[OSM-talk] simplifying mapnik layout definition

Dave Stubbs osm.list at randomjunk.co.uk
Tue May 27 23:50:17 BST 2008


On Tue, May 27, 2008 at 6:29 PM, Robert (Jamie) Munro <rjmunro at arjam.net> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Andy Allan wrote:
> | Now, if someone is volunteering to make a concise definition format
> | that can be pre-processed into the mapnik XML format (or mapnik python
> | code, or even just read by a modified mapnik directly, or whatever)
> | then I'd absolutely love to SEE THE WORKING CODE. That osm.xml is an
> | unwieldy beast isn't in question, nor are the myriad of possibilities
> | to improve it - what is lacking is working alternative.
>
> IMHO in this case, the code is the easy part - it's designing a good
> format that is hard. It has to be actually easier to read and edit in
> general, rather than easier for whoever designed it but just as hard (or
> even harder) for everyone else.
>
> Whether it should be discussed here or in Mapnik's lists depends on if
> the format will be generally useful, or specific to open street map's
> needs - perhaps it will be tied intimately to our system of tags.
> Perhaps the format could be used to generate osmarender formatting at
> the same time.
>
> If we can agree some basic ideas for how the format should work, and
> what won't work, someone may come along and try to implement it. If you
> just say "go away until you have finished the whole thing" every time,
> it's not surprising people don't write anything because they have no
> idea where to start.

I don't think this how it works. You don't have to "go away until you
have finished the whole thing". What you have to do is think about the
problem for a little longer than 15 minutes, come up with a reasonable
starting point for an idea, and prefix the whole e-mail with "I'm
going to do some work on making the mapnik style sheets easier to read
and understand. This is my initial idea, does anybody have any
comments or sugestions before I get started (just to check I'm not
wasting my time)?"

This simple formula takes a pie in the sky "wouldn't it be good if"
massive waste of time into the makings of a useful discussion. The
trick then of course is to ignore all the people suggesting you create
an uber GUI to unify mapnik and osmarender using the one true
formatting technique will all the bells, and just concentrate on the
useful feedback to make sure you're not going to be wasting your time.
You can then go away implement a first step, come back on the list and
ask for further feedback.

If you consider all of that a distraction then you can of course skip
those steps and just do it and convince people they like it (probably
not difficult if you do something useful).

What's not going to happen is us all discussing the perfect style
sheet format with the vague hope that at the end of it all someone
will actually come along and implement it for us. There are hundreds
of routes to take, but if something is going to work then it has to be
done by someone who already has a clue where to start and what they're
doing -- basically someone who was willing to actually properly
consider the problem and how it might be fixed.

Alternatively if you happen to know of a bunch of developers desperate
to get started working on writing some OSM software, but they just
can't seem to find themselves something to do, but would really really
like a couple of specs to get their teeth into, then by all means,
keep talking. But for as long as trac has quite so many untouched
tickets, I'm going to doubt that's really true.

>
> Here's some working code that simplifies the file:
> sed "s/(\[bridge] = 'yes' or \[bridge]='true')/isBridge/g" osm.xml
>
> And the reverse:
> sed "s/isBridge/([bridge] = 'yes' or [bridge]='true')/g"


It doesn't count if anybody with the vaguest clue can do it in 30
seconds for the simple reason that most of us probably already have.

Dave




More information about the talk mailing list