Re: Dojo tabs demo / using dojo in LSMB

[John Locke <..hidden..>]

On 07/29/2013 07:36 AM, Brian Wolf wrote:
> Perhaps the "common denominator" of an area of the application.  For 
> example, in AR lots of functionality surrounds the customer; in AP, 
> the vendor. There's probably much overlap in selecting an entity,  
> viewing (perhaps a dashboard) basic information about it, and 
> performing operations on that entity (eg, receiving a payment).  It 
> might be effective to have a one-page for a specific area of the 
> application.
> On 07/29/2013 09:49 AM, Chris Travers wrote:
>
> > One thing I would ask is that if we go with a multi-page design, what 
> > are the aspects of one-page design that we should be looking into 
> > incorporating?


I think the natural place to start is with overview lists -> detail 
views. In many cases, being able to see multiple transactions at once is 
very helpful. Having some panes for viewing/editing data, possibly some 
modal dialogs for data entry, and similar can be really helpful.

I've got a pretty sizeable single-page app we use for much of our 
business, but it keeps the page paradigm -- pages get loaded into tabs 
which may be opened or closed. We've pretty much ignored the challenges 
Brian pointed out with state across refreshes -- as an internal app, we 
just take you back to a launch tab on refresh, you have to open up 
whatever you need. I did have a browser history manager partially 
implemented that would re-open tabs you had closed when you hit the back 
button -- but it wasn't high enough value for us to get fully working.

In any case, converting the multiple pages into more usable standalone 
pages is clearly the next step. I think it's eventually worth 
experimenting with single-page apps, but not necessarily immediately, 
and make sure whatever UI we come up with ends up with a consensus 
before making it official -- probably try out a few different paradigms 
and see what we all like. Whatever we do, I think the app should be able 
to fall back to multiple pages, e.g. with js off, or with a lighter UI 
version for mobile, or whatever.

Brian mentioned earlier a drop-down menu instead of a tree -- I was 
planning to just replace the current tree with a Dojo tree, which does 
use a cookie to remember previous state of expand/collapse. This seems 
like a really good place to start our experiments -- once we have the 
menu in a store, it's pretty trivial to switch between a menu bar and a 
tree widget.

So is there any further objection to adding Dojo to the current trunk? 
The version I've put in my git repo does currently weigh in at 62M -- 
though currently it's adding around 140K to the page loads I've set up 
so far...

Happy to see a few other developers on the list with Dojo experience!

Cheers,
John Locke
http://www.freelock.com