Hello,
>I just realized that we're using simplified XML (SML) throughout the sys=
tem
>(except for the bc:module, bc:href, etc. tags). What do you think of kee=
ping
>it simple also with the bc:*-tags?
>
>-- the attributes way --
><bc:module id=3D"moo" name=3D"HelloWorld" package=3D"hello_world">
> <cache expires=3D"30" use_uri=3D"true" var=3D"$moo"/>
> <params>
> <param_name>value</param_name>
> <another_param_name>another_value</another_param_name>
> </params>
></bc:module>
><bc:href id=3D"href" href=3D"/archive/mp3/any_document.mp3" usedocroot=3D=
"true" />
>
>-- the sml way --
><bc:module>
> <id>moo</id>
> <name>HelloWorld</name>
> <package>some.package</package>
> <cache>
> <expires>30</expires>
> <use_uri>true</use_uri>
> <var>$moo</var>
> </cache>
> <params>
> <param_name>value</param_name>
> <another_param_name>another_value</another_param_name>
> </params>
></bc:module>
>
><bc:href>
> <id>href</id>
> <href>/archive/mp3/any_document.mp3</href>
> <usedocroot>true</usedocroot>
></bc:href>
>
>I'm kinda unsure. I like both and both is descriptive as well as
>standardized ;-) But, in the bcp we use the "simple" format and maybe we
>should keep this convetion throughout.
>
>What are you're thoughts on this?
I think you are right, because XML is only really extensible if you can l=
ater add tags to values to extend the meaning of those values. Since you =
can=B4t have tags in tag attributes, it is better to use SML.
That is the reason why Metabase XML format is also in SML. I realized the=
need to make it SML when I added support to specify external variables w=
ith tag values. That was right before I made first Metabase public releas=
e. I am glad I did it.
BTW, MetaL XML files are fully SML. I have forbidden non-SML constructs i=
n the XML parser, just in case somebody (I) forgets.
Regards,
Manuel Lemos
|