A good question
also. No, it doesn’t always work like that. Keep in
mind that XML and SGML split in optional and required features and this made a
difference to applications that had to bridge that chasm as well as
organizations.
It will definitely check
for well-formedness (Draconian errors). XML requires it. Validity
handling can vary because DTDs are optional in XML; so, when this is done can
vary. It possibly also varies by property settings of the editor
itself but someone else would have to confirm that. Items such as
id/idref checking can be reported when the user asks for them even if the
parser caught them on-load. This can be helpful when editing document
components such as work packages. IDs are not auto-generated
because local standards vary as to what constitutes an ID (for example, using a
letter for the package type (M for maintenance) and then appending the WP
number and TM series). This is the sort of thing regular
expression/microparsing can check but no the XML DTD-based validity checking.
Other things are
ecosystem-dependent. For example, is the FOSI putting spaces after list
items and in which contexts? Or is the tagger entering editor specific
processing instructions? These come under the ‘secret decoder ring”
knowledge category. FOSI post validity check contributions are a pure
bitch.
So, some of this comes
down to the features of generalized markup v.s. specific local applications and
tools. For example, does your editor of choice make use of the FPI for getting
the DTD or does it use the name in the quotes following? Little items
like that will confound. Catalogs can confound. Training for
specific editors may have walked out the door with the last architect and the
tagger left to do the job doesn’t actually know the details of XML well
enough to know what kinds of mistakes they are making, and the technical
writers don’t care.
I read recently where
TimBL said the web isn’t about documents. NOW he tells us.
len
From: Nico van de Water [mailto:nicovdwater@gmx.de]
Sent: Friday, August 10, 2012 2:23
PM
To: Len
Bullard
Subject: Re: [xml-dev] Who Is
Teaching XML To Technical Writers?
Hi Len, dear list
members,
My most pressing question in this case would be then: doesn't the XML editor
itself check for well-formedness and validity? So far (since roughly 1998) I
have always used XML Spy, and what I liked from the beginning is that this
program immediately flags any errors in the XML one opens. I have always
thought that this was a more or less default functionality in XML editors. Or I
am wrong here?
With thanks and best regards,
Nico.
Op 2012-08-10 20:12, Len Bullard
schreef:
537254D815AAA64492F9FA9A38ECB40DE83BEC@mail4.huntsville.ses-i.com"
type=cite>Good question, Mike. I'm not sure it is manual although that is how I
found it. I'm editing in Notepad++; faster and I can see the error
patterns; and parse checking on load and after changes. It's old school
but it has advantages and one is to find stuff like this.
Probably the editor isn't checking for this on-entry. Arbor Text allows
it and then it's caught if they use id checking (DUPLICATE). However, I
suspect the training is lacking in lots and lots of logistics
departments. This is the "tagging" culture and it is actively harmful.
[end of snip]
NOTES:
1. Please note my new address:
Groesbeeksedwarsweg 301, 6521 DK Nijmegen, Netherlands. My postal address
remains the same: Postbus 1367, 6501 BJ Nijmegen, Netherlands.
2.
3.
========================================================================
Nico van de Water Linguistic Services
For language matters...
Linguistic consultancy, technical and medical/scientific translations,
software and web site localization, technical authoring, and terminology
management
(Mr.) Nico van de Water
Postbus 1367
6501 BJ Nijmegen
Netherlands
VAT ID: NL076023308B02
Registered at Nijmegen Chamber of Commerce, dossier no. 511 63 217
[t] +31-6-300 444 52 (mobile)
[m] nicovdwater@gmx.de
[w] under development
Please consider the environment before printing (parts of) this mail!
