Re: [xml] suggestion with libxml (fwd)

[Hilaire Fernandes <hilaire@seul.org>]

Some suggestion about using XML with Dr Geo who may be usefull to other
project

Hilaire Fernandes 
Dr Geo project http://www.drgeo.seul.org

---------- Forwarded message ----------
Date: Mon, 28 Jun 1999 08:01:18 -0400
From: Daniel Veillard <Daniel.Veillard@w3.org>
Reply-To: xml@rufus.w3.org
To: xml@rufus.w3.org
Subject: Re: [xml] suggestion with libxml
Resent-Date: Mon, 28 Jun 1999 08:01:20 -0400
Resent-From: xml@rufus.w3.org


  Hi Hilaire,

> I may use libxml to parse xml file format for an interactive geometry
> application. 
> I've a couple of question.
>  Here are examples - valid ? - to define points :
>   
>   <freepoint label="A" x="1.3" y="10" />

  ok

>   <freepoint label="B" x="13" y="10" />

  ok
  
>   <middle label="I" point="A" point="B"/>

  Wrong.

>  The 1st things I don't like is I need to define different tags for each
>  kind of point (and there is a lot) althought this is always point I
>  define. Do I have the choice ?

  You can either express your semantic:
    - in the tag name
    - using an extra attribute defining the type, i.e.the kind of point
    - use the structure to express the semantic

>  The 2nd things. Is it valid, in the middle definition, to have twice the
>  point attribut ? Or should I write somethings like :
>  
>   <middle label="I" point1="A" point2="B"/>

   Yes you will have to change a given attribute can occur only once
   other (suggested) way:
   <middle label="I">
       <point value="A"/>
       <point value="B"/>
   </middle>
   Basically it uses the structure to express the relationship, that's
IMHO the best method to use, it scale and can be extended to other 
relationships.

>  This appear to be more easy to parse this with libxml but may be less
>  readble/conistant for users
>  
>  The situation is even worst with intersection point:
>  Here the intersection of line & cercle ; segment & half-line :
>  
>  <intersection label="A" line="(MN)" circle="(C1)" k="1" />
>  <intersection label="B" segment="[MN]" halfline="[OP)" />

  Same
    <intersection label="A" k="1">
       <line value="(MN)"/>
       <circle value="(C1)"/>
    </intersection>

>  This really look like a nightmare to parse with libxml as there is 10th
> kind of intersection point.
>  
>  Alternatively I could make a horrible things like this :
>  
>  <intersection_line_circle label="A" line="(MN)" circle="(C1)" k="1" />
>  <intersection_segment_halfline label="B" segment="[MN]" halfline="[OP)" />
>  
>  Horrible for the users but easy to parse.
>  
>  Any ideas, suggestion ?

  use the structure to express the logical concepts. It won't be harder to parse
is expendable, and basically makes sense ...

Daniel

-- 
	    [Yes, I have moved back to France !]
Daniel.Veillard@w3.org | W3C, INRIA Rhone-Alpes  | Today's Bookmarks :
Tel : +33 476 615 257  | 655, avenue de l'Europe | Linux, WWW, rpmfind,
Fax : +33 476 615 207  | 38330 Montbonnot FRANCE | rpm2html, XML,
http://www.w3.org/People/W3Cpeople.html#Veillard | badminton, and Kaffe.
----
Message from the list xml@rufus.w3.org
Archived at : http://rufus.w3.org/veillard/XML/messages
to unsubscribe: echo "unsubscribe xml" | mail  majordomo@rufus.w3.org