> From: [EMAIL PROTECTED] (Kevin D. Clark)
> Date: Tue, 14 Nov 2006 10:30:12 -0500
> Now suppose in version 2 of the protocol we want to add some more
> properties:
> <foo>
> <color>blue</color>
> <height>1.3m</height>
> <fav-dessert>pie</fav-dessert>
> <weight>22kg</weight>
> <fav-movie>Brazil</fav-movie>
> </foo>
That's great, but when v3 of the protocol comes out, and the message
is encoded
<foo>
<color>blue</color>
<height>1.3m</height>
<favorites>
<dessert>pie</dessert>
<movie>Brazil</movie>
</favorites>
<weight>22kg</weight>
</foo>
your v1 and v2 nodes are going to be incompatible with your v3 nodes.
Then you did it wrong in the first place. Now, the benny of XML is you would utilize namespaces to provide versioning and you could encode both formats within one. Occasionally, you have to duplicate information becouse the initial definition wasn't very extendable. But that's a new format.
It's all great and good that I could very well rename 'foo' to 'bar', but.. Why? ;-)
On the other hand, if the communication protocol was formed solely of
grammatical Lojban text, this problem would never happen:
la fus. se skari lo blanu
gi'e mitre li papici loka clani
gi'e rainei lo bavyvacysai poi se cmene zoi gy. pie .gy. ku'o
.e lo skina poi se cmene zoi gy. Brazil .gy. ku'o
gi'e ki'ogra li rere
Easy as that. :) The text will parse into a tree just like XML, but
with the advantage that standard rewriting rules can be used to
convert between different *representations* with the same *meaning*.
Hm. Sounds like one of the original goals of SGML, doesn't it?
....
That looks like french poetry to me.
Thomas
_______________________________________________ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/
