You are viewing a plain text version of this content. The canonical link for it is here.
Posted to docs@cocoon.apache.org by David Crossley <cr...@indexgeo.com.au> on 2003/09/05 10:05:16 UTC

maintaining consistent URLs

I notice that these old URLs, e.g.
http://xml.apache.org/cocoon/faq/index.html
now redirect to
http://cocoon.apache.org/2.0/faq/index.html

Does anyone know how this is being handled? Is it
a .htaccess or something?

How could we get it to always point to whatever the
current version is? And it would be nice if URLs
like http://cocoon.apache.org/faq/index.html would
also point to the current version.

Also, how do we propose to maintain consistent URLs
when the current doc reorganisation has happened, e.g.
userdocs/ directory is going away, so won't that break
all those other websites that already link to 2.1 docs.

Wouldn't it be terrible if Cocoon becomes laughing stock
... Cocoon is billed as "glue" but it does not stick.

--David




Re: maintaining consistent URLs

Posted by Vadim Gritsenko <va...@verizon.net>.
David Crossley wrote:

>On Tue, 2003-09-09 at 12:01, Vadim Gritsenko wrote:
>  
>
>>David Crossley wrote:
>>    
>>
>>>I found a .htaccess in the xml-site/targets/cocoon/ CVS,
>>>      
>>>
>>But you have found the answer, good :)
>>
>>
>>    
>>
>>>so i changed it to point to the 2.1 docs instead.
>>>I also manually tweaked that file on www.apache.org
>>>      
>>>
>>There is a *reason* why it's in CVS, and the reason is
>>........... so 
>>you don't have to tweak it by hands! :)
>>    
>>
>
>I did do cvs commit, however i thought that i needed to
>*also* do it by hand because i assumed that the old
>xml-site/targets/cocoon/ was not being automatically
>checked out anymore.
>  
>

No, it never (AFAIK) was checked out automatically. I did it for you - 
log was attached to the email.

Vadim


Re: maintaining consistent URLs

Posted by David Crossley <cr...@indexgeo.com.au>.
On Tue, 2003-09-09 at 12:01, Vadim Gritsenko wrote:
> David Crossley wrote:
> >I found a .htaccess in the xml-site/targets/cocoon/ CVS,
> 
> But you have found the answer, good :)
> 
> 
> >so i changed it to point to the 2.1 docs instead.
> >I also manually tweaked that file on www.apache.org
> 
> There is a *reason* why it's in CVS, and the reason is
> ........... so 
> you don't have to tweak it by hands! :)

I did do cvs commit, however i thought that i needed to
*also* do it by hand because i assumed that the old
xml-site/targets/cocoon/ was not being automatically
checked out anymore.

Thanks for your clarification.
--David



Re: maintaining consistent URLs

Posted by Vadim Gritsenko <va...@verizon.net>.
David Crossley wrote:

>David Crossley wrote:
>  
>
>>I notice that these old URLs, e.g.
>>http://xml.apache.org/cocoon/faq/index.html
>>now redirect to
>>http://cocoon.apache.org/2.0/faq/index.html
>>
>>Does anyone know how this is being handled? Is it
>>a .htaccess or something?
>>    
>>

Sorry, I missed this email...


>>How could we get it to always point to whatever the
>>current version is? And it would be nice if URLs
>>like http://cocoon.apache.org/faq/index.html would
>>also point to the current version.
>>    
>>
><snip/>
>
>I found a .htaccess in the xml-site/targets/cocoon/ CVS,
>  
>

But you have found the answer, good :)


>so i changed it to point to the 2.1 docs instead.
>I also manually tweaked that file on www.apache.org
>  
>

There is a *reason* why it's in CVS, and the reason is ........... so 
you don't have to tweak it by hands! :)

vgritsenko@minotaur /www/xml.apache.org/cocoon $ cvs update
cvs server: Updating .
RCS file: /home/cvspublic/xml-site/targets/cocoon/.htaccess,v
retrieving revision 1.2
retrieving revision 1.3
Merging differences between 1.2 and 1.3 into .htaccess
.htaccess already contains the differences between 1.2 and 1.3


Vadim



Re: maintaining consistent URLs

Posted by David Crossley <cr...@indexgeo.com.au>.
David Crossley wrote:
> I notice that these old URLs, e.g.
> http://xml.apache.org/cocoon/faq/index.html
> now redirect to
> http://cocoon.apache.org/2.0/faq/index.html
> 
> Does anyone know how this is being handled? Is it
> a .htaccess or something?
> 
> How could we get it to always point to whatever the
> current version is? And it would be nice if URLs
> like http://cocoon.apache.org/faq/index.html would
> also point to the current version.
<snip/>

I found a .htaccess in the xml-site/targets/cocoon/ CVS,
so i changed it to point to the 2.1 docs instead.
I also manually tweaked that file on www.apache.org

--David




Re: maintaining consistent URLs

Posted by David Crossley <cr...@indexgeo.com.au>.
Bertrand Delacretaz wrote: 
> David Crossley a écrit :
> > ...I think that it is vital for Cocoon's
> > reputation that we maintain all links. I know that we are still
> > a static website and it is a management nightmare.
> 
> It is certainly, I'm afraid that no one will maintain this if we start 
> doing it in a static way.

That is right, so there seems no point trying that.

> > That is why
> > URL-space needs careful planning....
> 
> Right - and now might be a good time, as Carsten started to reorganize 
> the docs.

One way to minimise the disruption would be to only make
radical changes to docs structure with major versions,
say in 2.2 rather than in 2.1.x

> But I'm sure the docs will keep moving around over time, so URL-space 
> planning has to be clever - and easy to maintain for us lazy (or too 
> busy) people.

We should be able to do it. There were various discussions in the past
on the structure of docs, tables of contents, and solid URLs.
With so much happening, perhaps we have all forgotten the outcome.

> I hate to talk once more about tools for the docs without knowing if 
> time will allow me to make it happen, but how about:
> 
> -giving each doc page a unique identifier at the start of the filename, 
> like
> 
>    00042-html-generator.xml

<snip implementation idea/>

> This would not be hard to implement (assuming the docs are served by a 
> live Cocoon at some point), and require zero maintenance.

Yes, but that might make it look like Cocoon needs to have
unique identifiers to be able to generate its output, like some
other clumsy systems. We have the Sitemap, surely there is a way.

Building on this flow of ideas ...

I wonder if there could be a special generator that built a registry by
gathering metadata content from the collection of xdocs. It could get
title, unique-identifier, current location, etc.

It would also have an existing registry that we provide it,
because we know all the URLs that we have published in the past.
We would need to manually build that registry to start, but then
the registry-generator would append to it and keep it maintained
thereafter.

So when the sitemap has no match, this registry is consulted,
and if still not found then generates a 404 and a page that
leads them on to some search solutions. 

Oh yeah, we need the live website.

> Or do you think it is actually possible to define permanent URLs for 
> our docs?

Well, from here on we should be able to define totally permanent URLs
for the conceptual stuff, .e.g. the sitemap conceptual document.

I suggest holding off the update of the website until 2.2 release,
i.e keep the 2.1 structure intact. (There you go Carsten, that
sounds like a "reason" for a new repository :-)

--David



Re: maintaining consistent URLs

Posted by Bertrand Delacretaz <bd...@codeconsult.ch>.
Le Vendredi, 5 sep 2003, à 17:15 Europe/Zurich, David Crossley a écrit :
> ...I think that it is vital for Cocoon's
> reputation that we maintain all links. I know that we are still
> a static website and it is a management nightmare.

It is certainly, I'm afraid that no one will maintain this if we start 
doing it in a static way.

> That is why
> URL-space needs careful planning....

Right - and now might be a good time, as Carsten started to reorganize 
the docs.

But I'm sure the docs will keep moving around over time, so URL-space 
planning has to be clever - and easy to maintain for us lazy (or too 
busy) people.

I hate to talk once more about tools for the docs without knowing if 
time will allow me to make it happen, but how about:

-giving each doc page a unique identifier at the start of the filename, 
like

   00042-html-generator.xml

-handling "not found" requests by searching for the document ID in the 
filesystem, for example "wrongdir/00042-html-generator.html" would be 
redirected to 
"someNewDir/someSubdir/00042-html-generator-with-a-new-name.html" if 
the document has been moved around.

This would not be hard to implement (assuming the docs are served by a 
live Cocoon at some point), and require zero maintenance.

Or do you think it is actually possible to define permanent URLs for 
our docs?

-Bertrand


Re: maintaining consistent URLs

Posted by David Crossley <cr...@indexgeo.com.au>.
Bertrand Delacretaz wrote:
> David Crossley a écrit :
> > ...Also, how do we propose to maintain consistent URLs
> > when the current doc reorganisation has happened, e.g.
> > userdocs/ directory is going away, so won't that break
> > all those other websites that already link to 2.1 docs.
> >
> > Wouldn't it be terrible if Cocoon becomes laughing stock
> > ... Cocoon is billed as "glue" but it does not stick.
> >
> 
> Do you mean
> a) handling obsolete links in a way that still shows the relevant page,
> or
> b) just helping the user find the page?
> 
> I think a) would be nice but a lot of work.

Yes i do mean a) sorry. I think that it is vital for Cocoon's
reputation that we maintain all links. I know that we are still
a static website and it is a management nightmare. That is why
URL-space needs careful planning.

> b) should be improved, the current error page is certainly not good 
> enough.
> 
> For b), the first step could be to configure a nicer error page at the 
> http server level, and the next step would be to show a search page to 
> help people find what they're after (but this requires a live Cocoon 
> instance if we want to do our own precise search).

The live Cocoon is surely getting closer, but gee, we would have
no excuse for broken links if we were live. Our reputation would
then be seriously on show.

> If the current Google search is good enough, an easy way would be to 
> configure an http error page which explains what happened and contains 
> a Google search box like on other pages of the site.

That is a good suggestion for a workaround.

--David


Re: maintaining consistent URLs

Posted by Bertrand Delacretaz <bd...@codeconsult.ch>.
Le Vendredi, 5 sep 2003, à 10:05 Europe/Zurich, David Crossley a écrit :
> ...Also, how do we propose to maintain consistent URLs
> when the current doc reorganisation has happened, e.g.
> userdocs/ directory is going away, so won't that break
> all those other websites that already link to 2.1 docs.
>
> Wouldn't it be terrible if Cocoon becomes laughing stock
> ... Cocoon is billed as "glue" but it does not stick.
>

Do you mean
a) handling obsolete links in a way that still shows the relevant page,
or
b) just helping the user find the page?

I think a) would be nice but a lot of work.
b) should be improved, the current error page is certainly not good 
enough.

For b), the first step could be to configure a nicer error page at the 
http server level, and the next step would be to show a search page to 
help people find what they're after (but this requires a live Cocoon 
instance if we want to do our own precise search).

If the current Google search is good enough, an easy way would be to 
configure an http error page which explains what happened and contains 
a Google search box like on other pages of the site.

-Bertrand