Re: [Gnu-arch-users] Re: arch with 'special files'

[Jan Hudec]

On Thu, Apr 07, 2005 at 11:54:53 +0200, address@hidden wrote:
> On Wed, Apr 06, 2005 at 01:38:16PM +0200, Jan Hudec wrote:
> > On Wed, Apr 06, 2005 at 11:30:20 +0200, address@hidden wrote:
> > > [XML inexact patching]
> > > 
> > > In fact, I do think it's possible. But to be really useful [...]
> 
> > For XML the equivalence is well defined. Two XML documents are
> > equivalent if and only if they parse to the same tree.
> 
> That's formally correct. But I was babbling about `useful´ (which is
> a bit unfair, because it's fuzzy and I can take it to mean anything
> after the fact ;-) I was trying to express my observation that XML
> the data description language (as opposed to the text markup language)
> is often used in instances where wildly different-looking representations
> mean the same thing. This is not a problem per se (this happens with
> C as well ;-), but the tools which mangle those XML files have a sick
> tendency to output wildly different forms (as opposed, for example,
> to a programmer working on a C file: if the order of some declarations
> doesn't matter she won't touch that without a good reason, and then
> it is transporting some semantics of its own).
> 
> Ugh. I don't know whether I've brought across my point.

I think I know what you mean. Well, yes, the point seems to be, that
order may or may not matter and that if it does not, the tools sometimes
reorder needlessly (because they convert it there and back to some other
representation that does not keep the ordering). So the trick is to let
the merge tool know, where the order matters. If the XML schema spec was
good, it would tell this, but I don't really believe it is as good.

-------------------------------------------------------------------------------
                                                 Jan 'Bulb' Hudec 
<address@hidden>

signature.asc
Description: Digital signature