Where does he mention the back button anywhere in his post?
What do you think he means when he says:
If they rank a page 1, and you click it, and immediately return to the search page, they penalize that page.
How would you immediately return to the last page you were on?
And if they did hit the back button and got a static page with no communication to the server they would have to click another link at some point, wouldn't they?
But if Google monitored clickthroughs for this purpose, then people like me who open half a dozen links from the results page in new windows/tabs all at once would screw up their calculations.
You are also forgetting the whole google bar thing.
Well no, as I specifically said that there was no reliable way of doing it. For instance, since the Googlebar only works with Internet Explorer, if they used it to track users in this way, it would bias rankings against sites that don't work well in Internet Explorer, but not bias rankings against sites that don't work well in other browsers (since only the first class would give negative feedback). Whether such biases are acceptable or not is obviously a matter for Google to decide, but as I said, it's a bit odd.
maybe if you thought out of the box for a second you wouldn't feel the need to doubt the original poster.
I thought it was odd, and asked for a reference. I then defended myself when somebody who missed my point entirely tried to patronise me. That is all. Nothing to do with "thinking outside of the box", I just wanted more information on it.
Before acting so condescending, please be familiar with the subject matter and think about what you are saying. It seems you are doing neither.
This way when you return to their site
My point is that clicking your back button shouldn't return you to their site in the way you are thinking of. It should show a stale copy - therefore no talking to the server, and no transmitting of cookies.
By the way, Google has attempted to acheive this concept of human ranking by watching to see how long you stay at a page you clicked on. If they rank a page 1, and you click it, and immediately return to the search page, they penalize that page.
Do you have a reference for that? According to the HTTP RFC, user-agents aren't supposed to talk to the server when users hit their back buttons, but rather display what the user last saw (despite it possibly being stale). It seems odd that Google would try to circumvent this, especially as there isn't a reliable way of doing so.
Lobby for draconian laws like the DMCA. You want OCR software to become tied up like DeCSS? They could always try printing their books on that "uncopyable" paper that they used for some computer game manuals in the early 90s:)
Hopefully, "book piracy" won't suddenly catch on. I suspect it will slowly rise, but a sharp increase will only prompt publishers to have a knee-jerk reaction and jump towards some kind of lock-down attempt. A slow increase will give publishers time to think about the most sensible way of altering their business model in the face of copyright infringement. Some have found that giving away electronic copies is profitable.
People don't really have structures in mind when they write, they arrange styles and whitespace to present their ideas as clearly as possible.
But if they are dividing their ideas up into discrete sections, that is structure, whether it's done with blank lines or <p> elements. The representation is less useful when it is meaningless outside of a human's perception though.
But here's a really simple counterexample: you're rereading something you just wrote, and you notice that the single space between two paragraphs is too narrow-- you want to widen it to express a logical jump, or a chronological one.
So should you have to go back and insert a new layer of structure between paragraphs and chapters?
The structure is already altered. You just need a way of expressing this in the written document. My argument is that basing this expression around a visual effect that means something to the author rather than an objective standard that is easy to learn is a poor representation of structure, as it's harder to process by a computer and not as useful in different contexts.
And you've just implied that books are not necessarily visual in nature. Do you have reasoning behind your assertion that the web is visual, or is it just that you've only ever experienced the web in a visual way?
It's millions of times cheaper and faster to have a single programmer update a popular screen reader or non-visual browser to look for that ID attribute in the tag instead of having millions of web designers in the entire world do a complete redesign the entire web site
"Looking for an id attribute in the tag" isn't any kind of solution. There just isn't anything sensible user-agents can do when faced with stupid markup, and even if there was, we'd end up with a completely screwed up language at the end of it. Consider this:
<img src="foo.png">
Now, without an alt attribute, what use is this to a user-agent that does not render images? What is the sensible thing to do here?
If that is what you are using as source material, no wonder you have the wrong idea. I've never seen so many blatant misrepresentations of the truth and other peoples arguments in one document before.
I always use ALT tags, but you know - I'll keep my nested tables, thanks. I see absolutely no reason to get rid of them when they've been working just fine for years.
Well that's your decision of course, but you are aware you are relying on non-standard behaviour, aren't you? There's nothing in the specifications that require a pixel-perfect representation of a table in the way you expect. That's the fundamental problem with abusing elements that denote meaning, rather than presentation.
What happens when browsers don't render tables the way they always have? The default look of table elements is quite dated now, perhaps browser vendors want to update the look a little. You think no browsers will ever stray from the table-layout tradition? They already have.
Imagine if there were 'security standards' that told you to leave your computer unplugged from the Internet, because it wasn't very secure.
Why should I have to replace my font tags and tables, that have worked absolutely well for *years*, with some new fangled technology that does the same thing?
You don't have to. It's just there are benefits to doing so. And I don't know where you get the idea that HTML does the same thing as CSS, they are completely different languages with completely different goals.
Dude, if you can't change the font of the site you're reading, try using a decent browser - maybe Internet Explorer.
Internet Explorer won't let you alter the font size when the author has specified it in pts or pxs. Doing the proper thing and using relative sizes makes things a lot better for many people. And no, I don't consider having to fiddle with my font settings every time I go to a new site to be pleasant.
Guess what? Plain old 'current' standards like HTML work well with the vast majority of browsers, are attractive, useable, and pretty damned accessible.
CSS 2 is five years old, it's used in many websites, I would describe it as a current, established technology. Whilst HTML-only sites may be attractive, usable, and accessible, this doesn't mean that this is due to the use of HTML, it means that the sites are that way despite this abuse of HTML.
This is one of my biggest bugbears with CSS design - creating pages that use the whole width of the browser window whatever the window size, and look good at all widths is just so hard.
Well that's just a general problem with web design, not with CSS, tables, or anything like that. At least with CSS you have a bit more flexibility, being able to use the width, min-width, and max-width properties, and being able to specify widths in a number of different units (ems, %, etc). You can "stack" columns with floating elements when the browser window gets narrow too.
Good design is hard, especially in a medium as flexible as the web.
They continued to ship the code in their distribution which from that point on puts it under the GPL.
No it doesn't. Two options exist:
SCO are distributing the code under the GPL
SCO are infringing on the copyrights of many people
Now, in the face of billions of dollars from a lawsuit against IBM, I can imagine them not particularly caring about infringing on copyrights, especially as most of the holders are unlikely to sue.
I don't know that KDE will ever attain the total degree of polish that XP has, simply because Linux/KDE is the result of efforts of hundreds of people working more or less independently
The average desktop distribution, however, is not. Using SuSE as an example, they don't blindly download the tarballs for KDE, compile them, and stick them into their distro. Part of their whole purpose is to refine the raw code into an end-user product.
Yes, it's good to have things polished by default, but the realities of open-source development means that these things often take a back seat to functionality. Getting a bunch of paid programmers to do the less fun stuff is taking advantage of the best of both worlds, commercial and non-commercial.
Making them conform to standards that tend to make sites look bad (by your own admission) just turns an already bad situation to shit.
Are you confusing me with somebody else? I am arguing that conforming to standards does not make sites look bad.
The rest of your argument is the tired old "lowest common denominator" straw man argument that has been refuted over and over again. I don't see the point in rehashing it once more, so just read this refutation instead.
I think you're making a cognitive claim, even there (about the cognitive relationship between structures and styles).
So you are arguing that HTML is bad because authoring in it isn't intuitive? This is a separate issue to the benefits of said structuring when applied as an interoperable file format. There is no requirement of actually authoring in HTML, you can author however you like as long as you can parse the information out of it appropriately with some program before publishing it.
But the 'value' you claim is derived purely from an arbitrary choice by an imaginary search-engineeer, so it holds no water for me.
There's nothing arbitrary about it. <em> means emphasis. <i> means italics. Following these definitions would obviously mean treating keywords residing within <em> elements as being more important. There's nothing imaginary about the search engineer, there are search engines out there that behave in this way.
There isn't a one-to-one relationship between <em> and <i> according to the specifications. It's fairly clear that lots of people have this confused, and ignore the specifications, but that doesn't mean there are no benefits to authoring correctly, or that the basic theory is unsound.
If someone authoring HTML by hand wants to skip some whitespace, multiple Ps (or BRs) is an entirely natural and reasonable way to do it.
You aren't thinking at the right level. The author doesn't want to add some whitespace because he has Enter Key Tourette's - there is a reason for adding the whitespace. The method of adding the whitespace depends on that reason.
For instance, if there is a menubar at the top, and the author wants to separate the text from the menubar, it makes far more sense to simply put a margin between the two:
#menu { margin-bottom: 3em; }
On the other hand, if the author merely prefers the look of spaced out paragraphs, it makes more sense to add margins to paragraphs:
p { margin: 1.5em; }
But, again, authoring in this style is not mandatory. It would be trivial to write a preprocessor that counted consecutive <br> elements, removed them from the stream, added an id to the following element, and included a { margin-top: xem; } in the stylesheet.
Goldfarb invented SGML, and initiated the theory that structures naturally precede styles.
I know who he is, I'm just unfamiliar with his Conjecture.
but the person you are replying to isn't talking about human cognition. It's about presenting content appropriately.
It's only appropriate if it conforms to human cognition.
But the presentation of the document to the end user is a completely different topic to the file format applications use to transfer information to each other.
All visitors do if the search engine they use pays more attention to emphasized text.
So you're trying to use this arbitrary choice to justify Goldfarb???
Please re-read my post, it seems you are arguing all by yourself. I'm saying nothing about Goldfarb. You asked if there was a benefit to using <em> for emphasis instead of misusing <i>. I pointed out one tangible benefit.
Series of BR elements should render as a single linebreak in conformant user-agents.
I'm sure you're thinking of Ps, but either way it's insanely anti-human. Multiple Ps should absolutely skip extra lines.
No, I'm thinking of the <br> element type. How is it "insanely anti-human"? End-users don't read the raw HTML, they care about how it is presented to them.
So... what kind of companies? All companies? Or just companies above a certain size?
In terms of the UK, feel free to read up on the law yourself. Or did you mean what should be the case? All the accessibility related laws I've heard about have some provision that it should be "reasonable" to provide services in an accessible manner.
And if all companies, you realise that screws small operations into the ground interms of their web presence costs.
Are you kidding? HTML is accessible by default. You have to actively screw up a document to make it inaccessible. The idea that accessible websites cost more to produce is an unfounded myth.
And if just above a certain size, what size? [...] Legislation should not be arbitrary.
Do you realise that the same argument would also apply to speed limits? Do you think the argument is valid?
Their clients (who do pay them) don't care about w3.org
They care when their website breaks. They care when they find out they are breaking the law (accessibility is legally required for many people).
I attemped to code to existing CSS "standards" 2 years ago only to discover that the browser interpretation of those "standards" was so unpredictable as to make the attempt pointless.
Times change. Netscape 4.x and Internet Explorer 4.x are close to being dead, which eliminates a hell of a lot of problems.
Actually, normal HTML is perfectly accessible to most people, including people with visual difficulties. It's when you start ignoring the specifications or adding extras like Javascript, CSS and Flash that you have the potential to screw up.
What spec are you talking about there? Last I knew Javascript(ECMAScript) and CSS were part of many specs.
I'm not following you. Normal HTML, such as HTML 4.01. Javascript, CSS and Flash are not part of normal HTML.
Ever heard of braille?
Braille isn't a visual medium, the visual artifact is a side effect of it's mechanism of transferral of information.
That was exactly my point. Read the exchange again:
I'm wondering what written text is, if not visual...
I've been hearing about accessibility and other potentially imposing guidelines for quite some time, and I've always been curious: is there any plan to try to enforce the guidelines?
There are many countries in which accessibility is a legal requirement for lots of organisations. For more information on these, please see WAI Policy.
But speaking of the private individual, should you and I also be subject to enforcement of web guidelines even in our personal, private web space?
I believe the most common point of view is that people who must cater to the needs of disabled people in the physical world must also do so on the WWW.
For instance, McDonald's are legally obliged to provide bathrooms that are specially equipped for people with mobility problems, at least in the UK. However private homes aren't required to provide them. It seems reasonable to draw the line at the same place on the web - so individuals would not be required to follow WCAG (or similar), yet service providers would.
Slashdot Mirror
What do you think he means when he says:
How would you immediately return to the last page you were on?
But if Google monitored clickthroughs for this purpose, then people like me who open half a dozen links from the results page in new windows/tabs all at once would screw up their calculations.
Well no, as I specifically said that there was no reliable way of doing it. For instance, since the Googlebar only works with Internet Explorer, if they used it to track users in this way, it would bias rankings against sites that don't work well in Internet Explorer, but not bias rankings against sites that don't work well in other browsers (since only the first class would give negative feedback). Whether such biases are acceptable or not is obviously a matter for Google to decide, but as I said, it's a bit odd.
I thought it was odd, and asked for a reference. I then defended myself when somebody who missed my point entirely tried to patronise me. That is all. Nothing to do with "thinking outside of the box", I just wanted more information on it.
Before acting so condescending, please be familiar with the subject matter and think about what you are saying. It seems you are doing neither.
My point is that clicking your back button shouldn't return you to their site in the way you are thinking of. It should show a stale copy - therefore no talking to the server, and no transmitting of cookies.
Do you have a reference for that? According to the HTTP RFC, user-agents aren't supposed to talk to the server when users hit their back buttons, but rather display what the user last saw (despite it possibly being stale). It seems odd that Google would try to circumvent this, especially as there isn't a reliable way of doing so.
Lobby for draconian laws like the DMCA. You want OCR software to become tied up like DeCSS? They could always try printing their books on that "uncopyable" paper that they used for some computer game manuals in the early 90s :)
Hopefully, "book piracy" won't suddenly catch on. I suspect it will slowly rise, but a sharp increase will only prompt publishers to have a knee-jerk reaction and jump towards some kind of lock-down attempt. A slow increase will give publishers time to think about the most sensible way of altering their business model in the face of copyright infringement. Some have found that giving away electronic copies is profitable.
But if they are dividing their ideas up into discrete sections, that is structure, whether it's done with blank lines or <p> elements. The representation is less useful when it is meaningless outside of a human's perception though.
The structure is already altered. You just need a way of expressing this in the written document. My argument is that basing this expression around a visual effect that means something to the author rather than an objective standard that is easy to learn is a poor representation of structure, as it's harder to process by a computer and not as useful in different contexts.
And you've just implied that books are not necessarily visual in nature. Do you have reasoning behind your assertion that the web is visual, or is it just that you've only ever experienced the web in a visual way?
"Looking for an id attribute in the tag" isn't any kind of solution. There just isn't anything sensible user-agents can do when faced with stupid markup, and even if there was, we'd end up with a completely screwed up language at the end of it. Consider this:
Now, without an alt attribute, what use is this to a user-agent that does not render images? What is the sensible thing to do here?
If that is what you are using as source material, no wonder you have the wrong idea. I've never seen so many blatant misrepresentations of the truth and other peoples arguments in one document before.
Well that's your decision of course, but you are aware you are relying on non-standard behaviour, aren't you? There's nothing in the specifications that require a pixel-perfect representation of a table in the way you expect. That's the fundamental problem with abusing elements that denote meaning, rather than presentation.
What happens when browsers don't render tables the way they always have? The default look of table elements is quite dated now, perhaps browser vendors want to update the look a little. You think no browsers will ever stray from the table-layout tradition? They already have.
Well CERT already advise switching off client-side scripting in web browsers.
You don't have to. It's just there are benefits to doing so. And I don't know where you get the idea that HTML does the same thing as CSS, they are completely different languages with completely different goals.
Internet Explorer won't let you alter the font size when the author has specified it in pts or pxs. Doing the proper thing and using relative sizes makes things a lot better for many people. And no, I don't consider having to fiddle with my font settings every time I go to a new site to be pleasant.
CSS 2 is five years old, it's used in many websites, I would describe it as a current, established technology. Whilst HTML-only sites may be attractive, usable, and accessible, this doesn't mean that this is due to the use of HTML, it means that the sites are that way despite this abuse of HTML.
They can. They just do it in a way you don't like, but in a way that is perfectly sensible when faced with properly-authored websites.
Well that's just a general problem with web design, not with CSS, tables, or anything like that. At least with CSS you have a bit more flexibility, being able to use the width, min-width, and max-width properties, and being able to specify widths in a number of different units (ems, %, etc). You can "stack" columns with floating elements when the browser window gets narrow too.
Good design is hard, especially in a medium as flexible as the web.
I take it you didn't bother reading the document then? Why should I even bother arguing? It just makes you look stupid and wastes our time.
Care to give an example?
Agreed.
For interoperability purposes.
Looking good and complying with W3C recommendations are not mutually exclusive.
No it doesn't. Two options exist:
Now, in the face of billions of dollars from a lawsuit against IBM, I can imagine them not particularly caring about infringing on copyrights, especially as most of the holders are unlikely to sue.
The average desktop distribution, however, is not. Using SuSE as an example, they don't blindly download the tarballs for KDE, compile them, and stick them into their distro. Part of their whole purpose is to refine the raw code into an end-user product.
Yes, it's good to have things polished by default, but the realities of open-source development means that these things often take a back seat to functionality. Getting a bunch of paid programmers to do the less fun stuff is taking advantage of the best of both worlds, commercial and non-commercial.
Are you confusing me with somebody else? I am arguing that conforming to standards does not make sites look bad.
The rest of your argument is the tired old "lowest common denominator" straw man argument that has been refuted over and over again. I don't see the point in rehashing it once more, so just read this refutation instead.
So you are arguing that HTML is bad because authoring in it isn't intuitive? This is a separate issue to the benefits of said structuring when applied as an interoperable file format. There is no requirement of actually authoring in HTML, you can author however you like as long as you can parse the information out of it appropriately with some program before publishing it.
There's nothing arbitrary about it. <em> means emphasis. <i> means italics. Following these definitions would obviously mean treating keywords residing within <em> elements as being more important. There's nothing imaginary about the search engineer, there are search engines out there that behave in this way.
There isn't a one-to-one relationship between <em> and <i> according to the specifications. It's fairly clear that lots of people have this confused, and ignore the specifications, but that doesn't mean there are no benefits to authoring correctly, or that the basic theory is unsound.
You aren't thinking at the right level. The author doesn't want to add some whitespace because he has Enter Key Tourette's - there is a reason for adding the whitespace. The method of adding the whitespace depends on that reason.
For instance, if there is a menubar at the top, and the author wants to separate the text from the menubar, it makes far more sense to simply put a margin between the two:
On the other hand, if the author merely prefers the look of spaced out paragraphs, it makes more sense to add margins to paragraphs:
But, again, authoring in this style is not mandatory. It would be trivial to write a preprocessor that counted consecutive <br> elements, removed them from the stream, added an id to the following element, and included a { margin-top: xem; } in the stylesheet.
I know who he is, I'm just unfamiliar with his Conjecture.
But the presentation of the document to the end user is a completely different topic to the file format applications use to transfer information to each other.
Please re-read my post, it seems you are arguing all by yourself. I'm saying nothing about Goldfarb. You asked if there was a benefit to using <em> for emphasis instead of misusing <i>. I pointed out one tangible benefit.
No, I'm thinking of the <br> element type. How is it "insanely anti-human"? End-users don't read the raw HTML, they care about how it is presented to them.
That's not true at all.
Accessible websites don't have to be ugly. What makes you think that they do?
This is already happening.
In terms of the UK, feel free to read up on the law yourself. Or did you mean what should be the case? All the accessibility related laws I've heard about have some provision that it should be "reasonable" to provide services in an accessible manner.
Are you kidding? HTML is accessible by default. You have to actively screw up a document to make it inaccessible. The idea that accessible websites cost more to produce is an unfounded myth.
Do you realise that the same argument would also apply to speed limits? Do you think the argument is valid?
For interoperability.
They care when their website breaks. They care when they find out they are breaking the law (accessibility is legally required for many people).
Times change. Netscape 4.x and Internet Explorer 4.x are close to being dead, which eliminates a hell of a lot of problems.
Well they don't have to compete with other sites, and their credibility is judged differently to commercial sites, so why waste the resources?
Then what is your point?
Why would you expect an accessibility requirement to improve the graphical appeal of a website?
I'm not following you. Normal HTML, such as HTML 4.01. Javascript, CSS and Flash are not part of normal HTML.
That was exactly my point. Read the exchange again:
There are many countries in which accessibility is a legal requirement for lots of organisations. For more information on these, please see WAI Policy.
I believe the most common point of view is that people who must cater to the needs of disabled people in the physical world must also do so on the WWW.
For instance, McDonald's are legally obliged to provide bathrooms that are specially equipped for people with mobility problems, at least in the UK. However private homes aren't required to provide them. It seems reasonable to draw the line at the same place on the web - so individuals would not be required to follow WCAG (or similar), yet service providers would.