[plug] Whats a good xml editor for linux?

[James Devenish]

In message <200312190954.10817.derekfountain at yahoo.co.uk>
on Fri, Dec 19, 2003 at 09:54:10AM +0800, Derek Fountain wrote:
> I find this very strange. XML is open, modern and very useful technology, and 
> the back end is well supported with expat, libxml2 and the like. For some 
> reason the OSS community doesn't seem too interested in writing an editor 
> capable of dealing with XML files nicely.

In message <200312191303.42184.derekfountain at yahoo.co.uk>
on Fri, Dec 19, 2003 at 01:03:42PM +0800, Derek Fountain wrote:
> OOo is an XML editor.

I'm sorry to hear that you don't think OOo can be classified as a "good"
or "nice" editor of OOo's XML documents. I, too, have had difficulty
being productive with OOo (see also the [OT] appendix, below).

In message <200312191303.42184.derekfountain at yahoo.co.uk>
on Fri, Dec 19, 2003 at 01:03:42PM +0800, Derek Fountain wrote:
> > > Actually, it seems you do understand the reasoning! Think of
> > > OpenOffice as an XML editor - that's exactly what it is.
> >
> > No it isn't an XML editor.
> 
> Open a new document in OOo. Write some text and save it in native OOo format. 
> Go to the disk and look at the file (unzipping it first). XML, right? Now 
> load it back into OOo and edit the document. Save it again and go and have 
> another look. Still XML, but now it's changed. You edited the XML using OOo.

OOo is very good at editing one "category" of XML documents (i.e. OOo
documents). I'm not aware whether it provides assistance for generic
editing of XML files. If not, it sounds dysfunctional (to me) to argue
for it to be classed as "an XML editor" (which is why I had not thought
it would be an answer to the original question). I wouldn't call vim a
CRT editor, yet its invocation is clearly correlated to changes in my
CRT image (create, save, open -- yes, invoking vim on the same file
clearly leads to reproducible effects on the CRT).

* [OT] rant. One problem that comes to mind for many GUI apps, and
office software included, is that people often aim for interfaces that
are comparable to those in use by Microsoft. I personally feel that
Microsoft tends to produce poor interfaces, yet OSS is driven toward
emulation of Microsoft's interfaces by a number of factors. For example:

 - Many GUI developers have only been exposed to Windows GUIs or X11
   GUIs, and therefore emulate those designs by default.
 - X applications tend not to be good examples of interface design, or
   at least it is not possible to look toward UNIX and say "here are a
   lot of applications that point me in a good, coherent direction for
   the design of my interfaces".
 - It is hard to feel comfortable aiming for general usability motifs in
   light of the vast and disparate "interfaces" that we find in the
   wider world (e.g. your fridge adjustment is done differently to your
   microwave adjustment, which is different to adjusting temperatures
   for your sink, which is different to the volume for your TV). It is
   also hard to resist "specialising" too much.
 - Interfaces are *damn annoying* to produce. Factors affecting this
   include: many users don't think like programmers, and won't use
   software in a fashion that is sympathetic to the underlying
   programming language or data structures; basic human interaction
   features often require a lot of code, or a great deal of branched
   execution, time wasting calculations and silly bit shifting; most
   people don't have regression tests for computations, let alone
   interfaces. Text interfaces seem just as much of a challenge as
   GUI applications, to me.

My brain melts whenever the word "innovation" gets near the word
"interfaces". Typically, we would probably find that "the interface
is probably wrong," and yet interface design is an incredibly common
activity. People, hey? Perhaps Derek feels that XMLSpy has an optimal
interface as an XML editor. Perhaps it does!