[tei-council] invalid examples allowed in Guidelines?

Tue Jan 11 10:30:09 EST 2011

(sorry to come to this discussion late)

I think this discussion is missing a rather important point about what 
the function of an example is or should be.
The examples in the body of the text of the Guidelines are for the most 
part intended to illustrate a specific point about how a specific 
element should be used. They are not meant to be complete documents in 
themselves, nor exemplary in any sense other than that of the point 
being illustrated. They should be valid, but not if making them valid 
necessitates introducing confusing irrelevancies.

So, in this case, I would ask whether the example is trying to tell us 
that a <div> can contain just <entry>s, or whether it is trying to tell 
us something about the contents of the <entry>s.  I think  it's the 
former rather than the latter. Adding content to the <entry>s would in 
fact confuse the issue (does it mean that a <div> can contain only empty 
entries, or only entries which contain just <gap>s, or any kind of 
entry?) Hence, I think this example is in the category of "deliberately 
invalid schematick" and I would wrap it in a CDATA marked section 
accordingly.

On 11/01/11 13:00, James Cummings wrote:
> On 11/01/11 12:49, Sebastian Rahtz wrote:
>> Here's an interesting example:
>>
>> <div><head>English-French</head>
>> <entry><!-- ... --></entry>
>> <entry><!-- ... --></entry>
>> <entry><!-- ... --></entry>
>> </div>
>>
>>
>> Suppose you _do_ want to make this minimally valid, how
>> do you do so?<entry>'s content model says
>>
>>       ( hom | sense | model.entryPart.top | model.global) +
>>
>> so which does one pick to make a toy example?
> <note>  seems reasonable. ;-)
>
> (I was tempted to suggest<gap>  also valid at this point...)
>
> -James