[uf-rest] AHAH! Not yet.

[Kevin Marks]

On Jan 4, 2006, at 12:38 PM, Dimitri Glazkov wrote:

> I am all for simpicity, but one thing that bugs me about AHAH is that
> the payload is usually an HTML fragment, not a complete document.

Right, becasue it is meant to replace part of a page.

> In my (progressive enhancement junkie) view, an application must work
> even with JS disabled. This means that we have three options:
>
> 1) Devise a parallel UI for JS-disabled friends (thus violating the
> you-know-what principle)

No it doesn't - the data is not repeated (it's in the AHAH fragments).

This does depend on your backend implementation, but you can 
concatenate chunks of HTML in any decent system. So, you can omit the 
non-essential aspects for non-JS users; you may need to maintain a 
little more state, or do things one at a time instead of having a list 
of elements to act on and their results in one place

> 2) Make all AHAH-served HTML fragments accessible by following a link
> or making a form submission (thus serving invalid HTML)

So, make a page that collates the structure and AHAH fragments. You 
will likely want more than just the fragment in any case - you will 
almost certainly want some higher-level site navigation, and you may 
want a different CSS for these users.


> 3) Somehow figure out how to server complete HTML documents as AHAH 
> payload.

You can do this, but then you need to make the client-side parsing code 
smarter, and you make it harder to re-use the fragments in the way 
described above.