cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ross Gardler <rgard...@apache.org>
Subject Re: [docs] test publish from daisy
Date Thu, 13 Oct 2005 09:56:58 GMT
Stefano Mazzocchi wrote:
> Ross Gardler wrote:
> 
>> Vadim Gritsenko wrote:
>>

...

>>> Another problem I noticed - all URLs are number.daisy.html - we can't 
>>> publish these to official website...
>>
>> Which part(s) are you objecting to?
>>
>> The number is because daisy uses numbers rather than names to identify 
>> every page. There are ways around this in either Forrest or Daisy but 
>> it is quite a bit of work.
>>
>> Related to this is the hierarchical layour of docs. i.e. there is no 
>> directory structure. Again this is a "feature" of Daisy (one that I 
>> have grwon to love). Documents are stored in a "flat" structure. The 
>> illusion of hierarchy is created by the navigation documents. 
>> Currently the Forrest plugin does not reproduce this (although it could).

...

> I still think a flat URL space for documents is a "good thing"(tm). See 
> wikipedia.
> 
> As for non-numeric, well, I personally like numeric, because otherwise 
> it forces you to have a language identifier (again, see wikipedia) and 
> disambiguation pages, but given that we won't have 300K pages, I do 
> agree that makes the user experience a little more comfortable.

I agree with respect to both flat and numeric URLs. Having an URL space
that defines what is in a document assumes that all documents fit neatly 
into a single "classification". This is not often the case.

Of course, you can argue that nobody will remember a number in a URL,
but do they really remember full URLs?

No, they don't they remember key facts about the page and use the Google
search (or thiner bookmarks) to find it again. In other words the URL
(with the exception of the domain part) is irrelevant with respect to
finding information.

However, as Vadim said (in another mail in this thread) we have to be
careful not to break current links and search engine indexing. This can 
be done by forcing the rewriting of links to mirror the existing 
structure, but that assumes the existing structure is good. I don't 
think it is, some of the stuff in user docs, for example, is valuable to 
developers and vice versa.

An alternative would be to create a set of rewrites to maintain the
existing links.

Ross


Mime
View raw message