Thank you, Lou. It is what I thought. I asked because the "suggested  
values" in the description of the element don't include it.  
Suggestions of course are just that, and the suggested value "section"  
with its definition of "sections of any kind" can refer to sentences.

But if the milestone element is indeed the recommended element for  
marking sentence boundaries in some environments, why not add it to  
the suggested values? It is not necessary, but it makes an obvious  
point in the most obvious place, and the obvious cannot be repeated  
often enough.

MM

On Mar 3, 2008, at 9:06 AM, lou burnard wrote:

> Martin Mueller wrote:
>> Is there some fundamental objection to using a milestone element  
>> to  mark sentence boundaries with an appropriate @unit attribute?
> Certainly not. You can use milestone elements for anything you please.
>
>> I know  there is a special <s> element.
> Yes, though note that <s> is NOT necessarily for "sentence"s -- just  
> for an end-to-end segmentation. Which. Might Not. Necessaily. Mean  
> Sentences.
>
>> But if your corpus contains a lot of  verse sentences are likely to  
>> cross <l> elements, and using a  milestone may be the simplest way  
>> of managing the problem of  concurrent hierarchies.
>>
>
> The recommended method for dealing with this is precisely....  
> milestones!
>
>
>> MM
>>