Re: GNU Herds' FS pledges, and a related experiment

[inimino]

Davi Leal wrote:
> Hi again inimino,
> 
> The pledges idea is cool, but so that it works we should all work on the same 
> platform.

Hi Davi,

I expect it may take several experiments to learn what works or what doesn't.  
I haven't tried this before so I am looking to gather information by trying a 
variety of different approaches.

I believe that a convincing and effective answer to these questions will help 
a lot to fill some of the current gaps in free software.  I think we can have 
multiple platforms and try to see what works.  I'm happy to help work on moving 
the existing GNU Herds' platform forward at the same time as trying my own 
experiments.

I'm also very interested in learning from the existing work that's been done at 
GNU Herd, and if there's any resources you could recommend on the history of 
the project I'd like to read about it.

>>   1. Suggest a few small software projects that might be interesting
>>      or generally useful.
> 
> That is already supported at GNU Herds.  The minimun donation required to 
> create a donation-pledge-group is 2 cents.

The key difference as far as I can tell between GNU Herds' approach and the 
approach I've taken in this experiment is that the Herds pledges are open-
ended, and pledges may be made for as long as necessary until someone chooses 
to take the project on.  In this experiment, I've already committed the time 
up front, and the pledges run for a limited time in order to determine what 
I'll be working on.

Most of the differences you highlight arise from this difference in approach.

I suspect both approaches may work best in different circumstances, but I 
don't have any real theory about what those might be just yet.

>>   2. Allow a week for interested people to pledge donations to the
>>      project of their choice. 
> 
> GNU Herds does not force such limitation on period, "one week".

The benefit of this is that people who pledge will know that they can see 
results (either their project wins or it does not) at a known point in the 
future.

Also, since I am running this as an experiment, it is necessary for me to 
have an upper bound set on the time the experiment will run.

>>   4. Work full time for two weeks on the selected project, publishing
>>      daily progress updates, with code published as often as possible.  
> 
> This requirement is too hard.

I think it will work for me (the "daily" may slip now and then) but there 
is no reason it has to be done this way, it's just a choice I've made for 
this experiment.

>>   6. Release a summary of the project, lessons learned, and future plans. 
> 
> This requirement is too hard.

Again I think it will be OK for me, but as a general model this may not 
be appropriate.  This is more a feature of the experiment itself than of 
the funding model.

After this experiment runs and I publish what I've learned, perhaps the GNU 
Herds pledge system can adapt some of the lessons learned so we all benefit.

>> There would seem to be some opportunity for cross-pollination of ideas
>> at least, and maybe we can share some code too.
> 
> The GNU Herds source code is free software, licensed under AGPLv3 to try to 
> avoid the ASP loop hole other licenses suffer.

I'm not familiar with this loop hole and unfortunately have not had time to 
get a firm grasp of the AGPLv3 yet.

> You could try to add your 3 new donation-pledge-groups at GNU Herds to keep 
> track of your pledge proposals: Tree Explorer, lgrep, grapher.

I think it would confuse the current experiment to accept pledges in two ways, 
but I may run future experiments (or larger projects) in that way.



-inimino