Ben Bucksch ben.bucksch at
Wed Apr 6 09:57:09 UTC 2011

On 06.04.2011 04:19, Justin Wood (Callek) wrote:
> On 4/5/2011 6:42 PM, Ben Bucksch wrote:
>> Going back to that quoted line, I guess one way to express it would 
>> be "go try to build a XUL-shim over HTML, and see where it fails, 
>> there'll be tons of problems".
> ... writing an XUL shim is bound to fail as well; since it has its own 
> edge-cases and oddities that one wouldn't expect. And to try and mimic 
> those oddities in HTML might be hard/impossible in some cases. Without 
> those differences actually being helpful in many cases

Yes, there are differences that are coincidental. However, many other 
differences are critical and non-obvious, and I mentioned some. I merely 
proposed to "write a XUL shim" as a means to discover where the 
differences are, and then you can sort out those which 1. are 
unimportant and those which 2. make app development less nice (but can 
be worked around), and those which 3. significantly limit app development.

I claim that 3. *and* 2. must be fixed, not just 3.. I agree with you 
that 1. can be neglected.

> The idea would be to write an app in HTML and see where you have problems.

I did, and I had tons of problems, enough for me to give up on it. With 
that approach, you run into problems only coincidentally, though, during 
debugging. By attempting to write a XUL shim, you get much better coverage.

More information about the tb-planning mailing list