Re: Semantic linking (Re: Cinclude issues...)
Hey Jeff,
I've got cocoon-2.1 CVS up and running. I'm getting ready to play with
some of the LinkRewriter stuff. I think the semantic linking will be a
huge help in certain areas of our site. I'm glad I can continue to use
the old relative linking too.
I'm still trying to figure out the cleanest way to implement all these
features into the site map. We're at a place in our site development
where what we do now will affect our predicessors positively or
negatively for a good time to come.
Since I can't bank that those who come after will be visionaries I
have to try and get things in place that make since and work well by
themselves. The less of a learning curve there is the better. I'm sure
you understand. 8o)
I love the concepts of the semantic web and I want to get as close to it
as possible. Currently, though, I'm surrounded with a trillion options
and trying to dig through to what exactly it is that I need to do next
in order to get there. asideIt seems to be a long sentence day.
8o)/aside
Thanks for listening to my rant. 8o) I'll probably post some RT's in the
future about some of what I've been thinking.
I don't know that there was a call to action in any of that, but any
thoughts you (or anyone else) may have would be *greatly* appreciated.
Thanks,
Ben
Jeff Turner [EMAIL PROTECTED] 02/11/03 21:19 PM
On Tue, Feb 11, 2003 at 03:22:53PM -0500, Ben Young wrote:
Ok, Jeff. I've been pondering this site.xml thing all day.
Btw, the LinkRewriterTransformer is currently only in 2.1. Although as
the InputModules it depends on are also in 2.0.4, it could be ported
fairly easily.
asideBTW, thanks for the upgrade suggestion. 8o)/aside I have a
couple questions about semantic linking that maybe you can field.
I don't want to give up the old linking method just yet, but I would
like to use semantic linking for certain scenarios. Is this posible
with the linkmap implementation?
Yes, it's not an either/or choice. If a link happens to start with
'site:' (or somesuch prefix) it will be translated. Pretty much any XML
format can be used, by changing the XPath prefix and suffix.
In order to manage the 7,000+ pages I would definitely need the
XInclude ability. Would that be very hard to add?
Just a matter of adding map:tansform type=xinclude/ to the pipeline.
Though, the XInclude transformer isn't very good. XPointer support is
dodgy ('/site/samples' doesn't work, but '/site/samples/*' does), and
there is no support for the xmlns() scheme, meaning if your XML uses
namespaces like Forrest's site.xml, you have to use
*[local-name()='foo']. You might be better off using XML entities;.
How hard is it to set up one's own link schemas?
Requires defining a few input modules in cocoon.xconf and the sitemap.
See CVS Cocoon's linkrewriter block
(http://localhost:8080/cocoon/samples/linkrewriter/). Actually, wait
for
me to commit a fix for the bookdemo sample..
--Jeff
Thanks for all your help, Jeff. This look a lot more promising and
full
featured then the way I was headed.
...
-
Please check that your question has not already been answered in the
FAQ before posting. http://xml.apache.org/cocoon/faq/index.html
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
Re: Semantic linking (Re: Cinclude issues...)
On Thu, Feb 13, 2003 at 09:09:21AM -0500, Ben Young wrote: Hey Jeff, I've got cocoon-2.1 CVS up and running. I'm getting ready to play with some of the LinkRewriter stuff. I think the semantic linking will be a huge help in certain areas of our site. I'm glad I can continue to use the old relative linking too. Actually, indirect linking is a better name than semantic. Effectively, you're linking to an XPath node in an intermediate doc, rather than a specific location. I'm still trying to figure out the cleanest way to implement all these features into the site map. We're at a place in our site development where what we do now will affect our predicessors positively or negatively for a good time to come. I suppose KISS is the best approach then. For a first iteration, it might be best to forget about site: linking and concentrate on the menus. It can be added later very easily, and be made to accommodate whatever XML format you choose. Here is a diff of the sitemap changes made in Forrest, to use site.xml for both 'site:' linking and menu (book.xml) generation: http://cvs.apache.org/viewcvs.cgi/xml-forrest/src/resources/conf/sitemap.xmap.diff?r1=1.45r2=1.45.2.1diff_format=h Since I can't bank that those who come after will be visionaries I have to try and get things in place that make since and work well by themselves. The less of a learning curve there is the better. I'm sure you understand. 8o) Oo.. one thing to consider when managing big URI spaces: what happens when they change? Do the old URLs just break when a page reshuffle is required? I have a vague plan on how to solve this (mark obsolete pages and generate redirects) in Forrest. I love the concepts of the semantic web and I want to get as close to it as possible. Currently, though, I'm surrounded with a trillion options and trying to dig through to what exactly it is that I need to do next in order to get there. asideIt seems to be a long sentence day. 8o)/aside :) Well sounds like many of the issues are similar to Forrest's. Feel free to mail me offlist about specific stuff. --Jeff Thanks for listening to my rant. 8o) I'll probably post some RT's in the future about some of what I've been thinking. I don't know that there was a call to action in any of that, but any thoughts you (or anyone else) may have would be *greatly* appreciated. Thanks, Ben ... - Please check that your question has not already been answered in the FAQ before posting. http://xml.apache.org/cocoon/faq/index.html To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
Semantic linking (Re: Cinclude issues...)
On Tue, Feb 11, 2003 at 03:22:53PM -0500, Ben Young wrote:
Ok, Jeff. I've been pondering this site.xml thing all day.
Btw, the LinkRewriterTransformer is currently only in 2.1. Although as
the InputModules it depends on are also in 2.0.4, it could be ported
fairly easily.
asideBTW, thanks for the upgrade suggestion. 8o)/aside I have a
couple questions about semantic linking that maybe you can field.
I don't want to give up the old linking method just yet, but I would
like to use semantic linking for certain scenarios. Is this posible
with the linkmap implementation?
Yes, it's not an either/or choice. If a link happens to start with
'site:' (or somesuch prefix) it will be translated. Pretty much any XML
format can be used, by changing the XPath prefix and suffix.
In order to manage the 7,000+ pages I would definitely need the
XInclude ability. Would that be very hard to add?
Just a matter of adding map:tansform type=xinclude/ to the pipeline.
Though, the XInclude transformer isn't very good. XPointer support is
dodgy ('/site/samples' doesn't work, but '/site/samples/*' does), and
there is no support for the xmlns() scheme, meaning if your XML uses
namespaces like Forrest's site.xml, you have to use
*[local-name()='foo']. You might be better off using XML entities;.
How hard is it to set up one's own link schemas?
Requires defining a few input modules in cocoon.xconf and the sitemap.
See CVS Cocoon's linkrewriter block
(http://localhost:8080/cocoon/samples/linkrewriter/). Actually, wait for
me to commit a fix for the bookdemo sample..
--Jeff
Thanks for all your help, Jeff. This look a lot more promising and full
featured then the way I was headed.
...
-
Please check that your question has not already been answered in the
FAQ before posting. http://xml.apache.org/cocoon/faq/index.html
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
