Re: Poll results available; Temporary quit as potential employee

Arne Babenhauserheide <arne_bab-S0/[email protected]> Fri, 09 Dec 2016 20:25:32 +0100
Newsgroups	gmane.network.freenet.devel
Message-ID	<[email protected]>
Matthew Toseland writes:
> The final results seem surprisingly sensible, though clearly there is
> overlap and lack of clarity around the darknet enhancements stuff.

Yes, it’s pretty solid. Though the only reason that fixing WoT is not at
the top is that xor pulled out his vote for that (and marked it as
non-distributed — he did *not* use it for something else).

However keep in mind that the money will only last for 20 weeks of
fulltime employment, not for 50 (and we did not offer 50 weeks: xor said
that he would need 27k for one year employment, but the cost estimates
used fulltime weeks).

I did an evaluation of the full set of CSVs xor distributed with the
results, and that shows that from the top 20 tasks, 14 are robustly in
the top 20, regardless of the evaluation I choose:

- Web of Trust: Finish first iteration of most critical speed fixes; ranks: 2 0 17; cost estimate: 4.1
- Darknet invitation bundles; ranks: 1 1 1; cost estimate: 1.125
- Improve FProxy CSS3 support to allow better Freesite UI; ranks: 0 2 2; cost estimate: 1.75
- Short node references; ranks: 6 3 12; cost estimate: 0.8
- Legal docs: Instead of the above suggestions, just link the Tor legal FAQ; ranks: 4 4 15; cost estimate: 0.5
- Document how Freenet works, with a infographic; ranks: 10 5 7; cost estimate: 1.625
- Friend requests, like in Facebook; ranks: 3 7 0; cost estimate: 1.375
- Keepalive; ranks: 13 8 3; cost estimate: 1.75
- Friend-of-a-friend connection suggestions (FOAF); ranks: 14 9 4; cost estimate: 2.25
- Darknet chat improvements; ranks: 17 10 5; cost estimate: 2.2
- Single use node references; ranks: 16 11 6; cost estimate: 2.4
- how to install Freenet; ranks: 7 13 8; cost estimate: 0.9
- Migrate to the Gradle build tool; ranks: 12 15 9; cost estimate: 1.166
- Do not reject Freemail messages when an ID is not known; ranks: 8 18 16; cost estimate: 1.0

I plan to do a few other evaluations to test the robustness of the
results and upload them over the weekend. One of them is to use a metric
which reduces the effect of different scaling of the development time
per feature.

> Whoever implements it, or buys it in, should go carefully over
> everything that's been said about it, especially on the bug tracker, and
> check my old FOAF branch. It may not make sense to adhere too rigorously
> to the specific division of jobs given here, especially as they overlap.

Yes. It’s a list of things we’d like to pay for, not a list of clearly
isolated tasks.

> I do think that Xor having to deal with this while potentially being a
> beneficiary was unfair.

That was quite a risky situation, yes, and it had the bad effects we
could have expected :(

I had hoped that me doing an independent evaluation would reduce the
problem, but that seems to not have been enough.

> I for one appreciate all he's done for the
> project, mostly as a volunteer. Unfortunately even for volunteers
> Freenet has very little personpower at the moment...

Yes.

> Legally I probably could work on Freenet a bit as a volunteer. When I
> will have the time and energy to do so remains to be seen. I will update
> my blog sometime soon to reflect this.

That would be great.

> How are we going to move forward on converting this money and
> rough-roadmap into code? Are any of the volunteer devs available? (I'm
> not)

Maybe we can put out that question via all Freenet channels we have to
first see whether someone from our community is up for taking money for
improving Freenet.

> Do we want to try just doing bounties, or put it out to tender? Is
> there enough money for that to be viable? Even if we do that, somebody
> will need to be responsible for supervising them, verifying that the
> code works and is reasonable etc - getting good results from a third
> party developer requires that *we* know what we're doing (see e.g. the
> London Ambulance Service disaster - the contractor was a bunch of
> idiots, but worse, the buyer had no idea what was going on).

There’s a real danger there, yes. This is the weakest spot I see in our
community: We do not have good structures to ensure that money we spend
is used to the best effect. Though if this poll gives social weight to
our roadmaps, then it might help us in that task.

Best wishes,
Arne
-- 
Unpolitisch sein
heißt politisch sein
ohne es zu merken

_______________________________________________
Devl mailing list
[email protected]
https://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
signature.asc (application/pgp-signature, 800 B)