RE: [xml-dev] Six Reasons Not to use XML Attributes

["Len Bullard" <cbullard@hiwaay.net>]

Where agreement holds, not a bad one.  Once out of standards work, I very
rarely design the XML.  I use it.   Agreements like that help with the
memory caveat.

len

-----Original Message-----
From: Cox, Bruce [mailto:Bruce.Cox@USPTO.GOV] 
Sent: Monday, March 05, 2012 4:55 PM
To: Len Bullard; Christopher R. Maden; xml-dev@lists.xml.org
Subject: RE: [xml-dev] Six Reasons Not to use XML Attributes

Some years ago we learned that different XML-able or -aware systems treated
attributes very differently.  In one case, a useful and powerful search
engine, it completely ignored them.  Consequently, we've agreed
internationally that content should never appear in an attribute, only in an
element.  We use attributes for meta data, defined as data that the business
doesn't care about, possibly doesn't even know about.  One experienced
developer put it this way: if the system won't work without the information,
do not put it in an attribute.  We are rarely that extreme in practice, but
it's not a bad starting point.

Bruce B Cox
OCIO/AED/Software Architecture and Engineering Division
571-272-9004


-----Original Message-----
From: Len Bullard [mailto:Len.Bullard@ses-i.com] 
Sent: 2012 March 2, Friday 13:04
To: Christopher R. Maden; xml-dev@lists.xml.org
Subject: RE: [xml-dev] Six Reasons Not to use XML Attributes

It's a good theory but in practice (to paraphrase, "theory is what we know
but can't use; practice is what we do but can't prove"), attributes are used
for displayable content, a simple example being the TM numbers in the
mil-spec doc headers.

pubno="DTM 1-1520-280-13"

Which goes to the point there are so far no rules for using these for which
there are no exceptions.  Therefore as Mike said, no right answers although
as added, some good reasons.

The problems I hit are seldom using these or not but the secret decoder ring
knowledge someone needs to apply them correctly in situations where the DTD
does not actually govern the final output (say XSLT
contributions) or they are badly documented (say almost any DTD of some size
and age).

len


From: Christopher R. Maden [mailto:crism@maden.org] 

There is still a presumption in XML default processing (e.g., the XSLT
default templates) that content is visible and attributes are not.  When
working with machine interchange languages, this doesn't matter, but for
human documents, it is a good principle to follow.

~Chris
--
Chris Maden, text nerd  <URL: http://crism.maden.org/ > "Be wary of great
leaders.  Hope that there are many, many small  leaders." - Pete Seeger