Gnucash XML Data File XSLT Transformation steps

Derek Atkins warlord at MIT.EDU
Mon Aug 11 01:50:15 CDT 2003


Matthew Vanecek <mevanecek at yahoo.com> writes:

> What's the purpose for the prodigious number of namespaces we use? 
> Can't we find a way to simplify and reduce the number of namespaces?

Because each object should have its own...

> I realized that the write_namespaces...() function would probably be
> inadequate.  What would be handy is a dynamic list (or hash) that is
> processed at file save time.  If you add an object, you add the object's
> namespace to the hash.  Being a hash, duplicates are prevented, although
> the same namespace could be added many times.  Trade-offs, I guess.  I
> just came up with this while poking through some XSLT stuff I'm doing
> for other stuff.

Well, that's sort of what happens now -- each object registered with
the XML backend has a callback to "output your namespaces to the data
file".  Granted, this means the namespace will be added even if there
are no instances of that object type, but I consider that ok.

> Also, I think we should maintain the XML output format, because it is
> WAY handy for transforming data to other formats (e.g., HTML for
> reports, etc.), and easier to write than, say, Scheme (for
> non-developers to write reports, etc.).

I never said the XML data format was going away, just that it wasn't
going to be the main data format.  I've always maintained that the XML
format should remain for "export" and "import" purposes --
"interchange"..

But I want to stop seeing XML for day-to-day datastore use.

-derek
-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available


More information about the gnucash-user mailing list