Re: SD1 - Short End Tags [fmt]

lee@sq.com wrote:
> 
> > I would much rather have 8 versions of:
> > > <author>
> > >     <firstname>foo</firstname>
> > >     <lastname>bar</lastname>
> > > </author>
> >
> > than have each server alpha-rename the tags to different
> > incomprehensible tags,
> 
> Another approach is to have a header,
> <header>
>  <field element="a" name="firstname">First Name</field>
>  <field element="b" name="lastname">Last Name</field>
>  <field element="c" name="author">Warpper for Author record</field>
>  <field element="d" name="id">Unique ID for database key</field>
> </header>
> <body>
> <c><a>Parcival</a><b>Snitterpoop</b><d>1</d></c>
> <c><a>Priscilla</a><b>Winterbottom</b><d>2</d></c>
> 
> and so forth.
> 
> This maps conveniently into typesetting tables, and also into the
> common variant of CSV files in which the first line lists the field names:
>     firstname,lastname,author,id
>     Parcival,Snitterpoop,1
>     Priscilla,Winterbottom,2
> in which you can't actually distinguish the first line from the others
> except by human understanding and inspection (ugh!).
> 
> I have added textual descriptions in the header; these could be omitted if
> unavailable, with the possibility of making the "field" element EMPTY.
> 
> Lee

Umm, what does one do for variants such as the multiple ways 
different cultures specify a date?  It is not a problem 
in a class (make a subclass), but how do you recommend 
doing that in the header without an extra level of field tagging?

It is a problem in some database applications.  In fact, the 
customization problems for these are getting to be enormous.

len

Received on Tuesday, 20 May 1997 19:31:14 UTC