Re: Nearly The Whole Of The Internet Is NOT W3C Valid

Alberto wrote: 
But Rob, this is not something between me and you. I understand I
started the thread taking as an occasion your reply, but the scope we
are dealing with, and which merely springs from that occasion, spans
beyond it by such a degree and in so obvious a manner, that in no way
you should misread it as something between you and me.

I also understand that you may think terms like "uncorrectly" are in a
non formally correct spelling, but you also have to realize that here
on these groups, which are open to an international audience and that
so many persons from all around the globe read, it is just impossible
that we are all native english speaker. I am not in fact, so you should
be indulgent with my possible mistakes, and you shouldn't use my
possible grammatical mistakes as an argument to gain a point in a cause
that cannot be won anyway.

That was not my intent, I thought you were attributing its use to me - clearly a misunderstanding.



I am not here to invent a thesis. Like yourself, I am too a person who
is adult and who can produce an intellectual effort in order to stress
the importance of something that is corroborated by facts, not
uncorroborated. As said I do not mean I can be beyond critics, but I
find it intellectually impossible to dismiss a list like the one I
provide upon the bases you attempt to propose.

What I criticise is that conclusions are drawn from the presentation of the list that are not supported by a more rigorous analysis.



Mozilla validated, true: yet, as they can surely confirm to you if they
are honest, when I did the list on the nefarious day sept 11 2005, it
didn't. Now, in such long a list, if all we are left with is the
desperate search of a link incorrectly listed as invalidated so to
uphold it as an alleged disqualification of the  whole list, we have
been left with very little, and we are clearly scrambling for a
desperate line of defence at the bottom of the barrel. If we are left
with that, we are actually proving the strenght of the list, NOT its
weakness.

We have a list that not only is long, but that by its very same length
could have been longer.
Most importantly, the names listed in such list are stunning ones: they
are not a part of the internet, they are the internet.


They are web sites, not 'the internet'.

Can you reasonably criticise Apple for having 3 trivial errors? Or even MSDN their 70 or so and from that draw the conclusion that standards compliance doesn't matter?

Does the fact that perhaps 95% (complete guess) of markup *is* compliant count for nothing?



Now, relieved of any burden of proof as you deem yourself while,

The burden of proof is yours, not mine. You are proposing a theory, I am saying your 'proof' is insufficient.

[...]

The W3C is not God. But you deal with it as if it were.

Not at all, not ever. I have only ever used 'W3C' in the context of their standards, I don't think I have ever commented on their competence or omnipotence.


The W3C is wrong. That list proves it.

It proves nothing of the sort! Are the police wrong because crime still occurs? Should all laws be thrown out because everyday people break laws every single day?

Have you never, *ever* broken any law? No matter how trivial?

That is the standard you would hold W3C standards too. Incidentally, there are a number of fundamental internet and web standards that are not controlled by the W3C, ECMAScript being an example.


We cannot declare invalid the
whole of the world after rules we ourselves made and clearly nearly
everybody violates, and evryobdies that MATTER, and at the same time
say that we are right. It's a tautology that has to be corrected.
The W3C validation, as it is now, makes NO sense.

And all the chicanery of the world can't change this fact, which that
liost proves beyond any doubt, beyond any reasonable doubt, befeore
whateve court of unbiased men and women.

Your basic premise is that complaining about invalid sites is pointless because most sites are invalid. I can accept that as a point of view.

In order to argue the point, I would:

1. Define what 'standards compliant' means. Are we talking just HTML? Or are CSS, DOM, ECMAScript included? Most readers of this forum would include at least those when discussing the standards compliance of hosted pages. Let's restrict ourselves to HTML as defined by the W3C.


2. Establish a framework for determining the relevance of non-compliances. For example, the use of a single deprecated tag in a page of several hundred tags should be treated as trivial, whereas forgetting mandatory closing tags or incorrect nesting of block elements inside inline elements is much more serious.

Then you respond to individual complaints of 'Oh, Google isn't compliant because of ...' by indicating whether it really matters or not.

Of course purists will never be satisfied with anything less than 100% compliance (that goes beyond simple DTD validation), but the vast majority of surfers will be happy with 'fit for purpose' compliance.


3. Determine a good compliance methodology. The W3C validator only checks against a DTD and does not correctly report some markup (e.g. HTML inside script document.write statements). This results in spurious reporting of errors - validator results *must* be analysed further.


4. Having regard to the above, determine the consequences of compliance/non-compliance for various classes of common errors.

The usefulness of standards compliance can only be evaluated against its alternatives - the consequences of not being compliant. In the extreme, without standards the web would not exist, so we are only discussing what level of compliance is reasonable.



ciao, and btw it has been a pleasure to talk with you 

Likewise, cheers.


[...]


--
Rob
.

Relevant Pages

  • Re: ISO 27001 mapping to PCI
    ... Where standards 'overlap' might create opportunity for enterprise to have a policy that minimizes this complexity and thereby help promote ongoing compliance and mitigate maintenance. ... Unless the ISO 27001 accreditation is scoped to "only" systems that collect or process payment card data, ... Any views expressed in this message are those of the individual sender and not necessarily endorsed by BDO Kendalls. ...
    (Security-Basics)
  • Re: Why will we never see anything even remotely resembling decent standards compliance in SQL Serve
    ... > it is a considerable undertaking to cram in all the features that would ... *Anyone* who writes any code for SQL Server should demand standars compliance ... SQL Server already uses such mechanisms to govern code-breaking standards ...
    (microsoft.public.sqlserver.server)
  • Re: Websites that dont work with Linux
    ... Colin Brough wrote: ... compliance and why they shouldn't write for one OS/browser? ... it doesn't comply with the relevant standards. ...
    (uk.comp.os.linux)
  • Re: Free Microsoft Visual C++ Toolkit 2003
    ... good compliance to C95 for over a decade, but they have done very little ... I thought they had got a lot better with C99 compliance of late? ... *not* a rubber stamp of the product Microsoft is now widely shipping. ... ECMA/ISO standards I am not sure if you would say that MS is standards ...
    (comp.lang.c)
  • Re: [fw-wiz] Stats on how common NAT is?
    ... network topology, which connects to the Internet, IMHO. ... This is a poor response to meet HIPAA compliance ... balance vulnerabilities, threats and attacks with risks, ...
    (Firewall-Wizards)