By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
446,228 Members | 1,492 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 446,228 IT Pros & Developers. It's quick & easy.

link (rel and rev)

P: n/a
Hi,

After Brian mentioned the use for <link rel=..> for navigational
purposes in another thread, I've been looking into it and found that
HTML 3.2 has two other recognized link types than HTML 4.01, which are
"top" and "search". I compared these two pages:

http://www.w3.org/TR/REC-html32-19970114#link
http://www.w3.org/TR/html4/types.html#type-links

I suspect that "Start" in HTML 4.01 is the same as "top" in HTML 3.2, is
this correct?

Is there for HTML 4.01 something like "search" in HTML 3.2?

Since rel="prev" defines a link to a previous page, is this the same as
the rev attribute?

--
Regards
Harrie
Jul 23 '05 #1
Share this Question
Share on Google+
26 Replies


P: n/a
Harrie wrote:
Hi,

After Brian mentioned the use for <link rel=..> for navigational
purposes in another thread, I've been looking into it and found that
HTML 3.2 has two other recognized link types than HTML 4.01, which are
"top" and "search". I compared these two pages:

http://www.w3.org/TR/REC-html32-19970114#link
http://www.w3.org/TR/html4/types.html#type-links

I suspect that "Start" in HTML 4.01 is the same as "top" in HTML 3.2, is
this correct?
Not sure - when I was experimenting with these, I found browsers didn't
implement them consistently. I've ended up using both, but have
duplicated the links in the document body. See, for example:

http://tranchant.plus.com/web/html-tutorial/headings
Is there for HTML 4.01 something like "search" in HTML 3.2?
Pass. There's no reason why you can't use it anyway. Lynx will still
display it in its header bar.
Since rel="prev" defines a link to a previous page, is this the same as
the rev attribute?


I wondered that too. I think not, but I can't really explain why. <link
rev="made" ...> is a common way to list the author's mail address, and
Lynx uses that to send mail to the author if you press C (contact). See
the above URL for an example of that, too.

--
Mark.
http://tranchant.plus.com/
Jul 23 '05 #2

P: n/a
Harrie <di******************@hotmail.com> wrote:
After Brian mentioned the use for <link rel=..> for navigational
purposes in another thread, I've been looking into it and found that
HTML 3.2 has two other recognized link types than HTML 4.01, which
are "top" and "search".
Technically, HTML 3.2 mentions them as "proposed relationship values",
after stating that "HTML link relationships are as yet unstandardized,
although some conventions have been established." The word "proposed" is
much closer to truth than "established".

HTML 4.01 is more vague, though misleading in a different way, when it
says: "Authors may use the following recognized link types, listed here
with their conventional interpretations." This (probably intentionally)
leaves it open whether the list is a normative definition of the meanings
of some rel attribute values or a descriptive presentation of common (or
"conventional") browser interpretation. In reality, it's a little bit of
both, but mostly neither. It's not exact enough to be meaningfully
normative, and it's far too far from reality to be descriptive.

In particular, IE ignores all the values, except rel="stylesheet" in
<link> elements.
I suspect that "Start" in HTML 4.01 is the same as "top" in HTML 3.2,
is this correct?
We only have two vague definitions. In practice, Mozilla treats "start"
effectively as "top". But "start" is defined oddly. For example, if you
have a set of pages linked together as a simple linked list, with no
main or contents page (bad practice, but possible) then "start" would
naturally refer to the first in the sequence. But it would not be "top".
Is there for HTML 4.01 something like "search" in HTML 3.2?
Not really. This illustrates how the rel values are just a soup, mixed
differently in different versions. Note that "search" was fairly vaguely
defined in HTML 3.2: "The link references a page for searching material
related to a collection of pages." Why that "related to" instead of "in"?
Because they couldn't decide.
Since rel="prev" defines a link to a previous page, is this the same
as the rev attribute?


No, rel="prev" means the same as rev="next". But the rev attribute is
virtually unused, except for the rev="made", which is a rare idiosyncracy
for some browsers. Generally, the concept of rev attribute just confuses
people.

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

Jul 23 '05 #3

P: n/a
On Tue, 28 Sep 2004, Jukka K. Korpela wrote:
In particular, IE ignores all the values, except rel="stylesheet" in
<link> elements.


Jim Ley wrote a bit of Jscript which would pop them up as a usable
list of links (at least in IE5 and 5.5, relying on the machinery in
MS's own IE5WA web accessories). So I think it's fair to say that IE
is not entirely ignoring them, even though it - by default - isn't
rendering them in any useful way.
I suspect that "Start" in HTML 4.01 is the same as "top" in HTML 3.2,
is this correct?


We only have two vague definitions.


From the W3C themselves, agreed.

But various studies have been published on the web;
http://www.subotnik.net/html/link.html is useful in its own right, and
also cites quite a list of links to relevant material.

Jul 23 '05 #4

P: n/a
Tim
On Tue, 28 Sep 2004 09:29:16 +0200,
Harrie <di******************@hotmail.com> posted:
I suspect that "Start" in HTML 4.01 is the same as "top" in HTML 3.2, is
this correct?


Generally, they're not treated as the same. Start is usually the starting
page for a set of documents (which could be a subset of another set of
documents - family tree style). Some people regard "top" as the server
root (e.g. href="/" for the top of the site), some as the parent of the
current location (e.g. href="./" for the top of this branch of the tree),
i.e. there's no consensus. Whatever you do, do it consistently throughout
a site, in what seems the most intuitive manner possible (i.e. what doesn't
need explaining).

You can avoid using top, if you feel it's confusing. For instance, you can
use "home" for the server homepage (i.e. href="/"), and "up" for traversing
up the tree (or avoid "up" and go for plainly obvious things like
"homepage" and "contents").

--
If you insist on e-mailing me, use the reply-to address (it's real but
temporary). But please reply to the group, like you're supposed to.

This message was sent without a virus, please delete some files yourself.
Jul 23 '05 #5

P: n/a
On Tue, 28 Sep 2004 22:49:47 +0930, Tim <ti*@mail.localhost.invalid>
wrote:
On Tue, 28 Sep 2004 09:29:16 +0200,
Harrie <di******************@hotmail.com> posted:
I suspect that "Start" in HTML 4.01 is the same as "top" in HTML 3.2, is
this correct?
Generally, they're not treated as the same.


Generally they are. Mozilla/Firefox maps both of them to the 'top'
entry in the link bar and Opera maps both of them to the 'home' entry
in its link bar.
Start is usually the starting
page for a set of documents (which could be a subset of another set of
documents - family tree style). Some people regard "top" as the server
root (e.g. href="/" for the top of the site), some as the parent of the
current location (e.g. href="./" for the top of this branch of the tree),
i.e. there's no consensus.
It's always seemed to me that 'start' is a horizontal relationship
(i.e. similar to 'first') whereas 'top' is a vertical relationship.
Whatever you do, do it consistently throughout
a site, in what seems the most intuitive manner possible (i.e. what doesn't
need explaining).
Good advice.
You can avoid using top, if you feel it's confusing. For instance, you can
use "home" for the server homepage (i.e. href="/"),
Except that Mozilla doesn't support 'home' so it ends up stuffed away
in the drop down menus at the end.

Opera maps 'top', 'start' and 'home' to the 'home' button.
Mozilla maps 'top' and 'start' to the 'top' button and doesn't map
'home' to any of the default buttons.

Hence I decided to use 'top' for the links that point to my home page
and to avoid 'start' and 'home'.
and "up" for traversing
up the tree (or avoid "up" and go for plainly obvious things like
"homepage" and "contents").


'homepage' would end up not being used at all by Opera and being
stuffed away in the drop downs in Mozilla. For an important link like
that it's probably better to aim for something that will appear in the
basic buttons.

'contents' is mapped to the contents button in Opera and to the Table
of Contents item in the drop down in Mozilla. A ToC may or may not be
the same page as the home page so this may be appropriate for some
sites.

Useful reference: http://webcoder.info/reference/LinkBars.html

Steve

Jul 23 '05 #6

P: n/a
Steve Pugh wrote:
Tim wrote:

Harrie <di******************@hotmail.com> posted:

I suspect that "Start" in HTML 4.01 is the same as "top" in HTML
3.2, is this correct?


Generally, they're not treated as the same.

Generally they are. Mozilla/Firefox maps both of them to the 'top'
entry in the link bar


I don't think that's true in all versions. Mozilla treats them the same.
But Firefox is another story.

The browser-only release has no native site nav bar. Users can get an
extension that adds a site nav bar to the status bar. IIRC, at least one
version of this extension maps 'top' to top, but 'start' to first. I had
to change my <link> elements to avoid confusion.

--
Brian (remove "invalid" to email me)
http://www.tsmchughs.com/
Jul 23 '05 #7

P: n/a
Jukka K. Korpela wrote:
In particular, IE ignores all the values, except rel="stylesheet" in
<link> elements.


IE does, IIRC, add them to the DOM tree. Also, IE does use rel="shortcut
icon".

--
Toby A Inkster BSc (Hons) ARCS
Contact Me ~ http://tobyinkster.co.uk/contact

Jul 23 '05 #8

P: n/a
On Tue, 28 Sep 2004 19:39:31 GMT, Brian
<us*****@julietremblay.com.invalid> wrote:
Steve Pugh wrote:

Generally they are. Mozilla/Firefox maps both of them to the 'top'
entry in the link bar
I don't think that's true in all versions. Mozilla treats them the same.
But Firefox is another story.

The browser-only release has no native site nav bar. Users can get an
extension that adds a site nav bar to the status bar.


Mea Culpe. I did my research with Mozilla and trusted the site that I
quoted in the previous post when it said that Firefox (plus rleevant
extension) acted the same as Mozilla in this respect.
IIRC, at least one
version of this extension maps 'top' to top, but 'start' to first. I had
to change my <link> elements to avoid confusion.


I wrote my links as an Opera user and originally used 'home' for my
links to the site's home page. I later changed that to 'top' following
experiments with Mozilla and threads in this newsgroup. I feel that
the semantics of 'start' are confused and the browser implementations
support this feeling so I'd recommend avoiding 'start'.

Steve

Jul 23 '05 #9

P: n/a
Jukka K. Korpela said the following on 28/09/2004 13:25:
Harrie <di******************@hotmail.com> wrote:
Since rel="prev" defines a link to a previous page, is this the same
as the rev attribute?
No, rel="prev" means the same as rev="next". But the rev attribute is


At http://www.w3.org/TR/html4/struct/links.html#rev-link there's an
example which looks like prev could be the same as rel="prev", but I'm
probably overseeing something.
virtually unused, except for the rev="made", which is a rare idiosyncracy
for some browsers. Generally, the concept of rev attribute just confuses
people.


Count me in. I've seen an example somewhere (I'm a bit glad I lost the
URL to it) which had both rel and rev in one link element. It just
doens't make sense to me.

--
Regards
Harrie
Jul 23 '05 #10

P: n/a
Harrie said the following on 28/09/2004 09:29:
After Brian mentioned the use for <link rel=..> for navigational
purposes in another thread, I've been looking into it and found that
HTML 3.2 has two other recognized link types than HTML 4.01, which are
"top" and "search". I compared these two pages:

http://www.w3.org/TR/REC-html32-19970114#link
http://www.w3.org/TR/html4/types.html#type-links


Thanks to anyone who responded. I've examined the URLs and have
installed/enabled to toolbars for Mozilla, Firefox, Opera and IE
(couldn't find it for Netscape 7.2 and Lynx has it default) to see how
each of them uses it.

--
Regards
Harrie
Jul 23 '05 #11

P: n/a
Harrie <di******************@hotmail.com> wrote:
At http://www.w3.org/TR/html4/struct/links.html#rev-link there's an
example which looks like prev could be the same as rel="prev", but
I'm probably overseeing something.


It's a confusing formulation: "The rel and rev attributes play
complementary roles -- the rel attribute specifies a forward link and the
rev attribute specifies a reverse link." The second sentence tries to
explain the first, but probably manages to obscure things.

At the risk of confusing even more...

The words "forward" and "reverse" need to be taken _very_ abstractly
here. If we have, say, in document A the element <link rel="copyright"
href="B">, then it says that B is information about copyright to A. And
in that case, B might contain <link rev="copyright" href="A">, which says
"this document contains information about copyright to B". There's no
inherent sequentiality, no "forward" (or "backward") involved.

The word prev does not mean anything per se, only as an attribute value.
And rel="prev" and rev="prev" express the same relationship, but in
opposite directions. The word prev is not permissible as an attribute
specification as such, since no attribute is _declared_ (in the formal
syntax, the DTD) with an enumerated list as values. Sorry, I think this
may have added to the confusion.
Generally, the concept of rev
attribute just confuses people.


Count me in. I've seen an example somewhere (I'm a bit glad I lost
the URL to it) which had both rel and rev in one link element. It
just doens't make sense to me.


It's not common, and it's most probably not useful in any way, but by the
specifications, it would be sensible to write e.g.
<link rel="next" href="section42.html" rev="prev">
which says that "section42.html" is the next one (after the current one)
in some author-designed sequence of documents, and the current one is the
document preceding "section42.html" in such a sequence. This would be
logically redundant, of course, by the definition of "next" and "prev".
But we could also have, in theory that is, rel and rev attributes that
express unrelated relationships, e.g.
<link rel="appendix" href="AppA.html" rev="translation">
which says that "AppA.html" is an appendix to the current document and
that the current document is a translation of "AppA.html". (This could
mean that the current document is e.g. an English translation of a
document and the original document is presented as an appendix. Maybe not
a very good example, and "translation" is a value not listed in the
specifications though mentioned in some proposals.)

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

Jul 23 '05 #12

P: n/a
Harrie wrote:
At http://www.w3.org/TR/html4/struct/links.html#rev-link there's an
example which looks like prev could be the same as rel="prev", but I'm
probably overseeing something.


<link rel="prev" href="x"> on page Y is the same as <link rev="prev"
href="y"> on page X.

--
Toby A Inkster BSc (Hons) ARCS
Contact Me ~ http://tobyinkster.co.uk/contact

Jul 23 '05 #13

P: n/a
"Harrie" <di******************@hotmail.com> a écrit dans le message de
news:41***********************@news.xs4all.nl
After Brian mentioned the use for <link rel=..> for navigational
purposes in another thread, I've been looking into it and found that
HTML 3.2 has two other recognized link types than HTML 4.01, which are
"top" and "search". I compared these two pages:

http://www.w3.org/TR/REC-html32-19970114#link
http://www.w3.org/TR/html4/types.html#type-links


Juste a question that comes in my mind after reading this very interesting
thread : if you want to include the whole "site navigation structure" (home
page, sub categories, ...) in each of its page, is it necessary to add all
the necessary <link> in each page, or is there a mean to link to a single
external file that would countains all these <link> references ? (yes I
know, it's just a few lines of html, and you can use inclusion techniques,
but I was wondering if there is something else ?)

Jul 23 '05 #14

P: n/a
On Wed, 29 Sep 2004 06:19:47 +0000 (UTC), "Jukka K. Korpela"
<jk******@cs.tut.fi> wrote:

[...]
...The word prev is not permissible as an attribute
specification as such, since no attribute is _declared_
(in the formal syntax, the DTD) with an enumerated
list as values...
Hmm...

<!ENTITY % LinkTypes "CDATA"
-- space-separated list of link types -->

<!ELEMENT LINK - O EMPTY -- a media-independent link -->
<!ATTLIST LINK
%attrs; -- %coreattrs, %i18n, %events --
charset %Charset; #IMPLIED -- char encoding of linked resource --
href %URI; #IMPLIED -- URI for linked resource --
hreflang %LanguageCode; #IMPLIED -- language code --
type %ContentType; #IMPLIED -- advisory content type --
rel %LinkTypes; #IMPLIED -- forward link types --
rev %LinkTypes; #IMPLIED -- reverse link types --
media %MediaDesc; #IMPLIED -- for rendering on these media --


....seems to me that from the view of a validating parser, anything goes
as a rel/rev attribute value, even...

rel="prev"

....is all Ok.

--
Rex
Jul 23 '05 #15

P: n/a
Jan Roland Eriksson <jr****@newsguy.com> wrote:
...The word prev is not permissible as an attribute
specification as such, since no attribute is _declared_
(in the formal syntax, the DTD) with an enumerated list as values...
Hmm...

- - ...seems to me that from the view of a validating parser, anything goes
as a rel/rev attribute value, even...

rel="prev"

...is all Ok.


Yes, exactly. And this implies that
prev
or
"prev"
alone is not permissible as an attribute specification. _If_ the rel
attribute were declared e.g. as

<!ATTLIST LINK
...
rel (prev,next,foo,bar,whatever) #IMPLIED
...>

and there were no other attribute (for the element) declared with an
enumerated list of values containing the same values, _then_
<link prev href="zap">
would be correct by classical (SGML-based) HTML rules and equivalent to
<link rel="prev" href="zap">.
This is of course just theoretical. Browsers don't even get <h1 center>
right, although it _is_ permitted (and equivalent to <h1 align="center">)
by HTML 4.01 syntax.

--
Yucca, http://www.cs.tut.fi/~jkorpela/
Pages about Web authoring: http://www.cs.tut.fi/~jkorpela/www.html

Jul 23 '05 #16

P: n/a
Pierre Goiffon said the following on 29/09/2004 10:48:
Juste a question that comes in my mind after reading this very interesting
thread : if you want to include the whole "site navigation structure" (home
page, sub categories, ...) in each of its page, is it necessary to add all
the necessary <link> in each page, or is there a mean to link to a single
external file that would countains all these <link> references ? (yes I
know, it's just a few lines of html, and you can use inclusion techniques,
but I was wondering if there is something else ?)


You could use SSI or some other means of "including" (PHP,
preprocessing) for that, but that will probably only be of use for some
of them (home, copyright) since prev and next will change on other pages.

I've gathered some links about including files, but this is far from
complete and PHP is not included in those since I don't use it (but I
probably should).

http://www.htmlhelp.com/faq/html/des...l#include-file
http://www.cs.tut.fi/~jkorpela/html/cpre.html
http://www.allmyfaqs.com/faq.pl?Incl...ile_in_another
http://html-faq.com/utilities/?htmlpreprocessor

--
Regards
Harrie
Jul 23 '05 #17

P: n/a
Toby Inkster said the following on 29/09/2004 09:20:
Harrie wrote:
At http://www.w3.org/TR/html4/struct/links.html#rev-link there's an
example which looks like prev could be the same as rel="prev", but I'm
probably overseeing something.

I made a typo: "like prev could" should be "like rev=.. could".
<link rel="prev" href="x"> on page Y is the same as <link rev="prev"
href="y"> on page X.


To me this looks the same as:

<link rel="prev" href="x"> on page Y is the same as <link rel="next"
href="y"> on page X.

So it looks like rel="next" or rev="prev" is the same, is this correct?

So for page Y one could write rel="prev" href="x" aswell as rev="foo"
href="x"?

--
Regards
Harrie
Jul 23 '05 #18

P: n/a
Jukka K. Korpela said the following on 29/09/2004 08:19:
Harrie <di******************@hotmail.com> wrote:
At http://www.w3.org/TR/html4/struct/links.html#rev-link there's an
example which looks like prev could be the same as rel="prev", but
I'm probably overseeing something.
Sorry for the typo, I made it more confusing:

"like prev could" should be "like rev=.. could".
It's a confusing formulation: "The rel and rev attributes play
complementary roles -- the rel attribute specifies a forward link and the
rev attribute specifies a reverse link." The second sentence tries to
explain the first, but probably manages to obscure things.
This part I get (I think).
At the risk of confusing even more...

The words "forward" and "reverse" need to be taken _very_ abstractly
here. If we have, say, in document A the element <link rel="copyright"
href="B">, then it says that B is information about copyright to A. And
in that case, B might contain <link rev="copyright" href="A">, which says
"this document contains information about copyright to B". There's no
inherent sequentiality, no "forward" (or "backward") involved.
Ok, this still sounds logical to me.
The word prev does not mean anything per se, only as an attribute value.
Yes, that was my typo. What I ment was that rel="prev" href="x" on page
Y looks to me the same as "rev="foo" href="x" on the same page (and I
use "foo" here because that's what the example on the URL I posted uses
and I don't want to use "prev" or "next" here).
And rel="prev" and rev="prev" express the same relationship, but in
opposite directions. The word prev is not permissible as an attribute
So far I understand, but so does rel="next" and rel="prev", or am I
still not getting it? I understand that next and prev only make sense
for a logical order like pages in a book and that rel and rev could be
used for other things (like your copyright example above).
specification as such, since no attribute is _declared_ (in the formal
syntax, the DTD) with an enumerated list as values. Sorry, I think this
may have added to the confusion.
No, was already confused ;)

Thanks for this anyway, I'll sleep over it a few night, read it twenty
more times and one day it'll become clear to me.
Generally, the concept of rev attribute just confuses people.


Count me in. I've seen an example somewhere (I'm a bit glad I lost
the URL to it) which had both rel and rev in one link element. It
just doens't make sense to me.


It's not common, and it's most probably not useful in any way, but by the
specifications, it would be sensible to write e.g.
<link rel="next" href="section42.html" rev="prev">
which says that "section42.html" is the next one (after the current one)
in some author-designed sequence of documents, and the current one is the
document preceding "section42.html" in such a sequence. This would be
logically redundant, of course, by the definition of "next" and "prev".


Yes, it's redundant, but with your explenation it does make sense.
But we could also have, in theory that is, rel and rev attributes that
express unrelated relationships, e.g.
<link rel="appendix" href="AppA.html" rev="translation">
which says that "AppA.html" is an appendix to the current document and
that the current document is a translation of "AppA.html". (This could
mean that the current document is e.g. an English translation of a
document and the original document is presented as an appendix. Maybe not
a very good example, and "translation" is a value not listed in the
specifications though mentioned in some proposals.)


I think I also understand this, but I have to give it some time so my
mind will clear up.

I think I won't use rev= on my pages (at least not for navigation) and
stick to rel="next" and rel="prev" (and top, first, last, etc.). Is that
a good idea or am I in violation of something then?

--
Regards
Harrie
Jul 23 '05 #19

P: n/a
Harrie wrote:
So it looks like rel="next" or rev="prev" is the same, is this correct?


To a sentient being, yes. But, for example, automated agents might not
know that the following pair are links in opposite directions:

rel="next"
rel="prev"

But the following are *defined* to be opposites:

rel="next"
rev="next"

--
Toby A Inkster BSc (Hons) ARCS
Contact Me ~ http://tobyinkster.co.uk/contact

Jul 23 '05 #20

P: n/a
Toby Inkster said the following on 29/09/2004 21:12:
Harrie wrote:
So it looks like rel="next" or rev="prev" is the same, is this correct?
To a sentient being, yes. But, for example, automated agents might not
know that the following pair are links in opposite directions:

rel="next"
rel="prev"


I want to use this for navigation toolbars. One can probably use
rel="first" and rev="first", but I would prefer rel="first" and
rel="last" (and next/prev as above).
But the following are *defined* to be opposites:

rel="next"
rev="next"


I get your point, but I don't know of any use at the moment why I want
to use rev instead of rel. And Jukka said earlier that rev is "virtually
unused" and the site Brian posted (http://www.julietremblay.com/ which
has very nice photographs) as an example for this kind of things also
only uses rel, so I think I forget about rev for the moment.

--
Regards
Harrie
Jul 23 '05 #21

P: n/a
"Harrie" <di******************@hotmail.com> a écrit dans le message de
news:41***********************@news.xs4all.nl
is there a mean to link to a single external file that
would countains all these <link> references ? (yes I know, it's just
a few lines of html, and you can use inclusion techniques, but I was
wondering if there is something else ?)
You could use SSI or some other means of "including"


Thanks for answer. But as I mentionned, I was asking for a solution apart of
inclusion techniques.

Jul 23 '05 #22

P: n/a
On Thu, 30 Sep 2004 01:12:03 +0200, Harrie
<di******************@hotmail.com> wrote:
I want to use this for navigation toolbars. One can probably use
rel="first" and rev="first", but I would prefer rel="first" and
rel="last" (and next/prev as above).
rev="first" and rel="last" are very different.

Let's say you have a set of four documents.
1.html, 2.html, 3.html 4.html

1.html could contain the following:
<link rel="next" rev="prev" href="2.html">
<link rel="last" href="4.html">
<link rev="first" href="2.html">
<link rev="first" href="3.html">
<link rev="first" href="4.html">

Because 2 is the next document and 4 is the last, but 1 itself is the
first document for all the other three. Also see how the link to
2.html has rev="prev" as well as rel="next".

2.html could have these:
<link rel="prev" rev="next" href="1.html">
<link rel="next" rev="prev" href="3.html">
<link rel="last" href="4.html">
<link rel="first" href="1.html">

Note that here there can be no rev="first" or rev="last" because 2
itself is not the first or last document.

3.html could have these:
<link rel="prev" rev="next" href="2.html">
<link rel="next" rev="prev" href="4.html">
<link rel="last" href="4.html">
<link rel="first" href="1.html">

4.html could contain the following:
<link rel="prev" rev="next" href="3.html">
<link rel="first" href="1.html">
<link rev="last" href="2.html">
<link rev="last" href="3.html">
<link rev="last" href="4.html">

I get your point, but I don't know of any use at the moment why I want
to use rev instead of rel. And Jukka said earlier that rev is "virtually
unused" and the site Brian posted (http://www.julietremblay.com/ which
has very nice photographs) as an example for this kind of things also
only uses rel, so I think I forget about rev for the moment.


Quite right. rev is pointless as far as the current generation of web
browsers is concerned. Unless you have some other reason to use rev
(some meta-data gathering tool you use on your sites for instance)
there's little point wasting time adding rev relationships.

Steve

Jul 23 '05 #23

P: n/a
Steve Pugh said the following on 30/09/2004 09:55:
On Thu, 30 Sep 2004 01:12:03 +0200, Harrie
<di******************@hotmail.com> wrote:
I want to use this for navigation toolbars. One can probably use
rel="first" and rev="first", but I would prefer rel="first" and
rel="last" (and next/prev as above).


rev="first" and rel="last" are very different.

Let's say you have a set of four documents.


[snipped a very good example]

Thanks for this, this realy cleared things up for me, I think I see now
where my mind played tricks on me. I'll save the Message-ID of your post
for future reference.
I get your point, but I don't know of any use at the moment why I want
to use rev instead of rel. And Jukka said earlier that rev is "virtually
unused" and the site Brian posted (http://www.julietremblay.com/ which
has very nice photographs) as an example for this kind of things also
only uses rel, so I think I forget about rev for the moment.


Quite right. rev is pointless as far as the current generation of web
browsers is concerned. Unless you have some other reason to use rev
(some meta-data gathering tool you use on your sites for instance)
there's little point wasting time adding rev relationships.


Thanks for confirming that, I was a bit afraid I was going to ignore an
attribute because I just didn't understand it.

--
Regards
Harrie
Jul 23 '05 #24

P: n/a
Pierre Goiffon said the following on 30/09/2004 09:50:
"Harrie" <di******************@hotmail.com> a écrit dans le message de
news:41***********************@news.xs4all.nl
is there a mean to link to a single external file that
would countains all these <link> references ? (yes I know, it's just
a few lines of html, and you can use inclusion techniques, but I was
wondering if there is something else ?)

You could use SSI or some other means of "including"


Thanks for answer. But as I mentionned, I was asking for a solution apart of
inclusion techniques.


The URL's I posted contain some solutions about including, but maybe I
don't understand your question?

--
Regards
Harrie
Jul 23 '05 #25

P: n/a
"Harrie" <di******************@hotmail.com> a écrit dans le message de
news:41***********************@news.xs4all.nl
You could use SSI or some other means of "including"


Thanks for answer. But as I mentionned, I was asking for a solution
apart of inclusion techniques.


The URL's I posted contain some solutions about including, but maybe I
don't understand your question?


Maybe I can't manage to write it in english :D
Well, I was thinking about something like :

<link src="mylinks">

instead of :

<link rev="first" href="home.html">
<link rel="last" href="chap10.html">
<link rel="next" href="chap09.html">
....
Jul 23 '05 #26

P: n/a
Pierre Goiffon said the following on 04/10/2004 10:58:
"Harrie" <di******************@hotmail.com> a écrit dans le message de
news:41***********************@news.xs4all.nl
You could use SSI or some other means of "including"

Thanks for answer. But as I mentionned, I was asking for a solution
apart of inclusion techniques.
The URL's I posted contain some solutions about including, but maybe I
don't understand your question?


Maybe I can't manage to write it in english :D


Or I can't read/understand English properly.
Well, I was thinking about something like :

<link src="mylinks">

instead of :

<link rev="first" href="home.html">
<link rel="last" href="chap10.html">
<link rel="next" href="chap09.html">


Not with <link>, since it doesn't have a src attribute:

http://www.w3.org/TR/html401/struct/...html#edef-LINK

But you could try a SSI (.shtml or whatever your server is configured to
treat as SSI) file like so:

<html>
<head>
<title>Link SSI test</title>
<!--#include virtual="mylinks" -->
</head>
<body>
<p>Test page with included &lt;link&gt; menu</p>
</body>
</html>

But again, take note that "next" is not a god choice here (first and
last are o.k.) since "next" (and "prev" for that matter) will change
from page to page.

You could also work with HTML preprocessing to accomplish the same, but
I have only tested a few things with it. I haven't figured out how to
use it properly (probably with a Makefile, but that's not my strength).

--
Regards
Harrie
Jul 23 '05 #27

This discussion thread is closed

Replies have been disabled for this discussion.