Re: Very special idea

Thorsten Scherler <[email protected]> Tue, 19 Nov 2002 23:44:05 +0100
Newsgroups gmane.comp.cms.wyona.user
Message-ID <[email protected]>
Hi everybody,

why did I set up a db? It is a nice example for xsp and db-connections 
that's for sure.

But maybe it is better to store the bookmarks in xml directly. Why? 
Consider a directory with subdirectories and files. That are our 
bookmarks. The directories are main categories, the subdirectories are 
sub categories and the files are our uri. A subdirectory can have x 
other directories and then a bookmark.

In XML speaking:
<section title="Main Categorie 1">
	<section title="sub Categorie 1">
	<section title="sub2 Categorie 2">
	<bookmark id="1" .../>
	</section>
	</section>
</section>
<section title="Main Categorie 2">
<bookmark id="1" .../>
</section>

That would be quit hard to store in a RDB.

Speaking of standards:
Have a look at http://gazonk.org/~eloj/projects/xmlbookmarks.html there 
is an example for a bookmark project. But due to 
http://pyxml.sourceforge.net/topics/xbel/docs/html/xbel.html this 
project was freezed. The DTD can be found @ 
http://pyxml.sourceforge.net/topics/xbel/docs/html/public-text.html.

So what do you all think about the db or the xml. or a combination...

King regards

Thorsten Scherler wrote:
> Hi Markus, hello group,
> 
> I did some scratchpad work on that:
> 
> 1. set up a db
> 2. add_link.xsp
> <esql:query>
>  insert into links values
>  (
>  <xsp-request:get-parameter name="id"/>,
>  '<xsp-request:get-parameter name="uri"/>',
>  '<xsp-request:get-parameter name="title"/>',
>  '<xsp-request:get-parameter name="description"/>'
>  )
>  </esql:query>
> 3. add_link.xml
> <root>
>     <!-- field definitions -->
>     <parameter name="id" type="long" min="1" max="99999" nullable="no"/>
>     <parameter name="uri" type="string" min="1" max="100" nullable="no"/>
>     <parameter name="title" type="string" max-len="50" nullable="no"/>
>     <parameter name="description" type="string" max-len="255" 
> nullable="yes"/>
>     <constraint-set name="add_link">
>         <validate name="uri"/>
>         <validate name="title"/>
>         <validate name="description"/>
>     </constraint-set>
> </root>
> 4. add_link html
> <html>
> <body>
> <h1>Add a link</h1>
> <form action="add_link.xsp">
> <table>
> <tr>
> <td>id</td><td><input type="text" name="id"/></td>
> </tr>
> <tr>
> <td>uri</td><td><input type="text" name="uri"/></td>
> </tr>
> <tr>
> <td>Title</td><td><input type="text" name="title"/></td>
> </tr>
> <tr>
> <td>Beschreibung</td><td><input type="text" name="description"/></td>
> </tr>
> </table>
> <input type="submit" value="insert"/>
> </form>
> </body>
> </html>
> 
> ...and put it up to a cocoon pub.
> 
> King regards
> 
> Thorsten Scherler wrote:
> 
>> Hello group, hi Markus,
>>
>>> I guess, if the system is in use, you don't have to import very often 
>>> bookmarks.
>>
>>
>>
>> Ok, I think that is has to be a application that is running in the 
>> background (1) or integrated in another application (2).
>>
>> Let me explain that:
>> Consider a bookmark page (html-Form) where you have to fill out a form 
>> to submit a bookmark. Speaking fill out 2 fields the uri and the title 
>> for the uri. Then submit that. This link-page has to be open all the 
>> time because "u never know when 2 link a page". (1)
>>
>> Good news  go to http://cvs.apache.org/~nicolaken/stuff/index.html
>>    :-)     with Mozilla and click on the _Forrestbar XPI Install link_
>>            It will automatically install the bar :-) (2)
>> This could be an example how to integrate it. Like the Google Toolbar 
>> you will request your information from a server (bookmark would be 
>> stored in a filesystem xml or a database. The Toolbar should be the 
>> aim. Then (and I think only then) your are right.
>>
>> Most additions occure while browsing on a one by one level.
>>
>>> However, to start using the application, you need to set up quickly a 
>>> first set of bookmarks. So an easy way to import the bookmarks of one 
>>> browser will do best in this case (IE and Mozilla are both able to 
>>> export an html file which should be "importable"). 
>>
>>
>>
>> Then you need to write a stylesheet that can transform that:
>> Mozilla:
>> <h3 ADD_DATE="1036725843" LAST_MODIFIED="1036726275" 
>> ID="NC:BookmarksRoot#$3b5de9af">XML</h3>
>> <dd>XML solutions focused on the web
>>         <dl><p>
>>             <dt><a HREF="http://xml.apache.org/#N100CB" 
>> ADD_DATE="1036726009" LAST_CHARSET="ISO-8859-1">Xindice: A native XML 
>> database</a>
>>
>> IE6:
>> <dl><p>
>>     <dt><a HREF="http://www.rpbourret.com/xml/XMLAndDatabases.htm" 
>> ADD_DATE="1027165414" LAST_VISIT="1037568191" 
>> LAST_MODIFIED="1027165415">XML and Databases</a>
>> </dl>
>>
>> I think the system should store the uri and the name of the link.
>>
>>> Actually I consider the need for an export interface with higher 
>>> priority. But, I've to think about it...
>>>
>> As soon the system has the import stylesheet the conversion shouldn't 
>> be a problem.
>>
>>> <snip/>
>>>
>>> --mv
>>>
>>
>>
>> _______________________________________________
>> WyonaCMS-users mailing list
>> [email protected]
>> http://mail.wyona.org/cgi-bin/mailman/listinfo/wyonacms-users
>>
> 
> 
> _______________________________________________
> WyonaCMS-users mailing list
> [email protected]
> http://mail.wyona.org/cgi-bin/mailman/listinfo/wyonacms-users
>