[uf-discuss] Re: Microformats for scientific papers

[Alf Eaton]

On 19 Feb 2006, at 15:54, Ryan Cannon wrote:

> I like your use of space-separated class names (e.g. citation  
> reference book), but you do have a little bit of redundancy:
>
>   - <abbr> implies an abbreviation, so class="journal-title-abbr"  
> could just be "journal-title"

That's true, but it's going to be easier for parsers to use the class  
name to differentiate rather than having to look at the node name as  
well, so the redundancy might be worthwhile.

>   - <li id="ref16"><a name="ref16"> is redundant and may be (not  
> sure) illegal.

THe (X)HTML validator doesn't seem to complain, so I think it's ok.  
The thing is, the <li> surrounds the citation, so needs an  
identifier, but there also needs to be an anchor link in the page. I  
think you can ignore the internal anchor links as far as semantics  
go: they're only really useful for page navigation, as they don't  
wrap anything useful.

> I also question the extensive use of IDs in general. Using ID  
> fields is extremely limiting. How, for example, could one page  
> contain multiple articles, or sections of an article, or only  
> article metadata (such as an abstract list), if these fields  
> require IDs?

That's a very good point. I was thinking that I could use ids for  
sections that commonly only appeared once per article, like  
abstracts, introduction, bibliography, etc, but you're right, with  
more than one article on a page that won't work. I guess it'll have  
to be classes throughout.

> It seems to me that the main part of getting a scholarly (not just  
> scientific) article microformat would be hCite/citation  
> microformat, and a metadata format (possibly related to Dublin  
> Core??).

There is more than that, particularly when you want to include the  
actual data, but yes, the main parts so far are a) metadata, b)  
citations and c) common article sections (ie the structure of the  
paper itself).

alf.