Working on old code is always difficult. IBM got to the point of
removing a bug to their mainframe operating system on average introduced
a new bug.
Then you get into the testing side of things.
The flash side of potlatch is one that given the number of editors using
it and alternatives available to them today may not be a good return on
investment and I think that should be weighed up.
Nomination I think is essential and if it can be expanded so much the
better.
osm2pgsql is not something I have direct experience with but I suspect
it is one of the infrastructure things that many other things depend on.
The learning curve on old code is steep and if you have someone who
knows the code then I think use them if you possibly can. I've seen a
consultant been brought in to make a change and on half way through the
second day one of the programmers walked up to him and asked him what
the change was. The consultant was pointed to the line of code that
needed to be altered and it took a few seconds to make the change. The
consultant was trying to understand what the entire program did before
making any changes in case it had an impact which was the correct thing
for the consultant to do but experience with the software makes things
much faster.
Oh and I've seen someone say we can do that in half the time and half
the cost. Problem was they didn't understand the problems involved or
what needed to be done. They were fired a week later when it didn't
work but that didn't solve the program problem.
Have fun
Cheerio John
Frederik Ramm wrote on 2020-08-01 19:40:
Hi,
nice to see you rescue a few worthwhile things that have fallen through
the cracks of the Microgrant programme.
During the Microgrants process, there were proposals that didn’t make
it, but would together be a good pilot for a “OSM infrastructure”
process,
Are you planning to take the funds for these projects out of the
"Pineapple Grant" money, or out of the regular budget?
The OSMF Board wants to fund a limited number of projects proposed by
trusted long-term volunteers whose work we know and enjoy.
I think that "trusted long-term volunteers" is key here, and somewhat of
a weak point at the same time.
I notice that all three proposals are very short on hard deliverables;
what they mostly promise is working a certain number of hours on a
certain thing but there is no guarantee that, or to what extent, the
thing is going to be achieved. Richard's proposal is the clearest here
("The result will be a version of Potlatch 2 that can be run on Mac and
Windows laptops"), whereas Jochen and Sarah only commit to working on
something, not to actually achieving it. This means we'll pay them no
matter what.
Now this is all fine because we have reason to believe that every one of
the three proposals will be a good investment and even if a goal could
not be achieved, the money would at least land with people who have done
a lot of volunteer stuff for OSM in the past. But the criteria are fuzzy
- why do we trust these three people that if we give them money to work
on something it will be worth it? Assume someone came along saying wait
a minute, I can do the same for half the money! And then we would say,
err, umm, sorry, no, we don't trust you in the same way we trust these
"trusted long-term volunteers".
Looking forward, it might become necessary to define deliverables more
clearly and make payment conditional on results having been achieved,
rather on time having been spent. But if you're lucky...
In the long term, we want to re-activate the Engineering Working Group
(EWG) by making it a place for decision making, project guidance and
budget management for such projects.
... the EWG can take over that job ;)
Bye
Frederik
--
Sent from Postbox <https://www.postbox-inc.com>
_______________________________________________
talk mailing list
talk@openstreetmap.org
https://lists.openstreetmap.org/listinfo/talk