Improving OpenStreetMap Follow-up

[Tom MacWright]

As gravitystorm pointed out on the wiki, there should probably be mention
of the wiki page elsewhere, and I see that Andy already did so. Somehow I
didn't know about rails-dev, and had only been in osm-dev and about 4
others, so I'm too late to respond directly to that thread.

To be clear about a few points:

First off, I don't mean to troll or flamebait. The problems stated in the
wiki page are singularly stated because they explain the need for the
changes that might be possible, and the TODO list aka 'next actions' to get
those done. Given that several of the projects that power OSM have been
massive efforts by a small number of people, they're rather attached to
people, and I don't mean to criticize those people, only to guess at how
awesome these projects could get if they had more contributors and
contributions, so that the maintainers don't have to spend all their time
just fixing bugs.

To avoid flamebait, I'm not discussing the typical hot-button issues of
imports, licenses, tagging for x, etc. Not that they don't matter, but that
they're timesinks to discuss.

Also: it's on the wiki page because it's more accessible to people who
aren't plugged into mailing lists, it's mean to be a changing document, and
the mailing list encounters I've seen haven't all been rosy. The wiki page
so far has gotten six great contributors and only one unrepentant
wiki-vandal: I consider that a success.

My goals for the Rails Port can be summed up as:


   1. Make it kind of pretty
   2. Make it more usable in terms of UI
   3. Make it easy for others to get started and contribute
   4. Then look at the long list of awesome additions people have been
   cooking up.

To these purposes, I agree with Andy's renaming suggestion and disagree
with HAML. The Rails Port should be as plain-vanilla code as possible,
understandable at first glance, and should employ the absolute minimum of
programming magic to get there. The problems with it are not primarily
performance, or code-elegance. Yes, definitely the views need clean-up
though.

I'm for moving to more gems, though I haven't entirely wrapped my head
around that situation. Where they provide more code clarity they'll be
great.

And re: an editor in JS. I'm not knocking Potlatch: it's an awesome piece
of work. I think that there's a place for a simpler editor that's more
targeted in terms of tasks. I only say that it might be in Javascript
because that's what people are likely to contribute to (see #3), not
because of any language wars.

And I'm for changes that are considered, in a very 99% of people sense,
'making this more like other Rails apps'. Just code additions to make it
more elegant or clever are cool, but not a priority for me at this point,
since it doesn't fit into those four points - people aren't contributing
for reasons other than absolute code elegance. But that's not a
prescriptive statement: if that's what people are interested in working on,
go for it!

Tom
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/rails-dev/attachments/20120226/2320c029/attachment.html>