[Cosmo-dev] XHTML strict and dojo

[Matthew Eernisse]

Jared,

When the base JSP page for the UI was originally created, it actually 
did validate, I promise. :)

The reason now for including a doctype declaration is to trigger strict 
standards-compliant rendering mode in our supported browsers. You're 
absolutely right -- switching to a transitional doctype will make no 
difference here. We wouldn't be conforming to that either.

There might be somethings we could do to move the addition of those 
non-valid attributes into script, so the base page validates. However, 
it's also worth noting that the validator 'sees' almost none of the 
actual markup that the browser is rendering, since it's all generated 
on-the-fly from script.

I'm sure there are now all kinds of non-valid warts in our 
script-generated UI code -- we could sure do a dump of our generated 
source, and try to validate that, but frankly right now I don't see much 
value in that.

I do see value in emitting valid markup -- but at this point, trying to 
validate our script-generated UI markup would be just another thing we'd 
have to test, *separate* from actually testing in all our supported 
browsers to see if the UI functionally works.

I wish there were a better route than misrepresenting the flavor of XML 
we're using -- but I don't know a way besides a using an XHTML DTD to 
tell the browser to use the W3C box model for rendering.


Matthew




Jared Rhine wrote:
> The Cosmo web UI claims XHMTL 1.0 strict compliance.  Validators fail to 
> pass Cosmo pages as XHTML (in part) because of the inclusion of 
> "dojoType" and "widgetId" attributes.
> 
> Any thoughts on this?  Surely XHTML compliance isn't a Preview goal, but 
> is there anything that can ever been done about this?  It's not like 
> switching to a transitional doctype will make Cosmo pass validation, and 
> I doubt there's any reasonable way to get rid of the dojo-specific 
> attributes.  Is our best route to claim XHTML 1.0 strict, but not 
> actually be strict?
> 
> -- Jared
> _______________________________________________
> cosmo-dev mailing list
> cosmo-dev at lists.osafoundation.org
> http://lists.osafoundation.org/mailman/listinfo/cosmo-dev
>