Forum OpenACS Q&A: Bazaar for packages and ideas, pre-TIP discussion

1: Bazaar for packages and ideas, pre-TIP discussion

Posted by Malte Sussdorff on 11/14/03 05:45 PM

TIP: to be defined
Title: Bazaar for packages and ideas
State: Draft
Type: Openacs.org Feature Area
Vote: Pending
Created: Thursday, 13 November 2003

Abstract

The idea is to have a bazaar where members of the community can share their ideas about existing packages and future ones. Using ETP or equal editing capabilities, they can describe their vision, invite others to collaborate and say, which features need funding. Interested parties in co-development as well as funding could step forward and state their interest / commitment.

Current situation

Currently the community mostly knows by luck and reading between the lines in the IRC or forums, if new features are beeing developed and may go to the homepage of the developer, to find out more about it. Furthermore, exchange of ideas on future enhancements is not facilitated (IRC and forums are not structured enough for this).

Organisations might have an itch to scratch, but not the ressources to do it on their own. It is really difficult at the moment to look for co-funding.

Developers might have some (spare-)time they would like to spend on something interesting. But currently it is either comming up with an idea of your own, go to bug-tracker and fix some bugs or talk to people whom you happen to know, if you can help out.

Improvement achievement

The bazaar would allow to share ideas about a certain features. Interested parties could rally around a "stall" and have a look at the nice features that are there or might be developed in the future. People coming to the bazaar would be:

Idealists: They have a really cool idea for a new feature on a certain package or a completly new package. Maybe they got it under the shower, maybe a client of theirs has been asking for it. At least there is an urgent need for sharing this idea and discuss it with fellow colleagues.
Developers: They like to develop. They might earn money with it, but primarily they like to develop. If they have spare time, they think, what shall I do next. By browsing through the various stalls in the bazaar he stops at the eye catchers and decides to talk to the idealists about the new features, think about implementation and maybe even start developing it.
Shoppers: Shoppers go to the bazaar to buy carpets. They look for the stalls that offer a carpets that fit nicely into their homes (or with which they could fly to the moon). Once they found a stall offering such carpets, they talk to the idealists there and the developers (carpenters) and try to negotiate on a price. Maybe the price for the carpet might be too high for them, so they talk to others whether they could share the costs and exchange the design (as there is this fantastic carpet copy shop at the exit of the bazaar that copies existing carpets to be used in your home for next to no money). In the end, they hook up with one or more of the developers to build a carpet (feature).

Proposed steps

Create a /bazaar link at openacs.org (could merge with projects)
Install the wiki functionality of Dave there.
Open it up so anyone could open a new stall
Come up with a color / character scheme for
- In development
- Needs funding
- Needs developer
- Needs funding and developer
- Exists in version xyz
Encourage people to visit the bazaar

2: Re: Bazaar for packages and ideas, pre-TIP discussion (response to 1)

Posted by Caroline Meeks on 11/14/03 06:24 PM

This is a great idea.

Any reason we would not want to include non-software items such as:

Creating a shared booth display or shared brochure resources.

Also do we want to support software projects that do not fit within one package? Examples: OSX install on CD. Setting up OACS to be skinable and creating some different designs. Custom installs for various uses.

I'm also not sure about the states.

What state does the following scenario fall into.

Your client has a requirement that you anticipate he will fund in the future but you have not yet done detailed design. Your goal is to gather similar requirements and design ideas and generally give a heads up that you will be working on this in the future.

Also if we include non-software projects, we may want to replace "developer" with "technical resources" or something like that.

3: Re: Bazaar for packages and ideas, pre-TIP discussion (response to 1)

Posted by Jade Rubick on 11/14/03 08:09 PM

How will this interact with TIP #28:

https://openacs.org/forums/message-view?message_id=142810

?

I like the idea, but perhaps these two ideas should be coordinated in some way?

4: Re: Bazaar for packages and ideas, pre-TIP discussion (response to 2)

Posted by Malte Sussdorff on 11/15/03 09:12 AM

Caroline, I utterly agree, it should not be limited to development and was not my intention, though from re-reading the posting, it is. Will definitly change this once it get's tipped.

Your description would fall under *need funding*. Maybe I'm too conservative here, but unless the client has committed himself to it, it would need funding. OTOH, if you are fairly sure about it, just put it under *in development*. But maybe we need to think of a new state, after all.

Jade, the TIP of Dirk arrived due to the discussions about the bazaar and the need for getting information on what is happening in the various packages. I do not want to get a structured approach, which is why the proposal is called Bazaar. It should facilitate easy and spontaneous collaboration and idea exchange. In this it might not be structured enough for a status page (as per TIP #28), especially if the move is going to integrate it into Bugtracker.

In any case, Dirk and I are heavily coordinating on this (the idea has been in our heads for about three months already), I just not wanted to make a huge ammendment on the TIP.

Once openacs.org has been upgraded I'm willing in helping the jumpstart of the bazaar.

5: Re: Bazaar for packages and ideas, pre-TIP discussion (response to 4)

Posted by Lee Denison on 12/11/03 07:13 PM

This is a great idea Malte. I'd be willing to put some time into it if it is approved.

6: Re: Bazaar for packages and ideas, pre-TIP discussion (response to 1)

Posted by Ben Koot on 12/12/03 12:18 AM

Hi folks,

If we would take a commerial approach , I think moonlighter is a great example. What about creating an openacs based counterpart? http://www.guru.com/ For starters we have a massive database of OACS adepts.

Just a thought

Ben
www.timedesk.nl

7: Re: Bazaar for packages and ideas, pre-TIP discussion (response to 1)

Posted by Ben Koot on 12/12/03 12:23 AM

What I would like to add to the idea is barter ( or member value, as some "oldtimers" may recognize from oacs 3,25.) wher we had a "member value"pacakge.

Cheers
Ben

8: Re: Bazaar for packages and ideas, pre-TIP discussion (response to 1)

Posted by xx xx on 01/28/04 11:56 AM

I would like to add existing open source bounty projects to the ideas of this bazar; for example:

The idea of open international collaboration and funding may be included when trying to obtain EU funding or while developing a .LRN consortium:

Multiple project ideas for OpenACS/.LRN have been proposed, so ideas are not our problem:

We may need the groupware functionality of OpenACS to restrict bounties to certain users.
Actually, this might be a good example to show what can be done with OpenACS groupware.

And last but certainly not least, I see talented developers looking for work (and/or willing to put time in the bazar). So they are available.
However, next to the commitment of some very talented coders, I am pretty sure that funding has been a major factor that allowed .LRN to develop as fast as it does. Funding defeats some of the presumed open source myths. As I see it, this bazar should become an incentive for talented open source developers to commit to OpenACS (instead of one of the many other open source projects they can commit to).

Malte are you still planning to TIP this? I realize that your proposal is somewhat different (work is offered to buy as opposed to work is offered to collect) but the two may work together well.

Frank are you willing to share you're thoughts on this? Do your ideas fit in?

(There are other ways to implement this - rentacoder or emoonlighter - but some are patented :) )

9: Re: Bazaar for packages and ideas, pre-TIP discussion (response to 1)

Posted by xx xx on 01/28/04 01:34 PM

"OpenACS sites as central login point" was a near simultaneous posting with my previous posting.

Would that be a feature-driven problem that a bazar could solve?

How should it be handled? Is the dotwrk project a positive experience that we should learn from?

10: Re: Bazaar for packages and ideas, pre-TIP discussion (response to 9)

Posted by Joel Aufrecht on 01/28/04 01:54 PM

"Is the dotwrk project a positive experience that we should learn from?" Since the current status is listed as, "dotWRK has not been released as a package yet, and that is still some way off in the future," I think it would be very useful to learn from the dotWRK people what has and hasn't worked and what the key barriers to success have been. I think the problems in building a successful consortium are related but distinct from the bazaar issues. I propose these criteria for a successful open-source consortium:

Paying members receive more functionality than they would have gotten seperately.
Participants doing work for the consortium are rewarded equitably
The resulting work is useful to the community.

Some issues to think about: (how have other consortia addressed them?)

What timeframes are reasonable?
Where else should we look for possible funding participants?
What kind of formal but simple governance can we use to make sure the process is fair, without adding overhead that swamps the budget or kills the schedule?
Should we start with a specific proposal that solves one party's problems, and then generalize it for the other parties?
How do we break out of chicken-and-egg - a state where everybody is waiting for somebody else to take the lead?
Will we have a fatal free rider problem, where parties don't participate because they think they can wait for the consortium to finish, but because everybody rides, nobody funds the consortium and it doesn't happen?

From the external-authentication consortium in 2003, I offer these suggestions, and I'm sure other participants have more!

Some face-to-face meetings of all participants are essential to build a foundation for communication.
All parties must remain responsive throughout the schedule.
The amount of work that can be completed within the budget is always less than you hope.
Acceptance criteria must be established early and tested continuously. One way to do this is for the work team to provide working code/instructions regularly and users to responsible for testing it within the same day/week.