<snip>

robert burrell donkin wrote:

> most times i've used generation before i've had a pretty good 
> understanding of the code. i really can't say that about ecs2. what i 
> need to move forward on generation is to focus on just one ecs package 
> - let's say rtf - and get templates that seem to work ok for that. 
> then they can be rolled out and refined as we go. what would be really 
> useful for me is just one or two typical example classes from rtf 
> hand-coded. i can then use these as a base-line and attempt to 
> generate to them (ie. have a template that produces identical output 
> for those classes). once i've got to that stage i'd need you to go 
> through and take a look at the complete rtf package and make sure that 
> the generation's working ok. we can then try to generate everything 
> and maybe create some junit tests for the major packages so we can 
> make sure that the stuff works. then we'd need to start tweaking by 
> hand - and we'd be committed to the generated code. 

Agreed.

>
>
> another issue is one of the main features of ecs are all those really 
> handy methods which can't really be generated. from what i can see, 
> the work on html seems to have stalled (i'm happy to be corrected, 
> though). so even once we've got a set of templates that you, we'll 
> still need to go through and transfer methods across.

Yeah, If we can cut down on 70% of the generation of the initial html 
element code the biggest barrier will be crossed,  It's alot easier/fun 
to code useful methods/constructors then it is a bunch of accessors. 
 Nope, I'd pretty much agree html is pretty much stalled.  

> if you could find the time just to code one or two typical rtf 
> elements (i shouldn't need more than that), then i'll start to work on 
> creating templates for the generation. 

Ok, I can do that their isn't much to rtf.

-stephan


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to