[whatwg] rel/rev for <form> ?

ROBO Design robodesign at gmail.com
Sat Nov 5 03:18:23 PST 2005


On Sat, 05 Nov 2005 12:30:11 +0200, Charles Iliya Krempeaux  
<supercanadian at gmail.com> wrote:

> Hello,
>
<...>
>
> I think I see one thing I didn't explain properly.  Web crawlers would
> NOT "submit" forms.  But they would discover them.  In essence, "rel"
> and "rev" based formats on the <form> element would allow web crawlers
> to discover API's.  And (web crawlers or user agents, etc) could
> present these forms to the user, and allow the user to fill them in
> and "submit" them outside of the web page they are embedding in.
> Essentially, "rel" and "rev" based formats on the <form> element could
> allow you to turn the <form> element into a kind of IDL.
>
> Consider a blog post.  We have rel-author to specify the author blog
> post.  We have rel-tag to specify the keywords phrases the author
> tagged the blog post with.  We have rel-license to license of the blog
> post.  (And all these can be done within the body of the blog post.)
> But how do you specify the API to post a comment to the blog post?
> How do you specify the API to change how you tagged the blog post?
> (Having "rel" and "rev" based formats on the <form> element could
> allow you to do this.)

This would be interesting and this is the reason I don't "refuse" these  
attributes for <form>s too. You'll also agree that this won't be  
perfection and most likely will not apply to many forms any time soon.

This would become a viable solution in the future, with stereotypical  
forms, such as contact (name, email, subject, body), post comment (name,  
email, url, comment), etc.

<...>
> Again, I am NOT suggesting that web crawlers fill them in.  I am
> suggestion that web cralwers use "rel" and "rev" formats on <form>'s
> to discover API's.

Good :).

<...>
>> Also, you need to take into consideration the implications of having
>> rel/rev for <form>.
>
> Could you elaborate on this please.

Well, for one ... presenting meaningful <form>s to users, without the  
actual site might not be that easy (you have to take styling and scripting  
into consideration too).

Also, having this might open doors to security issues.

Even if your intent is not for bots to submit those forms, that won't stop  
bad people from making bots that submit forms. With microformats defined  
for <form>s, spammers will find it much easier to develop targeted bots.  
It's the same they do now with online services like Blogger. Spammers  
don't usually dedicate their time to specific sites (for example, mine).  
They just run stupid bots, hoping they will work (and they didn't yet on  
my site, even if there were some attempts). Having a standard for post  
comments, for contact forms would be good from the developers and end  
users perspectives, but also from a spammer perspective.

To me, that's not a reason good enough for *not* adding rel/rev to <form>.  
I don't really care about spammers, they can be stopped :).

And lastly, having rel/rev for <form> might also make others ask for rel=  
attributes to just about any other tag (just like CLASS and ID). After all  
... any header/paragraph/input/form/link/anchor/label/style/script can  
have its relation defined somehow by some microformats of the future.

-- 
http://www.robodesign.ro
ROBO Design - We bring you the future



More information about the whatwg mailing list