[uf-discuss] Comments from IBM/Lotus rep about Microformats

Ryan King ryan at technorati.com
Fri Dec 8 09:56:05 PST 2006

On Dec 8, 2006, at 4:04 AM, Bruce D'Arcus wrote:
> Let's review:
> Microformts are a clever set of conventions to reuse existing HTML
> attributes to encode some sort of structured meaning.
> However, using "title" to encode machine-readable content is pretty
> much a hack. Title, after all, is really for human readable labels.

I agree, it's not the cleanest of possible solutions. It *is* a  
compromise. Find us something better and we'll switch.

> Likewise, using class to indicate both properties and, um, class, is
> also a hack.

I'm not sure how you can justify claiming this. You must have read a  
different HTML specification than I have, because in my reading,  
there are no limits on how the class attribute can be used in HTML.  
According to the spec, it's available  for "general processing by  
user agents". I've explained more fully before [http:// 

> These hacks are no doubt necessary and practical in the context of
> current HTML and I really have no problem with it in that context, but
> it's precseily why there can be no generic microformat parser.

You say that like it's a bad thing. Show me a generic parser that  
works on the scale of The Web. Even HTML parsers are not generic–  
browser vendors commonly have to vary their browser's behavior from  
site to site in order to deal with differences. At least the same  
microformat can be extracted from different sites in the same way  
(once you know how to parse the HTML :D).

Ryan King
ryan at technorati.com

More information about the microformats-discuss mailing list