[Cosmo-dev] Getting a handle on specific UI elements
br at osafoundation.org
Tue Dec 12 11:20:44 PST 2006
On Dec 12, 2006, at 11:01 AM, Matthew Eernisse wrote:
>> Nope, it's not random at all, it's the logical path to the element
>> on the screen.
>>> It sounds like y'all are talking about some kind of logical
>>> namespace that would also need to be kept up-to-date as the UI
>>> code evolves -- otherwise you eventually end up with paths to UI
>>> elements that don't really reflect what's going on in the actual
>>> UI anymore.
> Cool, so it seems like you understand what I'm saying here.
> Anything but totally random strings of text kept in a spreadsheet
> or something would have to be maintained to keep in sync with the
> code. We would have to inform QA of those changes so they can keep
> their tests up to date.
We would have to update QA no matter what (under your scheme or mine)
to tell them that there is a new UI element and where to find it. But
in my scheme that string would not change even if there were re-
factorings, and in your scheme it would.
> So given that those requirements are pretty much identical to those
> for the existing JS namespace, I'm kind of dubious about the value
> of coming up with a completely separate namespacing scheme just for
> DOM-node ids.
> It seems like the only tangible benefit is that it's decoupled from
> the JS namespace, and I wonder if that's a big enough benefit to
> warrant the extra overhead.
There's that benefit, and the fact that my scheme requires no coding.
More information about the cosmo-dev