First, most webpages are not structured into paragraphs and chapters
Your quote was about authors - people who write. (Clue: the word "author"). People who write for a living structure their work. So do people who seriously write for the web. We are not talking about webpages devoted to pet cockroaches here.
really simple counterexample...
Improving a logical structure does not mean the structure wasn't planned out in the first place. It just means an improvement to the structure was made. Sheesh, this ain't a binary construct!
But in actual fact, the style comes first, and the structure is a hypothetical afterthought.
This is quite false, as over a millenia of history has already proved. I've already explained this to you elsewhere on this page, and you have not produced anything remotely approaching proof of your ridiculous assertion.
Right. This is like saying "A car is not a means of transportation.
Wrong. Your argument is like saying that only cars use roads. Clearly false, since all manner of wheeled vehicles use roads - whether or not you personally have used one.
Mediocre authors won't mind, perhaps, but serious authors should care how the document looks, and not surrender the details to a mechanical whim.
Again, this is not even reflected in the real world. An article writer writes structured articles for publications. It is the designers job to style it as best fits the publication, not the author. This is what happens in the real world.
However, I don't think you can deny that if people would write structured markup I would be extremely useful. Jim's example that search engines could use it to rank hits by looking at the content of em elements (AltaVista did this in the old days, when people still wrote reasonably structured), is a good one.
Google is experimenting with a number of structual-based searches. On http://labs.google.com there are two searches based on structural markup.
Google Sets is based on list structures.
Google Glossary is based on definition lists, abbreviation, acronym and definition markup.
Your wish for a digest of articles is structurally nothing more than an outline created using the header elements.
The benefits of structural markup are there, there's just not enough structural markup around yet.
People don't really have structures in mind when they write
This is clearly not true. Every serious author has a plan of the structure before they start writing. Fiction is well known for the preplanning work required to structure the story around plots and subplots, plus the detailed working out of characters before the first chapters are written.
Two massive examples. Tolkien worked out most of Middle Earth before he started writing Lord of the Rings. He worked out life and time lines, histories and myths. Even invented languages and alphabets and songs before the first chapter was even written. The story followed a pre-set plan.
Example 2 is David and Leigh Eddings Belgariad and Mallorean sagas. All the preplanning and structure was described and laid out in their complementing work: "The Rivan Codex". They spent months detailling their created world working out histories, major events and timelines.
Most typical writers start their articles by creating an outline first. When thet write, they stick to this outline. Whatever James Joyce decides to do is not what typically happens in the real world.
Goldfarb invented SGML, and initiated the theory that structures naturally precede styles.
False. This practice of structuring first and styling after predates Goldfarb and even the print industry. I can trace an immediate path of this practice all the way back to the Middle Ages. I expect pre-historians and archeologists can trace it back even further. Even to way before the early Egyptian civilisation. Its nothing new.
that styles should naturally be layered on top of these semantics is just deluded, imho
Can you provide a cohesive and plain English explanation as to why. Certainly the entire print industry is fashioned along the same concept, and they seem to be running along all fine and dandy.
The best approach to semantic markup would even be visible to screen scrapers
An infeasible and unworkable solution with no realistic possibility of existing within our lifetime. The solution you mock (of semantic markup) is one that works right now and delivers benefits to users, businesses and governments today.
Better something that does the job today over something that will never exists. Pragmatism!
Topic usually doesn't change within a sentence, so I like to add a 'text button' at the end that augments the topic-info with the resource-type. [examples] [robotwisdom.com]
This breaches WCAG2.0 Guideline 3.3.1.d which requires that linked text make sense out of context. No doubt, you will say that people don't use links out of context, and this has already been proven false. So your theory is widely considered to be an accessibility obstacle.
Naturally, you will recognise that WCAG2.0 is not yet a technical recommendation, so now look at Checkpoint 13.1 of WCAG 1.0 (which became a technical recommendation in 1999), which requires "Link text should be meaningful enough to make sense when read out of context -- either on its own or as part of a sequence of links.".#
This means that when the links from your page is presented as a list by themselves, having multiple links identified with [examples] all pointing to different resources is an obstacle to accessibility which makes it very difficult (priority 2) for disabled people to use your pages.
So your theory is inaccessible, and should be avoided.
But are there ever semantic/structural/stylistic units that overlap paragraph/section/chapter boundaries?
Paragraphs, sections and chapters are structural units. Even a book is a structural unit. A page is also a structure. Sentences are structure, they contain a linear sequence of words. Even words themselves have structure. Language also has structure.
Without structure you have nothing that can be efficiently reused or passed on.
Even the priests of the Middle Ages understand the value of structured information.
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.
Empty paragraphs are meaningless - much like your "Goldfarb's Conjecture". Its a pity your "proposed solutions" are nothing more than unrealistic over-hyped expecations largely derived from the wasteland that is Artificial Intelligence. In case you hadn't noticed, people are rather tired of waiting for you AI wunderkinds to produce something the real world can use right now.
Your jealousy of Tim Berners Lee is evident enough - pity you are incapable of producing something the real world can use. That's something Tim's already done, and a decade later its still globally used and still gaining popularity in every corner of the world.
When can we expect your little brain child to materialise?
There's no reason to burden the published page with authoring hints.
That is indeed your choice which can be forced for your ecstatic pleasure by disabling stylesheets in your browser and supplementing it with your own stylesheet with your preferences to render content exactly how you prefer it.
However, I disagree with you entirely, yet I don't force my choice on you. Please consider respecting the readers of website before continuing your ongoing puffery.
Do writers and/or readers really organise a text as a hierarchy of nested structural containers
*pulls a random book off a shelf* Look, each page has a page number. There's a table of contents at the start. There are even breaks called chapters and sections. Paging through, there's different type of headings, and there's even an index at the back.
Heck, the book can actually be read in a sequential fashion, reading the pages in an ascending fashion. Heck every page can be linearly read starting from the top-left, reading each line from left to right and top-to-bottom.
Is that isn't organised and hierachial structured text, then you don't know what is.
Looking at other books, wow - the same style and organisation.
Finnegan's Wake is just one book written by a drug-induced fogey. The world doesn't revolve around it.
I don't think this should be ENFORCED legally by anyone. That is just silly.
What is silly is that web designers can't produce quality and valid work, and so people have resorted to legal enforcement to get them to do so. People are merely protecting their rights to participate in society.
as long as the mainstream webpage building programs (Frontpage, Dreamweaver,...) don't catch up, web is going to be a horrid place for non-visual media.
So what creates websites, the tool or the user of the tool? If the tool is creating the website, why is anyone paying a web designer? If the web designer is creating the site using tools that produce inaccessible markup, why is the designer using a broken tool.
To me it is the web designer's responsibility to select tools that do the job. Don't blame the tool, blame the web designer - he should know better.
Now, dial up the the net using your modem. request a page, and now pick up your connected phone. Listen. Notice you can hear the data being sent, although you can't see it. The TCP/IP stack takes the packets and reconstructs the requested resource. At this point, you get to choose how this resource is represented. you can pick a graphical editor which takes the resource data and display it on a screen, or you can decide to pipe it into a speech reader which then reads it to you.
You choose to see the text, you can also chose to listen to it. It is not forced, and your choice cannot be forced on others.
Are you saying it should be compulsory for web pages to be written with accessibility in mind?
As opposed to mindless drivel posturing as logical arguments, of couse.
Do you also think it should be compulsory for all books to have braille and speaking book versions?
Ahhh, so you can't tell the difference between a web page and a book.
who is going to pay for all that?
Where on earth do you get off delivering a half-assed job, and then _expect_ to get paid to fix it. Why not do it right the first time. Designing accessible websites ain't difficult, and not expensive. Only if you cock-up badly does it cost a lot to fix. Perhaps its time for you to learn how to do your job properly now.
This really annoyes me. The web is a visual medium.
Using my browser, you sound like an idiot. The web isn't a visual medium, a visual representation of a web page is only one representation, and not the only representation.
Providing speaking book or braille versions of books doesn't break the textual books. Yet we want to break webpages. Why?
The web is a more flexible and accessible medium than print will ever be. The web doesn't suffer print's shortcomings of inflexible page sizes, inking processes, and a lot of rules that work well in the print world fail when used on the web. The print world is inflexible, the web is not.
What can be done on the web for very low cost is extortionately expensive for the print world. Since the web is a more accessible medium, it is easier to create an accessible information flow here than in its print cousins. So there is really no relevant comparison between print and the Web. The Web is designed to be completely accessible to its audience, print fails with the same resources.
If we want to make sites accessible, then make dual sites
Well, if you are prepared to pay twice for the same website, and still get sued because your accessible version is out of step with the inaccessible geek-gaw, then you only have yourself to blame. It is possible, as already pointed out, to deliver good looking functional websites that are also accessible.
Most cacheing systems (expecially AOL's) are a little smarter than that.
I didn't make myself too clear there! One visit from an AOL user produces a range of requests from more than one IP address - probably as a result of load-balancing on a collection of proxy servers. Whether AOL decides to cache the page or not matters not. Since each request may come from a different IP address than the original request, then tieing a session id to an ip address is going to declare the session invalid for that particular user.
A better solution would be to tie the session ID to an ip range big enough to include all the AOL proxies a particular user could come through.
Does anyone know what kind of heuristics MSIE used to determine which cookies are good and which are bad?
Internet Explorer 6 uses the Compact Privacy policy as specified in the W3C P3P spec. It uses this to determine whether a cookie is unsatsifactory (different rules based on whether it is a third party cookie or not). MSDN has documentation covering Internet Explorer's decision matrix (unfortunately framed).
Wouldn't this present a problem where the user is behind a proxy ?
Indeed it does. AOL for example uses a number of caching servers, and one user uses a number of different caching servers during his visit. So by tieing a session id to an IP address effectively prevents users of AOL and other large ISPs from using a website.
Slashdot Mirror
Your quote was about authors - people who write. (Clue: the word "author"). People who write for a living structure their work. So do people who seriously write for the web. We are not talking about webpages devoted to pet cockroaches here.
Improving a logical structure does not mean the structure wasn't planned out in the first place. It just means an improvement to the structure was made. Sheesh, this ain't a binary construct!
This is quite false, as over a millenia of history has already proved. I've already explained this to you elsewhere on this page, and you have not produced anything remotely approaching proof of your ridiculous assertion.
Wrong. Your argument is like saying that only cars use roads. Clearly false, since all manner of wheeled vehicles use roads - whether or not you personally have used one.
Again, this is not even reflected in the real world. An article writer writes structured articles for publications. It is the designers job to style it as best fits the publication, not the author. This is what happens in the real world.
Google is experimenting with a number of structual-based searches. On http://labs.google.com there are two searches based on structural markup.
Google Sets is based on list structures.
Google Glossary is based on definition lists, abbreviation, acronym and definition markup.
Your wish for a digest of articles is structurally nothing more than an outline created using the header elements.
The benefits of structural markup are there, there's just not enough structural markup around yet.
This is clearly not true. Every serious author has a plan of the structure before they start writing. Fiction is well known for the preplanning work required to structure the story around plots and subplots, plus the detailed working out of characters before the first chapters are written.
Two massive examples. Tolkien worked out most of Middle Earth before he started writing Lord of the Rings. He worked out life and time lines, histories and myths. Even invented languages and alphabets and songs before the first chapter was even written. The story followed a pre-set plan.
Example 2 is David and Leigh Eddings Belgariad and Mallorean sagas. All the preplanning and structure was described and laid out in their complementing work: "The Rivan Codex". They spent months detailling their created world working out histories, major events and timelines.
Most typical writers start their articles by creating an outline first. When thet write, they stick to this outline. Whatever James Joyce decides to do is not what typically happens in the real world.
False. This practice of structuring first and styling after predates Goldfarb and even the print industry. I can trace an immediate path of this practice all the way back to the Middle Ages. I expect pre-historians and archeologists can trace it back even further. Even to way before the early Egyptian civilisation. Its nothing new.
Can you provide a cohesive and plain English explanation as to why. Certainly the entire print industry is fashioned along the same concept, and they seem to be running along all fine and dandy.
An infeasible and unworkable solution with no realistic possibility of existing within our lifetime. The solution you mock (of semantic markup) is one that works right now and delivers benefits to users, businesses and governments today.
Better something that does the job today over something that will never exists. Pragmatism!
This breaches WCAG2.0 Guideline 3.3.1.d which requires that linked text make sense out of context. No doubt, you will say that people don't use links out of context, and this has already been proven false. So your theory is widely considered to be an accessibility obstacle.
Naturally, you will recognise that WCAG2.0 is not yet a technical recommendation, so now look at Checkpoint 13.1 of WCAG 1.0 (which became a technical recommendation in 1999), which requires "Link text should be meaningful enough to make sense when read out of context -- either on its own or as part of a sequence of links.".#
This means that when the links from your page is presented as a list by themselves, having multiple links identified with [examples] all pointing to different resources is an obstacle to accessibility which makes it very difficult (priority 2) for disabled people to use your pages.
So your theory is inaccessible, and should be avoided.
Gee, linking the relevant text to it perhaps.
Paragraphs, sections and chapters are structural units. Even a book is a structural unit. A page is also a structure. Sentences are structure, they contain a linear sequence of words. Even words themselves have structure. Language also has structure.
Without structure you have nothing that can be efficiently reused or passed on.
Even the priests of the Middle Ages understand the value of structured information.
And your non-existant fantasy totally dependant on the unmaterialised AI solutions is better how?
I can use the web right now to search for stuff, and it returns me a decent amount of well structured information. It works.
Empty paragraphs are meaningless - much like your "Goldfarb's Conjecture". Its a pity your "proposed solutions" are nothing more than unrealistic over-hyped expecations largely derived from the wasteland that is Artificial Intelligence. In case you hadn't noticed, people are rather tired of waiting for you AI wunderkinds to produce something the real world can use right now.
Your jealousy of Tim Berners Lee is evident enough - pity you are incapable of producing something the real world can use. That's something Tim's already done, and a decade later its still globally used and still gaining popularity in every corner of the world.
When can we expect your little brain child to materialise?
That is indeed your choice which can be forced for your ecstatic pleasure by disabling stylesheets in your browser and supplementing it with your own stylesheet with your preferences to render content exactly how you prefer it.
However, I disagree with you entirely, yet I don't force my choice on you. Please consider respecting the readers of website before continuing your ongoing puffery.
*pulls a random book off a shelf* Look, each page has a page number. There's a table of contents at the start. There are even breaks called chapters and sections. Paging through, there's different type of headings, and there's even an index at the back.
Heck, the book can actually be read in a sequential fashion, reading the pages in an ascending fashion. Heck every page can be linearly read starting from the top-left, reading each line from left to right and top-to-bottom.
Is that isn't organised and hierachial structured text, then you don't know what is.
Looking at other books, wow - the same style and organisation.
Finnegan's Wake is just one book written by a drug-induced fogey. The world doesn't revolve around it.
What is silly is that web designers can't produce quality and valid work, and so people have resorted to legal enforcement to get them to do so. People are merely protecting their rights to participate in society.
So what creates websites, the tool or the user of the tool? If the tool is creating the website, why is anyone paying a web designer? If the web designer is creating the site using tools that produce inaccessible markup, why is the designer using a broken tool.
To me it is the web designer's responsibility to select tools that do the job. Don't blame the tool, blame the web designer - he should know better.
No it isn't. This is a common mistake.
Section 508 is actually Section 508 of the Rehabilitation Act. Americans with Disabilities Act also covers the accessibility requirement. There is also the Section 255 of the Telecommunications Act of 1996 also lays out accessibility guidelines.
Section 508 applies to federal agencies and companies providing service to federal agencies
ADA applies to businesses and government organisations providing public accommodation - so basically offering a service to the general public.
Section 255 applies to telecommunications companies.
Three separate pieces of legislation all covering the topic of accessibility.
The rest of your post is excellent. Kudos.
The last one is most certainly incorrect. It should be:
Monitor. Visual.
Speakers. Aural
Refreshable Braille Display. Tactile
Now, dial up the the net using your modem. request a page, and now pick up your connected phone. Listen. Notice you can hear the data being sent, although you can't see it. The TCP/IP stack takes the packets and reconstructs the requested resource. At this point, you get to choose how this resource is represented. you can pick a graphical editor which takes the resource data and display it on a screen, or you can decide to pipe it into a speech reader which then reads it to you.
You choose to see the text, you can also chose to listen to it. It is not forced, and your choice cannot be forced on others.
As opposed to mindless drivel posturing as logical arguments, of couse.
Ahhh, so you can't tell the difference between a web page and a book.
Where on earth do you get off delivering a half-assed job, and then _expect_ to get paid to fix it. Why not do it right the first time. Designing accessible websites ain't difficult, and not expensive. Only if you cock-up badly does it cost a lot to fix. Perhaps its time for you to learn how to do your job properly now.
Using my browser, you sound like an idiot. The web isn't a visual medium, a visual representation of a web page is only one representation, and not the only representation.
The web is a more flexible and accessible medium than print will ever be. The web doesn't suffer print's shortcomings of inflexible page sizes, inking processes, and a lot of rules that work well in the print world fail when used on the web. The print world is inflexible, the web is not.
What can be done on the web for very low cost is extortionately expensive for the print world. Since the web is a more accessible medium, it is easier to create an accessible information flow here than in its print cousins. So there is really no relevant comparison between print and the Web. The Web is designed to be completely accessible to its audience, print fails with the same resources.
Well, if you are prepared to pay twice for the same website, and still get sued because your accessible version is out of step with the inaccessible geek-gaw, then you only have yourself to blame. It is possible, as already pointed out, to deliver good looking functional websites that are also accessible.
I didn't make myself too clear there! One visit from an AOL user produces a range of requests from more than one IP address - probably as a result of load-balancing on a collection of proxy servers. Whether AOL decides to cache the page or not matters not. Since each request may come from a different IP address than the original request, then tieing a session id to an ip address is going to declare the session invalid for that particular user.
A better solution would be to tie the session ID to an ip range big enough to include all the AOL proxies a particular user could come through.
Internet Explorer 6 uses the Compact Privacy policy as specified in the W3C P3P spec. It uses this to determine whether a cookie is unsatsifactory (different rules based on whether it is a third party cookie or not). MSDN has documentation covering Internet Explorer's decision matrix (unfortunately framed).
Start > Run
Telnet to ftp.mozilla.org:80 and type