Re: validation in Opera

Håkon Wium Lie <howcome@opera.com> wrote:

>Dear validators, I come in peace.

:-)


>As you may know, the Opera browser offers to validate pages by
>uploading the source to the W3C validator. [...] This feature
>doesn't work very well at the moment. There are two problems:
>
> 1) lack of DOCTYPE in documents, and
> 2) lack of <meta http-equiv
>
>In the case of 1), the validator will refuse to process the
>document and the retuned page has no options for revalidation
>(which it has when the URL is sent). I think this problem could be
>improved if there was a way for us to tell the validator "please
>look for a DOCTYPE in what we send you, if you can't find any use
>HTML 4.01 transitional".
>
>2) is slightly trickier. Most pages include this information in the
>HTTP header, but Opera does not pass this information along with
>the source. Is there a way for us to do so? One that would be
>overridden by the META tag, if found?

What you are asking for is in essence that the Validator provide defaults
for these two values when one is not explicitly given. We've considered
this long and hard and come to the conclusion that this is the least
desireable solution; it gives users little or no incentive to fix this.

I might be persuaded that "hidden" options, that merely provide a default
fallback if the information is unavailable, is defensible. But I worry 
that
this would be taken as license to include these options in "This Page is
Valid Foo" links.

I do very much wish to facilitate the inclusion of this type of feature in
browsers and authoring tools[0]. This issue seems to stem from differing
contexts mainly; in the context of a browser and its current page these
options make perfect sense, but in the Validator -- in this sense an
application in it's own right -- this is exactly the kind of thing we do
not want to do. I'm not sure there is any way to reconcile the two...


However we are looking at other options for the future. Right now you can
get (beta) output in generic XML (and EARL/RDF and Notation3 for that
matter) from which you could pick whatever information you wanted, and for
the future we are looking at ways to provide a SOAP interface to the
service.

At that point the Validator is no longer an "Application", but rather an
API; letting the user of that API (e.g. Opera) pick and choose what
warnings and information s/he wants would be appropriate.

Would either of these options be of interest to you?




[0] - So please do let us know about anything we can do. I wasn't aware
      of the issue you brought up here so I've given it little thought
      as yet. It's a new perspective that's tremendously usefull to us.



-- 
Now Playing "Fimbul" by "To Rustne Herrer"",
 from the album "Damebesøk".


--------------------------------------------------------------

I know that this is getting a long email - I think it's very important.

I can see the argument from Terje and know where Håkon's coming from - I 
user Opera all the time and am promoting it as a tool for accessibility. I 
also teach computing and web programming in particular.

A suggestion: And it involves Opera or anyone else who writes inbuilt 
links to validators rather than W3C at this stage.

Is it possible to, when submitting a page to a validator that the 
submitting programme could, on request, insert a !DOCTYPE before 
transmitting the HTML that you're trying to check? Here I'm assuming that 
its someone else's page that's being submitted otherwise it would be 
correct in having a correct !DOCTYPE declaration.

This way strange pages could be validated other than the missing !DOCTYPE 
to see just where they're going wrong.

Or have I got the wrong end of the stick?

John

John Colby
Lecturer, School of Computing, Faculty of Computing, Information and 
English
Room F328a, Feeney Building, University of Central England,
Franchise Street, Perry Barr, Birmingham B42 2SU
Tel: +44 (0) 121 331 6937, Fax +44 (0) 121 331 6281, Mobile: 0771 114 1621