Need XML advice

Here is my idea: <project name="Southern Death Valley Gravity Survey">   <description>     <date>June 1989</date>   <datacollection>       <date>December 1988</date>     <date>March 1989</date     <date>May 1989</date>     </datacollection>   <surveyors>     <surveyor>lfs</surveyor>     <surveyor>rrm</surveyor>     <surveyor>ccg</surveyor>     <surveyor>lcm</surveyor>     <surveyor>tgp</surveyor>   </surveyors>   <descriptiontext>Data collected in the Southern Death Valley area trying to model the Garlock fault zone near the eastern terminus of the zone.</descriptiontext>     <referencebenchmark>Shoshone, California (36.2 N 116.4 W)</referencebenchmark>   <terraincorrection>true</terraincorrection>   </description>   <data modelname="SDV NS 1998fs">     <orientation>015</orientation>   <direction>NS</direction>   <startingpoint>     <lat>36.68N</lat>     <long>116.36W</long>   </startingpoint>   <datapoints>     <point seqno="1" id="001">       <distance>-12075</distance>     <gravity>9.8</gravity>     <magnetic>458</magnetic>     </point>     <point seqno="2" id="002">       <distance>-11684</distance>     <gravity>9.5</gravity>     <magnetic>423</magnetic>     </point>     <point seqno="3" id="003">       <distance>-10900</distance>     <gravity>8.6</gravity>     <magnetic>389</magnetic>     </point>     <point seqno="4" id="004">       <distance>-9986</distance>     <gravity>6.3</gravity>     <magnetic>453</magnetic>     </point> ...     <point seqno="n" id="nnn">       <distance>15698</distance>     <gravity>159.2</gravity>     <magnetic>1036</magnetic>     </point>   </datapoints>   </data> </project> > I need some advice on this.  How would you format the XML?  I have an > idea, but want a second opinion.  I'll post my idea tomorrow.  Also, > should I write an DTD or a schema?  One or the other is required.   > Which is more difficult and which is more extensible? > > Background: > > For an outside project I have been tasked with converting a number of > human readable geophysics survey reports to XML to be imported into a > modelling package.  The modelling package requires either a DTD or a > Schema and another text file that translates the XML fields into the > fields it requires.  Any fields tagged as a description or that it > does not recognize is added to a description section for the model. > > Here is a snippet of my human readable report: > > === > > Project Name:  Southern Death Valley Gravity Survey > Project Date:  June 1989 > Data Collection Date:  December 1988 > Data Collection Date:  March 1989 > Data Collection Date:  May 1989 > Surveyor:  lfs > Surveyor:  rrm > Surveyor:  ccg > Surveyor:  lcm > Surveyor:  tgp > Description:  Data collected in the Southern Death Valley area trying > to model the Garlock fault zone near the eastern terminus of the zone. > > Reference Benchmark:  Shoshone, California (36.2 N 116.4 W) > Terrain Correction:  true > Model Name:  SDV NS 1998fs > Orientation:  NS015 > Starting Point Lat:  36.68 N > Starting Point Long:  116.36 W > Data Point:  1, -12075, 9.8, 458 > Data Point:  2, -11684, 9.5, 423 > Data Point:  3, -10900, 8.6, 389 > Data Point:  4, -9986, 6.3, 453 ... > > Data Point: n, 15698, 159.2, 1036 > > === > > The modeling package requires the orientation, direction, starting > point and the data points (ID, location on a line in meters, gravity > reading and magnetic reading).  The data points may or may not be in > order of either ID or position on the line.  Additionally, the IDs are > aplha numeric as opposed to straight numeric.  Also, the modelling > application can accept a straight numeric additional ID for each point.