kind of off topic now. i would actually look at slide, even for an existing app (if it uses viewstacks and states) for navigation. it makes life much much easier.
Slide uri's map to paths inside a display tree so the url /app/state/stateb/statec tells you that you have set the selected child for 3 viewstacks (thus making complex nav easier). also these uri's are really xpath statements so /app/*/*/state will only filter on the last one. you can even use xpath functions to make decisions on your navigation. this is a gross simplification, but ping me offline, i have a plugin which generates the slide app structure for you. back to navigation. i also have a simple plugin which generates my nav commands for me (slightly slide specific, but easily modifiable), either with mememto or sans, which could help you in this. I use JET to generate code a lot at the moment (i need an abstract factory, i need a set of commands, i need an enum etc etc) which has kind of removed me from the doldrums of implenting paterns to more thinking about the application as a whole. of course generation is not a answer to all (i refuse to generate actual implementation for long and arguable reasons). On Thu, Jan 22, 2009 at 2:18 PM, nwebb <neilw...@gmail.com> wrote: > Thanks v much Johannes - Command was the only pattern I had considered > thus far because I know it's often used for history/undo functionality - > interested to look at memento in conjunction. Useful info as always :) > > On Thu, Jan 22, 2009 at 12:08 PM, Johannes Nel <johannes....@gmail.com>wrote: > >> Depending on the application type, the command pattern works quite >> nicely with navigation. You can then also implement a memento more easily >> and have the command support an "undo" function allowing you to go back and >> forth. >> >> we use slide (obviously) and the uri based navigation (based on states) >> work really well with this approach as well. >> >> >> >> On Thu, Jan 22, 2009 at 11:02 AM, nwebb <neilw...@gmail.com> wrote: >> >>> Thanks guys for all your responses so far. >>> To clarify, it is a multi-step process (with a few possible branches). >>> >>> Tracy/Jim, for the most part I agree. Ideal if I was given the scope to >>> re-design the application, but I've just been handed the task of re-writing >>> the step-process logic. The client is very happy with what they have, so it >>> is a decision that is out of my hands. >>> >>> I am mainly looking for suggestions for (code) design >>> patterns/micro-architectures out there that handle this sort of thing - not >>> a fully-fledged framework. >>> >>> I did like the look of the book Haykel recommended, even though it was >>> not code-based so I may well purchase it too - thanks. >>> >>> Cheers, >>> Neil >>> >>> >>> >>> >>> >>> >>> >>> On Wed, Jan 21, 2009 at 9:02 PM, jim.abbott45 <jim.abbot...@yahoo.com>wrote: >>> >>>> >>>> FWIW, I have to concur with the advice that Tracy gave you and I can >>>> also recommend--HIGHLY--the content at the link that Haykel gave you. >>>> >>>> In short, use Wizard-style navigation for >>>> infrequent/complex/inherently multi-step tasks. For the rest of >>>> (usually, most of) your tasks, use a more fluid navigational >>>> mechanism, such as Hub-and-Spoke. >>>> >>>> --Jim >>>> >>>> >>>> --- In flexcoders@yahoogroups.com <flexcoders%40yahoogroups.com>, nwebb >>>> <neilw...@...> wrote: >>>> > >>>> > Hi, >>>> > >>>> > We have a modular Flex project. >>>> > Each screen has "back" and "next " buttons. >>>> > >>>> > There are various routes through the application and I'm about to >>>> re-write >>>> > the logic which determines where the buttons take the user when they >>>> are >>>> > pressed (what is already in place is overly complex). >>>> > >>>> > I'm guessing that there are fairly established methods for achieving >>>> this >>>> > and would be interested to see what exists, rather than roll out a >>>> bespoke >>>> > solution. Can anyone point me in the direction of a good resource? >>>> > >>>> > Cheers, >>>> > Neil >>>> > >>>> >>>> >>> >> >> >> -- >> j:pn >> \\no comment >> > > > -- j:pn \\no comment
