[OSM-dev] scaling

Dermot McNally dermotm at gmail.com
Mon Jan 10 00:28:01 GMT 2011


On 10 January 2011 00:07, Frederik Ramm <frederik at remote.org> wrote:

> Certainly. But that would then not require the kind of massive scaling that
> Steve was hinting at, would it?

To take that level of newbies and point them at today's toolset might
indeed require a lot of scale, though I admit I was more concerned
about the social effects. As such, defining a simpler, less dangerous
workflow for such users helps to "buffer" the database - both from
sheer load and from potentially bad data.

We are already overdue such a process, TBH. Whereas we have some
pretty good tools, like the Mapzen POI collector, there isn't a
comprehensive process of the sort that would, for instance, help avoid
the use of a POI collector to upload yet another copy of the same
hotel.

Indeed, when I read your post, Frederik, I suddenly got a shock. You
do have a point, but if we follow your reasoning through to its
conclusion we should already be discouraging (most kinds of) newbies
from joining OSM, simply because we don't have the (social)
infrastructure to sustain them without damage to the map. Many of us
have observed the phenomenon in our own communities - a new mapper
arrives on the scene, craps all over the map and vanishes. You always
used to get an element of this. As time goes by we get more and more.
And that, as you would probably be the first to point out, is fine,
the Community can cope, and that's why we have to be cautious in our
growth rate.

I would counter that this is insufficiently ambitious. I'm not talking
here about gung-ho ambition to have the most impressive graph of
contributors. Simply that if you can't carry the water to the fire as
fast as the tap can dispense it then there are alternatives to closing
the tap a little. You could get a bigger buckets. Or more buckets and
some pals to help.

Or maybe get a hose.

I don't know yet what our hose would look like. But if we have empty
or broken bits on the map... And if we have the option of getting more
people to work on them... And if we find ourselves tempted to turn
them away or to stall them a bit... Then we would surely be better
served by diverting some effort to finding ways of boosting our social
scale instead.

Dermot

-- 
--------------------------------------
Igaühel on siin oma laul
ja ma oma ei leiagi üles



More information about the dev mailing list