Ask YC: Are Standards a joke?
What do you guys think ?
Google news has - 1310 Errors, 91 warning(s) http://tinyurl.com/5vne4e
Google.co.in has - 68 Errors, 13 warning(s)
http://tinyurl.com/5oznsp
News.ycombinator.com has - 136 Errors, 6
warning(s)
http://tinyurl.com/5l6ftf
Facebook has - 27 Errors, 6 warning(s)
http://tinyurl.com/62xcta
Yahoo has - 91 Errors, 64 warning(s)
http://tinyurl.com/5bqf9p
Youtube has - 70 Errors, 56 warning(s)
http://tinyurl.com/5gsap7
Live has - 4 errors
http://tinyurl.com/69x2ll
Blogger has - 28 Errors
http://tinyurl.com/5djo4q
8 comments
There is enough standardization happening to allow your browser to render those pages. That's a good start.
The problem historically with Web Standards has been 2-fold the lack of concrete implementations for vendors to work from. This leads to a variety of interpretations of the details of the standard.
Secondly their has been a lack of incentive for people to create standard markup. I hope things like Open Social (with Caja) and Yahoo's Search Monkey fix that. In the case of Search Monkey by using a Microformat (which must be valid HTML) then you get an enhanced listing in the Yahoo search engine. A definite benefit.
The problem historically with Web Standards has been 2-fold the lack of concrete implementations for vendors to work from. This leads to a variety of interpretations of the details of the standard.
Secondly their has been a lack of incentive for people to create standard markup. I hope things like Open Social (with Caja) and Yahoo's Search Monkey fix that. In the case of Search Monkey by using a Microformat (which must be valid HTML) then you get an enhanced listing in the Yahoo search engine. A definite benefit.
I didn't look at all the pages, but Google News' 1310 errors are caused in large part by the fact that they don't even declare a doctype. The use of the font tag (and not even quoting attribute values) is pretty embarrassing, but as previously stated, it's "good enough" for most browsers to render it properly.
There are dozens of very valid reasons for why that isn't really "good enough," but I won't get into those here. I don't think it's a question of whether or not standards are a joke, I just think there is disagreement among people as to whether or not they're worth adhering to. Having a few errors is no big deal; using deprecated HTML tags is a big deal (or at least I assume it will be at some point). I don't think the standards were created under the assumption that everyone would adhere 100% to them; that simply requires more effort than most people are willing to put forth.
This is further compounded by the fact that, yes, browsers do generally properly render even horrific markup such as that on Google News. I would argue that this is only done because it's a necessity and it shouldn't be viewed as a license to create deplorable markup, but what do I know. Until such a time that the advantages of proper standards (and extensions of them, such as microformats) become real benefits to people/businesses, many (most?) will continue to create junk markup -- because they're simply lazy, want to save a few bytes on the page load, or whatever other justifications they come up with.
That doesn't mean standards are a joke, it just means the benefits are too few, too poorly understood, or perceived as too insignificant (who cares about blind people on the internet?!?!)
There are dozens of very valid reasons for why that isn't really "good enough," but I won't get into those here. I don't think it's a question of whether or not standards are a joke, I just think there is disagreement among people as to whether or not they're worth adhering to. Having a few errors is no big deal; using deprecated HTML tags is a big deal (or at least I assume it will be at some point). I don't think the standards were created under the assumption that everyone would adhere 100% to them; that simply requires more effort than most people are willing to put forth.
This is further compounded by the fact that, yes, browsers do generally properly render even horrific markup such as that on Google News. I would argue that this is only done because it's a necessity and it shouldn't be viewed as a license to create deplorable markup, but what do I know. Until such a time that the advantages of proper standards (and extensions of them, such as microformats) become real benefits to people/businesses, many (most?) will continue to create junk markup -- because they're simply lazy, want to save a few bytes on the page load, or whatever other justifications they come up with.
That doesn't mean standards are a joke, it just means the benefits are too few, too poorly understood, or perceived as too insignificant (who cares about blind people on the internet?!?!)
It is perfectly legal in HTML to not quote attributes, and the FONT-tag is also a perfectly legal HTML-tag.
Purely presentational markup (like the FONT-tag) is legal and not a problem. It is only a problem (e.g. for accessibility) when information is conveyed only through presentational markup, or when semantic markup are misused for presentation (eg tables for layout or blockquote for margins). But validation cannot detect these kinds of problems (because it requires an understanding of the meaning of the document), so validation is pretty useless to determine if the HTML is "good" or "bad".
Validation is highly overrated as a means of determining the standards-compliance of a web-site.
Purely presentational markup (like the FONT-tag) is legal and not a problem. It is only a problem (e.g. for accessibility) when information is conveyed only through presentational markup, or when semantic markup are misused for presentation (eg tables for layout or blockquote for margins). But validation cannot detect these kinds of problems (because it requires an understanding of the meaning of the document), so validation is pretty useless to determine if the HTML is "good" or "bad".
Validation is highly overrated as a means of determining the standards-compliance of a web-site.
On the web real standards usually evolve out of best practices of hackers. The "standards" set by standards bodies are standard in name only, for the most part. They publish a large specification and people keep what is useful and discard what sucks. Or if you are Microsoft, Apple, Mozilla, etc you create your own proprietary stuff.
I used to worry about standards until I realized that it costs time and money when and as long as the browsers are following a single standard its a waste of energy. Now if it works in all the browsers I push it live, standard or not. If it breaks, I fix it and then push it again.
It is all a big coordination game really. An invisible hand works well enough in the web marketplace of ideas.
I used to worry about standards until I realized that it costs time and money when and as long as the browsers are following a single standard its a waste of energy. Now if it works in all the browsers I push it live, standard or not. If it breaks, I fix it and then push it again.
It is all a big coordination game really. An invisible hand works well enough in the web marketplace of ideas.
Validation ≠ Standards
All of these parsing 'errors' have clear and straightforward ways to handle them — to the point where the browser can handle an unclosed <p> much faster than it can download and parse '</p>'.
What's worse is that making a parse tree out of markup was never a problem! Consistent CSS layout has always had problems in all rendering engines, but parsing has never been problematic in the slightest.
Microsoft was absolutely right to cockblock XHTML — removing all parsing exception handling is so damn counter-revolutionary. It's pathetic that Microsoft understands the web better in this case than the W3C.
All of these parsing 'errors' have clear and straightforward ways to handle them — to the point where the browser can handle an unclosed <p> much faster than it can download and parse '</p>'.
What's worse is that making a parse tree out of markup was never a problem! Consistent CSS layout has always had problems in all rendering engines, but parsing has never been problematic in the slightest.
Microsoft was absolutely right to cockblock XHTML — removing all parsing exception handling is so damn counter-revolutionary. It's pathetic that Microsoft understands the web better in this case than the W3C.
An unclosed P is not an error. It has always been perfectly legal HTML and has an unambigous parse. There are a lot of cases in HTML where you are allowed to leave out tags, and the paser can infer then. This is part of the SGML nature of HTML.
However browsers also have graceful handling of invalid and ambigous syntax, and that is a much more controversial issue, because the resolutions have been browser specific and undocumented. This is obviously bad for interoperability.
Draconian error handling is the simplest possible solution to this problem, but HTML5 has a much better (but also much more complex) solution which is to openly specify rules for handling any conceivable syntax error. Microsoft has not peen part of this effort AFAIK.
However browsers also have graceful handling of invalid and ambigous syntax, and that is a much more controversial issue, because the resolutions have been browser specific and undocumented. This is obviously bad for interoperability.
Draconian error handling is the simplest possible solution to this problem, but HTML5 has a much better (but also much more complex) solution which is to openly specify rules for handling any conceivable syntax error. Microsoft has not peen part of this effort AFAIK.
Dont confuse standards and validation. Web standards are of course critical for the workings of the web. Validation is highly overrated though.
The reason people get hung up about validation is that it can be performed mechanically and gives an unambigous result. However it only check against a small (and pretty unimportant) part of the specs. For example a validator cannot check if you use semantic markup (which is pretty important), but will complain about superficial and harmless errors like leaving out the type attribute on a script tag.
The reason people get hung up about validation is that it can be performed mechanically and gives an unambigous result. However it only check against a small (and pretty unimportant) part of the specs. For example a validator cannot check if you use semantic markup (which is pretty important), but will complain about superficial and harmless errors like leaving out the type attribute on a script tag.
Not when you are developing for a government client that could hypothetically get sued if you don't use them. They don't find that possibility funny.
Standards are like international law -- only half the countries follow half the laws half the time.
Standards are crucial to the rendering of all those pages -- but de facto standards, rather than de jure standards.