Re: Considerations on RDF/XUL bindings

>> - What about an "order" statement somewhere in the template 
>> declaration ? as ORDER in SQL statement, it should be possible to sort 
>> answers. I know this is possible in trees, or in javascript, but why 
>> not give this possibility more low level.
> 
> In fact it's more powerful outside trees, because IIRC you can actually 
> specify primary and secondary sort resources.

I am surely completely out of understanding, but having a sort 
implemented in the C template code directly should be quicker and easier 
to use than a combination of XUL and javascript for the developper. And 
this doesn't remove the possibility to make your personnal sort in 
javascript.

Templates, unfortunately are really a pain to learn. Docs are quite old, 
and misses lots of details. This is bad, because much developper pass 
their way after trying the "Rapid Development" language that is supposed 
to be the XPFO and templates mechanism. Where is the Rapid Development 
if you have to spend month gathering docs that are hard to find or 
totally incomplete, discuss month with developpers, get tuned with last 
developpement, read and reread source up and down to try to figure out 
how all this work, try, and try again without having decent error log...

All this is a terrible pain that seems to slow down quite completely the 
developpement of applications based on Gecko.

>> - Xul templating mechanism allow some sort of conditionnal building, 
>> and a limited way of looping. We lack recursion and an efficient way 
>> of building template inside template.
> 
> We support conditions, as long as all you want is equality ;-)
> We definitely support recursion. How else do you think the bookmarks 
> toolbar works?

Well this is a discovery for me, I'm looking at the code at this moment, 
and I can see lots of undocumented attributes (well for me...) or 
complicated/strange use of them :
"infer"
"template"
"parent"
"rdf:type"

how come all these be completely forgotten in the references of Chris 
Waterson, or xul planet ? I don't know who writes the content of the RDF 
guide of XUL planet (I expect that one of the Neil on this newsgroup is 
the author of most of the Tutorial ?), but I noticed 30 new examples 
recently, why not comment complicated examples of firefox code as this 
one ?

>> - If there are syntax error, or logical error, these should be shown 
>> in a special debug mode (why not when attribute debug is set to 'true' 
>> in the template itself ?)
> 
> Logical error meaning an error in the programmer's logic? That would be 
> a very special debug mode :-)

Well I think we have understood each other. I'm not really confident in 
my english, so expect inconsistencies...

Sorry for this long answer, I hope this will be of some use...