Hi!
I'm thinking about making one directory per page. Inside this directory the metadata can be stored. Together with different versions (or diffs?) of the page.
Regards, Stefan
Using directories might be a good idea. I think that it could also enable us to implement categories/subcategories/sub*categories.
What kind of metadata do you think to be stored?
--vst
Stefan Scholl wrote:
Hi!
I'm thinking about making one directory per page. Inside this directory the metadata can be stored. Together with different versions (or diffs?) of the page.
Regards, Stefan
cl-wiki-devel mailing list cl-wiki-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/cl-wiki-devel
On 9/8/05, Vehbi Sinan Tunalioglu vehbisinan@gmail.com wrote:
Using directories might be a good idea. I think that it could also enable us to implement categories/subcategories/sub*categories.
Please no :) Categories, if they're required, can be implemented either through metadata, or automatically-managed "Catalog index pages". But not through a directory hierarchy; that just breaks the wiki flat-namespace model.
:) Ok.
"Catalog index pages"? It sounds like a manual way. Am I right?
--vst
Ian Clelland wrote:
On 9/8/05, Vehbi Sinan Tunalioglu vehbisinan@gmail.com wrote:
Using directories might be a good idea. I think that it could also enable us to implement categories/subcategories/sub*categories.
Please no :) Categories, if they're required, can be implemented either through metadata, or automatically-managed "Catalog index pages". But not through a directory hierarchy; that just breaks the wiki flat-namespace model.
On 9/8/05, Vehbi Sinan Tunalioglu vehbisinan@gmail.com wrote:
:) Ok.
"Catalog index pages"? It sounds like a manual way. Am I right?
I can imagine a scheme where, when you put a page into a category, an index page somewhere gets automatically updated by the wiki... I don't know if it's a good scheme, but I think you could do that sort of thing without a lot of manual intervention.
Ian Clelland wrote:
On 9/8/05, Vehbi Sinan Tunalioglu vehbisinan@gmail.com wrote:
:) Ok.
"Catalog index pages"? It sounds like a manual way. Am I right?
I can imagine a scheme where, when you put a page into a category, an index page somewhere gets automatically updated by the wiki... I don't know if it's a good scheme, but I think you could do that sort of thing without a lot of manual intervention.
Nice. But how to do that recursively or how to find the category of a given page? Actually, this is a question of whether we want to implement the idea of categories, and if yes, how comprehensive it should be. Stefan, the maintainer, should maybe decide it. But, I think that I'll need a category mechanism that can grow in depth as well, like subcategories.
--vst
On 2005-09-08 23:43:36, Vehbi Sinan Tunalioglu wrote:
Nice. But how to do that recursively or how to find the category of a given page? Actually, this is a question of whether we want to implement the idea of categories, and if yes, how comprehensive it should be. Stefan, the maintainer, should maybe decide it. But, I think that I'll need a category mechanism that can grow in depth as well, like subcategories.
Most wikis use categories, even if their wiki engines don't have propper support for it.
The most primitive wikis just put a wiki word of the category on the page. Then the listing of every page in that category is done with the backreference feature (mostly a link on the page name in the head of a page).
A propper way for categories could be a nice "USP" (hehe) for CL-WIKI.
But I think this feature is not one of the top priorities because with a nice storage system it should be possible to add this very easy.
An index for *all* pages would be nice (highlighting internal links for non existant pages). Categories could be added with a similar functionality.
The namespace should stay flat. A category could be another meta information about a page. Together with the last modified date, author, etc.
Regards, Stefan
On 2005-09-09 14:21:41, Stefan Scholl wrote:
An index for *all* pages would be nice (highlighting internal links for non existant pages).
Some global data structure that could contain all the names and the number of the current version. This number will be used as part of a file name within the page's directory.
All access to this global data structure must be done with locking.
Every write access to this data structure must write the data to a file, too, which can be loaded when you start CL-WIKI.
Scanning all directories for a file name with the highest number isn't very difficult. ==> A desaster recovery is trivial.
-
Ian Clelland -
Stefan Scholl -
Vehbi Sinan Tunalioglu