: hell no. That sort of thing leads to security holes. If I make a typo, I'm perfectly capable of going back and fixing it, should Apache notify me immediately. But if it misinterprets a typo as meaning something I didn't want, I won't know about it until it bites me in the ass. There is nothing wrong with strict syntax for config files.
The rest are relatively sane complaints and ones I've said myself in the past. Even if they are contained in a godawful PDF.
I suspect most people here are able to position the cursor over the article link and look in the status bar, note the.pdf at the end of the URL, and know that this is a PDF.
Assuming they are able to do it is one thing. Expecting them to do it every time they follow a link is another thing entirely.
Let me get this straight: you think that the Boston Tea Party was caused because you wanted the right to be innocent until proven guilty? Your teachers at school let you down badly.
You threw the tea away because you didn't want taxation without representation. It had nothing - nothing whatsoever - to do with innocent until proven guilty.
In fact, the USA's justice system is heavily based on the English justice system. That "innocent until proven guilty" you were talking about? Your cornerstone case that gave you that right (Coffin vs United States) directly referenced English precedent.
So not only are you completely wrong about why you threw the tea away, but you say it's because England denied you something when in actual fact part of the reason you have it is because of England.
the past two presidents have both tried Marijuana and/or cocaine -- you'd think they'd both know that neither is worth the billions of dollars spent annually to stop the inflow.
Isn't worth it to who? It's not their billions of dollars they are spending, it's the taxpayers'.
On the other hand, making any sort of movement towards legalisation of drugs will piss off two important groups of people - Christians and drug dealers.
Typical Christians of the sort Bush relies on won't vote for him if he says anything other than "drugs are evil".
Drug dealers collectively stand to lose billions should things like marijuana become legalised. They'll assassinate anybody who attempts to take their business away from them.
So the reason Bush and Clinton continue with the "drugs are bad, mmkay?" rhetoric is because a) they want votes and b) they want to live. The decision making process never involves doing the right thing for the country, because the democratic process and the black market don't allow for that.
The person you are replying to was pointing out that Sun's stance causes fragmentation of the Java platform.
Pointing out that Sun not wanting to fragment the Java platform is entirely beside the point. The best thing they could do to avoid fragmentation would be to make their implementation open-source. Anything else, and it will force many people to create new implementations rather than use Sun's.
Set your mail client to check for new email once an hour.
Switch your phone to voicemail.
If your boss won't let you, then it's an organisation problem, because your boss absolutely needs to understand that this is how to get you to work most efficiently.
Sigh. Now you are just trying to weasel out of the argument by playing with semantics.
A web page is on the web. A web page is part of the web. A mailbox is on the web. A mailbox is part of the web. A newsgroup is on the web. A newsgroup is part of the web.
In this context, "on" and "part of" are synonyms. Here's a quote:
The Web is a universe of resources. A resource is defined by [RFC2396] to be anything that has identity. Examples include documents, files, menu items, machines, and services, as well as people, organizations, and concepts.
Note the words "The web is...". Not "The web points to..." or "The web contains...". He's talking about the web being directly comprised of the resources that are linked together. You know, in a web-like manner? In other words, leaf nodes are still nodes.
Arguably, mailto: URIs do put email 'on the web'.
So the original complaint that it was not proper to refer to Mozilla as a "web suite" is without merit - even if you draw a distinction between "on the web" and "part of the web". Right?
email functions quite happily without URIs.
So what? Something doesn't have to rely on URIs to be part of the web. If you point to a resource with a URI, you make it part of the web, whether the resource wants to be part of the web or not.
I could put a bus journey to my house 'on the web' by inventing a new URI scheme that allows 'bus://my-house?via=the_chip_shop'.
Your URI scheme wouldn't be recognised by anybody but you though.
The wouldn't make the bus, my house or the chip shop part of the Web.
If you had broad agreement on the bus URI scheme (e.g. had it ratified as an IETF standard), then yes, it would.
The web is an abstract concept, not a protocol or network connection. It links resources together. If your bus journey is defined as a resource, and it is pointed to by a URI, then it's part of the web.
Your mistake is thinking that "URI == part of the web" is the extreme part of your example. It's not. Defining the bus journey as a resource is the extreme part of of that example.
You:I don't deny that URIs are part of the Web. That doesn't make the things they point to part of the Web
Tim Berners-Lee:An information object is "on the web" if it has a URI.
I already pointed you to that quote. Try reading what I link to - it directly contradicts what you are saying. I'm not linking to random documents you know, they actually have relevence.
I'm saying it's almost universally accepted that 'World Wide Web' means HTTP+HTML.
This thread started because an AC felt the need to "correct" somebody's "misuse" of the term 'web'. Obviously it's not that universally accepted, especially as the guy who invented the term doesn't agree with the definition.
I would not imagine that he considers email to be part of the Web.
What's the point in me posting links if you aren't going to read them? The reason URIs are called URIs and not UDIs is because the web includes non-document resources like email addresses. Axioms of Web Architecturewritten by Tim Berners-Lee explicitly mentions mailto.
You probably think that the only way to beat terrorism is to adopt it.
I can't see how you can remotely think that from what I posted.
I think if we truly wanted to win the war on terror we would tie enemy combatants up and execute them in public.
Hmm... let's see, who does things like that? Oh yeah, theTaliban.
And you say I want to adopt terrorism? You are the one advocating we take a leaf out of the Taliban's books!
This is how it has always worked in the past.
Except it doesn't work, does it? If it had worked in the past, we wouldn't still be dealing with it, would we?
Why should we think that all of a sudden being considerate and gentle is a valid tactic?
Straw man. I never said we should be "considerate and gentle".
Since you seem to have completely missed the point of my comment, I'll use smaller words and make it clearer: It's proven fact that a subset of the Coalition forces will happily torture people. Therefore the parent's claim that this weapon is relatively safe because the Coalition forces are nice guys is of no value.
No. "Web server" is actually a common term for an HTTP server (there's the misconception that the WWW==HTTP again), Sendmail is not a web server, because the term "web server", as everybody knows it, refers to software that exclusively deals with the HTTP portion of the web.
But yes, as there is a mailto: URI scheme, email is part of the WWW, and you could call Sendmail part of the WWW infrastructure.
Hosts whose names start with www can reasonably be expected to speak NNTP?
Of course not. A hostname starting with 'www.' might denote that it is part of the WWW, but that doesn't mean that it must necessarily speak every protocol that comprises the WWW, any more than a hostname beginning with 'mail.' means that there must be a POP3 server, IMAP4 server, webmail server and SMTP server all running on it.
Are you actually disagreeing with Tim Berners-Lee, or are you saying that he doesn't really mean it when he explains what the WWW is?
Web Browser is web. Mail is not. IRC is not. News is not.
That's a common misconception. According to Tim Berners-Lee, the guy that coined the term "World-Wide Web", anything addressable by a URL (now URI) is part of the WWW. This includes Usenet, as there are nntp: and news: URI schemes.
Have a little faith in the people serving on the front lines. We're professionals, just like you try to be at work. We care about honor, courage, commitment, etc.
It would be easier to have faith in you and consider you to be professionals if we hadn't seen photos of you guys torturing people and laughing about it.
You want us to give you the benefit of the doubt? Why should we? Some of you are good guys, sure, but some of you are evil bastards who shouldn't be trusted with a plastic fork, let alone anything like this. What, you want us to believe that only the "good guys" will be able to use this? Why would that be the case when the armed forces already allow the sickos access to guns, prisoners, etc?
Wonder if making people feel like they're being burned alive counts as torture?
If it leaves no marks on their bodies, does it make a difference? Call it torture or don't call it torture, nobody's going to get caught unless they are stupid enough to take photos of themselves doing it.
Just looking at my website statistics from people coming to my website via slashdot.org, I actually have a large number at the beginning of the workday, and towards the end, but during the day, it looks like most people stay pretty productive...
Or maybe it takes most of a day for your pages to expire from ISP's caches. You really shouldn't pay any attention to your logs, you can't count visitors with them, despite what many snake-oil sellers tell you.
What "beginning of the workday" are you talking about, anyway? You do realise that there is more than one timezone in the world, right?
Slashdot Mirror
Doesn't work for me: when I put in my address or postcode, it actually points to a road adjacent to mine.
I guess "regular" maps can now be officially declared dead. This is right on so many levels, and implementation appears to be flawless.
Regular maps don't need a computer and Internet connection. Those are two pretty big flaws.
The summary explicitly states that it is a "presentation".
No, it states that there was a presentation, i.e. somebody standing up in front of a room full of people. I expected a transcript of the presentation, not the slides that accompanied the presentation.
Couldn't you just have some extension to drop the credentials server-side so that the client would have to authenticate again if they clicked on such-and-such a link?
HTTP is stateless. The client has to supply authentication for each request anyway. The only reason you see a login the first time and then you are "logged in" for the remainder of your visit is because your browser hides that implementation detail for you.
Imagine if your browser asked you for your username and password each and every time you clicked on a link when you think you are logged in. That's the way it actually works, except your browser automatically retransmits the authentication using the information you typed in at the beginning of your session.
Basically it makes no sense to "drop the credentials server-side" because it's the client that continues to authenticate, over and over again, and the server has no official way of stopping it as per the spec.
I've never found a link so compelling that I would click on it without knowing something about where I was about to go and what the file type was.
You don't know the file type. The file extension is meaningless. You can follow a link to some resource named foo.html and it might be a PDF, and you can follow a link to some resource named foo.pdf and it might be an HTML document.
Same goes for the location, unless you disable client-side scripting or manually copy & paste the URIs instead of clicking on the links.
Unless you manually perform an HTTP HEAD request, you don't know the type of file you are requesting.
Shame you'll be using the TCP/IP/UDP protocols then :)
Some FTP servers can do this too - e.g. try downloading a directory name with .tar.gz appended.
I've never had a download from an FTP server ever fail. I've had *many* fail from HTTP served downloads that I really do try avoid downloading anything over about 3MB on HTTP.
I've downloaded many, many large files (e.g. ISOs) over HTTP with no problem and have done for years. If you are having problems downloading anything over 3MB, then I would guess that you are misconfiguring these computers. Really - you think the rest of the world is just putting up with flaky downloads?
TCP ensures an error-free connection for both FTP and HTTP. Neither FTP nor HTTP handle that part of the work. When you say "fail", what do you mean, exactly? Dropped connections? Corrupted files?
In any case, your personal experience and my personal experience is unimportant. That's what I was asking for stats. You are the one claiming that HTTP is unsuitable for large downloads; the burden of proof is on you to show that.
You keep skipping over torrents.
I think you are confusing me with somebody else.
Really, are you trying to attack one point by ignoring points you can't argue?
Er, what? I'm arguing that HTTP isn't as bad compared with FTP as you make it seem. BitTorrent doesn't factor into that argument whatsoever.
Or will you acknowledge that torrents can be far better than HTTP for downloads of large files?
I'll acknowledge that all three protocols have advantages and disadvantages. BitTorrent is not a silver bullet, the fact that users have to download and install additional software is a showstopper for many people, as is the fact that it's not simply a client downloading from a server (e.g. you have to open up ports and sacrifice upstream bandwidth to get a decent speed).
All three protocols "can be" far better than the other two. It depends on the circumstances. For large files, it depends on what servers are available, the update schedule, the bandwidth available, and so on. It's wrong to simply call one "far better" than another.
FTP has the ability to resume from a set point in a file
So does HTTP. Plus it has caching and virtual hosts.
My justification isn't its abbreviation, its the fact that when you're downloading files of larger than 2G, you don't want to risk anything going wrong with a HTTP download (the major risk of being unable to resume in most cases).
What? Only really ancient servers don't support byte range requests. Apache certainly does. Unless you are talking about dynamically generated files, in which case, how many 2GB dynamically generated files have you come across lately?
Do you have any stats to back up your claim that most servers don't support byte range requests, or are you just going on memories from years ago?
The basic auth logout: yes, people have been asking for it for years, but it's HTTP itself that doesn't provide a mechanism for logging out users, it's not Apache's fault.
The lax syntax: hell no. That sort of thing leads to security holes. If I make a typo, I'm perfectly capable of going back and fixing it, should Apache notify me immediately. But if it misinterprets a typo as meaning something I didn't want, I won't know about it until it bites me in the ass. There is nothing wrong with strict syntax for config files.
The rest are relatively sane complaints and ones I've said myself in the past. Even if they are contained in a godawful PDF.
I suspect most people here are able to position the cursor over the article link and look in the status bar, note the .pdf at the end of the URL, and know that this is a PDF.
Assuming they are able to do it is one thing. Expecting them to do it every time they follow a link is another thing entirely.
Innocent until proven guilty.
Let me get this straight: you think that the Boston Tea Party was caused because you wanted the right to be innocent until proven guilty? Your teachers at school let you down badly.
You threw the tea away because you didn't want taxation without representation. It had nothing - nothing whatsoever - to do with innocent until proven guilty.
In fact, the USA's justice system is heavily based on the English justice system. That "innocent until proven guilty" you were talking about? Your cornerstone case that gave you that right (Coffin vs United States) directly referenced English precedent.
So not only are you completely wrong about why you threw the tea away, but you say it's because England denied you something when in actual fact part of the reason you have it is because of England.
the past two presidents have both tried Marijuana and/or cocaine -- you'd think they'd both know that neither is worth the billions of dollars spent annually to stop the inflow.
Isn't worth it to who? It's not their billions of dollars they are spending, it's the taxpayers'.
On the other hand, making any sort of movement towards legalisation of drugs will piss off two important groups of people - Christians and drug dealers.
Typical Christians of the sort Bush relies on won't vote for him if he says anything other than "drugs are evil".
Drug dealers collectively stand to lose billions should things like marijuana become legalised. They'll assassinate anybody who attempts to take their business away from them.
So the reason Bush and Clinton continue with the "drugs are bad, mmkay?" rhetoric is because a) they want votes and b) they want to live. The decision making process never involves doing the right thing for the country, because the democratic process and the black market don't allow for that.
Talk about missing the point!
The person you are replying to was pointing out that Sun's stance causes fragmentation of the Java platform.
Pointing out that Sun not wanting to fragment the Java platform is entirely beside the point. The best thing they could do to avoid fragmentation would be to make their implementation open-source. Anything else, and it will force many people to create new implementations rather than use Sun's.
Set your IM to busy.
Set your mail client to check for new email once an hour.
Switch your phone to voicemail.
If your boss won't let you, then it's an organisation problem, because your boss absolutely needs to understand that this is how to get you to work most efficiently.
'On the web' is not the same as 'part of the web'
Sigh. Now you are just trying to weasel out of the argument by playing with semantics.
A web page is on the web. A web page is part of the web. A mailbox is on the web. A mailbox is part of the web. A newsgroup is on the web. A newsgroup is part of the web.
In this context, "on" and "part of" are synonyms. Here's a quote:
The Web is a universe of resources. A resource is defined by [RFC2396] to be anything that has identity. Examples include documents, files, menu items, machines, and services, as well as people, organizations, and concepts.
Note the words "The web is...". Not "The web points to..." or "The web contains...". He's talking about the web being directly comprised of the resources that are linked together. You know, in a web-like manner? In other words, leaf nodes are still nodes.
Arguably, mailto: URIs do put email 'on the web'.
So the original complaint that it was not proper to refer to Mozilla as a "web suite" is without merit - even if you draw a distinction between "on the web" and "part of the web". Right?
email functions quite happily without URIs.
So what? Something doesn't have to rely on URIs to be part of the web. If you point to a resource with a URI, you make it part of the web, whether the resource wants to be part of the web or not.
I could put a bus journey to my house 'on the web' by inventing a new URI scheme that allows 'bus://my-house?via=the_chip_shop'.
Your URI scheme wouldn't be recognised by anybody but you though.
The wouldn't make the bus, my house or the chip shop part of the Web.
If you had broad agreement on the bus URI scheme (e.g. had it ratified as an IETF standard), then yes, it would.
The web is an abstract concept, not a protocol or network connection. It links resources together. If your bus journey is defined as a resource, and it is pointed to by a URI, then it's part of the web.
Your mistake is thinking that "URI == part of the web" is the extreme part of your example. It's not. Defining the bus journey as a resource is the extreme part of of that example.
You: I don't deny that URIs are part of the Web. That doesn't make the things they point to part of the Web
Tim Berners-Lee: An information object is "on the web" if it has a URI.
I already pointed you to that quote. Try reading what I link to - it directly contradicts what you are saying. I'm not linking to random documents you know, they actually have relevence.
I'm saying it's almost universally accepted that 'World Wide Web' means HTTP+HTML.
This thread started because an AC felt the need to "correct" somebody's "misuse" of the term 'web'. Obviously it's not that universally accepted, especially as the guy who invented the term doesn't agree with the definition.
I would not imagine that he considers email to be part of the Web.
What's the point in me posting links if you aren't going to read them? The reason URIs are called URIs and not UDIs is because the web includes non-document resources like email addresses. Axioms of Web Architecture written by Tim Berners-Lee explicitly mentions mailto.
You probably think that the only way to beat terrorism is to adopt it.
I can't see how you can remotely think that from what I posted.
I think if we truly wanted to win the war on terror we would tie enemy combatants up and execute them in public.
Hmm... let's see, who does things like that? Oh yeah, the Taliban.
And you say I want to adopt terrorism? You are the one advocating we take a leaf out of the Taliban's books!
This is how it has always worked in the past.
Except it doesn't work, does it? If it had worked in the past, we wouldn't still be dealing with it, would we?
Why should we think that all of a sudden being considerate and gentle is a valid tactic?
Straw man. I never said we should be "considerate and gentle".
Since you seem to have completely missed the point of my comment, I'll use smaller words and make it clearer: It's proven fact that a subset of the Coalition forces will happily torture people. Therefore the parent's claim that this weapon is relatively safe because the Coalition forces are nice guys is of no value.
So, Sendmail is a Web server?
No. "Web server" is actually a common term for an HTTP server (there's the misconception that the WWW==HTTP again), Sendmail is not a web server, because the term "web server", as everybody knows it, refers to software that exclusively deals with the HTTP portion of the web.
But yes, as there is a mailto: URI scheme, email is part of the WWW, and you could call Sendmail part of the WWW infrastructure.
Hosts whose names start with www can reasonably be expected to speak NNTP?
Of course not. A hostname starting with 'www.' might denote that it is part of the WWW, but that doesn't mean that it must necessarily speak every protocol that comprises the WWW, any more than a hostname beginning with 'mail.' means that there must be a POP3 server, IMAP4 server, webmail server and SMTP server all running on it.
Are you actually disagreeing with Tim Berners-Lee, or are you saying that he doesn't really mean it when he explains what the WWW is?
Web Browser is web. Mail is not. IRC is not. News is not.
That's a common misconception. According to Tim Berners-Lee, the guy that coined the term "World-Wide Web", anything addressable by a URL (now URI) is part of the WWW. This includes Usenet, as there are nntp: and news: URI schemes.
More info, with cites.
it's scheduled for release in (brace yourself): "??? 2005".
They must be planning to profit somehow.
Have a little faith in the people serving on the front lines. We're professionals, just like you try to be at work. We care about honor, courage, commitment, etc.
It would be easier to have faith in you and consider you to be professionals if we hadn't seen photos of you guys torturing people and laughing about it.
You want us to give you the benefit of the doubt? Why should we? Some of you are good guys, sure, but some of you are evil bastards who shouldn't be trusted with a plastic fork, let alone anything like this. What, you want us to believe that only the "good guys" will be able to use this? Why would that be the case when the armed forces already allow the sickos access to guns, prisoners, etc?
Wonder if making people feel like they're being burned alive counts as torture?
If it leaves no marks on their bodies, does it make a difference? Call it torture or don't call it torture, nobody's going to get caught unless they are stupid enough to take photos of themselves doing it.
Just looking at my website statistics from people coming to my website via slashdot.org, I actually have a large number at the beginning of the workday, and towards the end, but during the day, it looks like most people stay pretty productive...
Or maybe it takes most of a day for your pages to expire from ISP's caches. You really shouldn't pay any attention to your logs, you can't count visitors with them, despite what many snake-oil sellers tell you.
What "beginning of the workday" are you talking about, anyway? You do realise that there is more than one timezone in the world, right?