Hi All,
I'm still unsure about my diagnosis and fix of the problem I was
seeing with John's metadata plugin. His idea of modifying a core
jQuery function to allow for automated metadata extraction was pure
genius in my mind. It was also very instructive for me as a person
trying to gain more
Meta Plugin possible bug:
The meta plugin appears to assume that the old $().get function will
always return a jQuery object because it calls apply to the cached
function _get and calls each on the returned object. This doesn't
seem to be the case if get is called via the methods described in
, Jörn Zaefferer [EMAIL PROTECTED] wrote:
Original-Nachricht
Datum: Thu, 2 Nov 2006 02:08:26 -0500
Von: John Resig [EMAIL PROTECTED]
An: jQuery Discussion. discuss@jquery.com
Betreff: Re: [jQuery] jQuery Metadata Plugin
From John's example, the following classes could
Keep in mind that there are characters allowed in classes that are not
allowed in the CSS spec for selection of classes. In other words, if you
include a paren or a bracket in a class, it can't be styled. Therefore,
you can use it for script, without worrying that someone will accidently
style
Discussion. discuss@jquery.com
Betreff: Re: [jQuery] jQuery Metadata Plugin
From John's example, the following classes could be applied:
someclass, {some: (illegal), 'data'} (illegal), anotherclass
Yeah, the whole embedding-data-in-classes idea is just messy to begin
with. Another option
Dave Methvin schrieb:
I got on the w3c validator discussion list because of
target being dropped from xhtml 1.1
...
I changed target into class=target somewhere
and let jquery do the rest.
Were you trying to open a new window on some links? One easy way to do that
is to set the target
yes, rel=external is a great, 90% , except dealing with multiple
windows... but that's neither here or there... lots of work arounds.
On 11/2/06, Dave Methvin [EMAIL PROTECTED] wrote:
I got on the w3c validator discussion list because of
target being dropped from xhtml 1.1
...
I changed
On 11/2/06, Ⓙⓐⓚⓔ [EMAIL PROTECTED] wrote:
sure, slip in a different namespace and you can make your own tags
but it won't validate unless you do the dtd for it too.
I'm wondering, about validation, does a page must validate before or
after it is parsed by javascript? Curently, w3c
hee hee... the validator is not involved is serving the page, it's an
additional check that we use to 'validate', it doesn't do javascript,
that's how we all get the monkey patching!
Any browser can act as a screen reader... so they might do the
javascript... they don't do css except the aural
Very nice!BlairOn 11/1/06, John Resig [EMAIL PROTECTED] wrote:
Hi Everyone -I just finished reading over the massive thread that discussesembedding metadata into elements for later extraction. Well, I justfinished a plugin to handle all three metadata-extraction methods.
You can see a demo here:
John Resig schrieb:
Hi Everyone -
I just finished reading over the massive thread that discusses
embedding metadata into elements for later extraction. Well, I just
finished a plugin to handle all three metadata-extraction methods.
You can see a demo here: (Uses Firebug debug statements!)
of course the data option is easiest but will break validation
the class works too, but what if you define a css class called some or
even data?
script is the most verbose but will work every time and with
type=application/json it sure looks like the best solution.
On 11/1/06, Jörn Zaefferer
In the case of adding metedata to divs, are these approaches any better than just using nested hidden spans?--ErikOn 10/31/06, John Resig
[EMAIL PROTECTED] wrote:Hi Everyone -I just finished reading over the massive thread that discusses
embedding metadata into elements for later extraction.
it was pointed out that hidden text fields may prove problematic to
non-visual presentation.
On 11/1/06, Erik Beeson [EMAIL PROTECTED] wrote:
In the case of adding metedata to divs, are these approaches any better than
just using nested hidden spans?
--Erik
On 10/31/06, John Resig [EMAIL
of course the data option is easiest but will break validation
the class works too, but what if you define a css class called some or
even data?
I don't follow. The class technique doesn't break existing
addClass/removeClass functionality - everything works as you might
expect.
script is the
I don't think it will break jquery. but by adding code into a class
css will still treat each word as a class and apply that style, for
your example, it would invoke a class called some or a class called
data.
On 11/1/06, John Resig [EMAIL PROTECTED] wrote:
of course the data option is easiest
A short explanation of this module and what it is for, appreciated. This
is way over my head but seems to be very useful if I understand it. Is it
a way to add a data= attribute to any HTML construct and then have
jquery give JS code access to the content of that attribute?
--Jacob
of
I don't remember the original thread, but a developer wanted a way to allow
designers to pass arguments to a jQuery plugin to create a widget. Which was
activated by assigning an agreed upon class name.
--
View this message in context:
Or just adding an xml data island to your document that has everything
you want it it and then parsing that using xpath. This is approach is
much cleaner, should validate, and we already have the tools to parse it.
-Steve W.
Erik Beeson wrote:
In the case of adding metedata to divs, are
sure, slip in a different namespace and you can make your own tags
but it won't validate unless you do the dtd for it too.
On 11/1/06, Stephen Woodbridge [EMAIL PROTECTED] wrote:
Or just adding an xml data island to your document that has everything
you want it it and then parsing that
Or just adding an xml data island to your document that has everything
you want it it and then parsing that using xpath. This is approach is
much cleaner, should validate, and we already have the tools to parse it.
The problem with that technique is that, in the end, you still have to
do a
Oops, I forgot about that.
The guys at work do this, but they don't use a dtd and in fact don't
even use a doctype and clearly don't validate. Oh well, it was a
thought, if only half baked. :)
-Steve
Ⓙⓐⓚⓔ wrote:
sure, slip in a different namespace and you can make your own tags
but it
hey it's a thought thats why we discuss before committing to an idea.
of course the old tried and true method of a simple class and doing a
switch in the jq code always has been available. Much better if you
have many identical 'meta-infos'.
On 11/1/06, Stephen Woodbridge [EMAIL PROTECTED]
This is great John and will be great for the community and widget
developers! I hope this will be put in SVN soon.
--
Brandon Aaron
On 10/31/06, John Resig [EMAIL PROTECTED] wrote:
Hi Everyone -
I just finished reading over the massive thread that discusses
embedding metadata into elements
Ⓙⓐⓚⓔ schrieb:
I don't think it will break jquery. but by adding code into a class
css will still treat each word as a class and apply that style, for
your example, it would invoke a class called some or a class called
data.
From John's example, the following classes could be applied:
Ⓙⓐⓚⓔ schrieb:
sure, slip in a different namespace and you can make your own tags
but it won't validate unless you do the dtd for it too.
If I got that right, the W3C validator ignores any custom DTD you give
him, therefore they won't help to validate your doc.
--
Jörn Zaefferer
I got on the w3c validator discussion list because of target being
dropped from xhtml 1.1
the only suggestions were to use a different dtd, there were samples
thrown around, none to my liking, so I changed target into
class=target somewhere and let jquery do the rest.
alernative DTDs is such a
exactly, they would be applied but the idea was not to apply css
styles , but to piggyback code into the class. John's class will also
apply an innocent unrelated class with each of those names.
style
.data { font-family: monaco,courier,monospace; }
/style
would apply
On 11/1/06, Jörn
From John's example, the following classes could be applied:
someclass, {some: (illegal), 'data'} (illegal), anotherclass
Yeah, the whole embedding-data-in-classes idea is just messy to begin
with. Another option that I plan on working in is so that you could
add classes in the format:
Erik Beeson schrieb:
In the case of adding metedata to divs, are these approaches any better
than just using nested hidden spans?
Yes, this was discussed exhaustively in the other thread and brandmarked
(by me) as Obtrusive 2.0.
Aaaah, not again...
-- Klaus
Sorry, I missed that. New here :) I think I like the embedded script approach. The data and custom class approaches seem so hackish (in a broken sort of way).--ErikOn 11/1/06,
Klaus Hartl [EMAIL PROTECTED] wrote:
Erik Beeson schrieb: In the case of adding metedata to divs, are these approaches
31 matches
Mail list logo
