They're a useful case-study for talking about ideas around direct democracy. See http://www.adelaide.edu.au/apsa/docs_papers/Others/Gauja.pdf
The causes of the party's decline is controversial. My take - that they ran into scaling problems when their parlimentary party grew large. Parts of their parliamentary party felt hamstrung by a membership who were free to take positions on issues that weren't well-considered or practical. They needed to move quickly on issues, and the direct democracy process stalled them. The federal parliamentary party went rogue, and a large section of the membership felt betrayed. This killed the collaborative culture that was the heart of what they stood for, and the party collapsed in infighting.
I'm pretty frustrated with the way freedoms are being eroded in the west, but direct democracy is not a strong response to this. It's complicated, ripe for gaming and diffuses responsibility - bad qualities. When it fails, people will look to a "strong leader" who "gets things done" to replace it. Danger.
Keep thinking though, because there will be better models out there. I suspect that if you could draft a new constitution, you could produce a free society with strong real democratic qualities using a structure similar to the old Icelandic Commonwealth - basically lots of statelets.
The major weakness of this kind of system is that it's hard to fight big wars, even to defend yourself. That's a reason that the awesome original Articles of Confederation were replaced with the current US system.
There was no mechanism for the rank and file to state their resistance in a timely fashion in a way that the leadership could hear, and of course the membership itself was deeply divided on the issue - once again, there was no mechanism for dissension amongst the membership to be dissipated. The result was that the party blew up in spectacular fashion at the next elections, unable to get the membership enthusiastic enough to put boots on the ground to run election activities, with a consequent massive drop in the polls, leading to a drop in funding, leading to extinction.
I was in the NSW branch and as a result of this mess became policy manager and tried to put in place what we would call a crowd-sourced policy creation system. But this was in the early 2000s, the Internet was still not entirely ubiquitous, and management in NSW decided that we had to do this thing on paper, not on the Web. It was slow, painful, and I moved countries before getting to the end (and anyway, it was too little too late, the writing was on the wall).
All of that to say that I don't think we can use the Australian Democrats as a counterexample for direct democracy. They never really had the processes in place to give such a system a chance. Today the story might very well be different if a direct democracy party was tried with a correct set up of tools to allow the rank and file a chance to meaningfully participate in policy creation. I've often thought about trying to set one up on that basis, but since I live in a country now where I'm not even a citizen, I've never tried.
You realize you are essentially proposing that we switch from the Republican Party vs. the Democratic Party, to the Vim Party and the Emacs Party- right? We all know how vim vs. emacs debates go down.
There are some interesting parallels there.
[0] http://en.wikipedia.org/wiki/Delegative_democracy [1] http://liquidfeedback.org/
Most citizens will not participate in day-to-day votes, because they will have proxied their votes to other party members they trust. This will bear some resemblance to the power structure of modern day political parties, but it will be transparent, accountable, and easy to change. Some people will still be influential and hold big swaths of votes in their district, and then continue to command this power after the election.
In some cases, the candidate might have such faith that he/she holds the majority of the proxies in the district, meaning they have broad power to make all decisions.
The German Pirate Party already uses this as their decision-making platform.
EDIT: Just posted this and discovered that quadhome and I had the exact same idea at the exact same time. :-) Whoops!
Regardless of your personal coding preferences (my next project might be in golang, tbh), choosing niche technology for what should be a mass open-source effort is not the most promising start. No wonder in Italy they're having trouble implementing it, knowledge must be thin on the ground.
A software like this must be built on EASY technology so that every organisation can adapt it to their own local needs for cheap; it has to be the web equivalent of VB6/VBA, something everyone can pick up fast, something widespread like PHP/Java/Python/Ruby/.Net.
You may really, really enjoy David Graeber's recent book, 'the Democracy Project.'
http://www.holisticpolitics.org/NewParty/NewPartyManual.php
It basically outlines strategies for how a new political party can realistically gain traction. As you allude to, an important factor is focusing on elections that can be won rather than just running to make a statement.
Perhaps it's possible to write software to analyze past election results, polling, etc. to find races where a third party has a chance and would be the best to allocate resources to, kind of like a Moneyball of politics.
Basically, I imagine a political party, defined by the software it runs on. The first incarnation of the software (website) focuses on organizing party members (users) to win elections, but it then gets iterated to include tools for distributed, direct democracy (i.e. the site will let the party members cast votes on what the elected leader votes on, but I also imagine most people proxy their power to others, which is a feature of the site).
The website first needs to get a few local leaders elected... city council, uncontested state positions, maybe a state senator if the v1.0 launch goes well. We need to both recruit people to run for office, then build the software to help them get elected, then build tools to allow transparent governance with direct oversite from the people.
It's hard to imagine capturing national congressional seats or the presidency on the first go, but I think that applying iterative processes to the formation of a political party and its infrastructure is one possible way to overcome the existing machine.