lists in attributes

[Russ (3)]

Can lists be used in attributes? For example, suppose I have an element like this:

<x tol="-2.0 2.0">

where I am specifying a lower and an upper tolerance on x. Or, for another example,

<poly coeff="-3.23 4.11 8.32"

where I am giving the coefficients of a polynomial. Are these legal?
Are they a good idea? Thanks.

 

[DFN-CIS NetNews Service]

On 12/02/2004, around 09:10, Russ wrote:
R>Are these legal?
Yes.

R> Are they a good idea?
No.

R> Thanks.
You're welcome.

Seriously, I would do something like

<x>
<lowerTol>-2.0</lowerTol>
<upperTol>-2.0</upperTol>
</x>

and

<poly>
<coeff>-3.23</coeff>
<coeff>4.11</coeff>
<coeff>8.32</coeff>
</poly>

 

[Russ (3)]

On 12/02/2004, around 09:10, Russ wrote:
R>Are these legal?
Yes.

R> Are they a good idea?
No.

Why not?

R> Thanks.
You're welcome.

Seriously, I would do something like

<x>
<lowerTol>-2.0</lowerTol>
<upperTol>-2.0</upperTol>
</x>

and

<poly>
<coeff>-3.23</coeff>
<coeff>4.11</coeff>
<coeff>8.32</coeff>
</poly>

To my untrained eye, these appear more cluttered to me, without adding any additional information.

 

[Patrick TJ McPhee]

% > R>Are these legal?
% > Yes.
% >
% > R> Are they a good idea?
% > No.
%
% Why not?

The argument usually goes that you should use XML to mark each individual component of your data, since then your XML parser can give you each individual component, and you then don't have to parse out the values you want from your attribute list.

% To my untrained eye, these appear more cluttered to me, without adding % any additional information.

The argument here goes that it doesn't matter because you would never look at the XML without some XML-specific tool. It says so right on the first page of the standard, although not in so many words.

You may balance this against the fact that DTDs and XML Schemas provide a type (NMTOKENS) which is essentially intended for a white-space delimited list of numbers, and that XML Schemas goes on to provide mechanisms for defining your own lists of white-space delimited things, all of which suggests that somebody who's thought about it a lot thinks that your approach isn't so bad after all. And your approach surely is more readable for some purposes.

 

[Arnold M. Slotnik]

(Patrick TJ McPhee) wrote in

The argument here goes that it doesn't matter because you would never look at the XML without some XML-specific tool. It says so right on the first page of the standard, although not in so many words.

Okay, I'll bite--where are the words that imply that?

Design goal #6 supports the opposite supposition.