These are chat archives for fieldpapers/fieldpapers

24th
Feb 2016
Kate Chapman
@wonderchook
Feb 24 2016 17:05
just a reminder that the Field Papers steering chat is in a little less than a hour
Ian Ross
@ian-ross
Feb 24 2016 17:30
We've set up the "latest and greatest" version of Lindsey's "big map" changes at http://fieldpapers-demo.cadasta.org -- this is more or less what we'd like to propose for integration into the main Field Papers site in the next week or two. It would be great if people could use the demo site to make some atlases and do the sorts of things they normally do on fieldpapers.org, just to try to shake out any bugs in parts of the code we're not familiar with. We think that the front-end code is now pretty good, I know that the "normal" route to making atlases works, and I know that canned atlases work, but I don't know about any "weirder" things that might exercise other parts of the back-end Rails code.
Seth Fitzsimmons
@mojodna
Feb 24 2016 17:35
:wave:
@ian-ross i’ll think through it. have you tried the GeoJSON input? that’s about as weird as it gets
Ian Ross
@ian-ross
Feb 24 2016 17:36
I tried not to disturb that much. What's the best way to test it?
Seth Fitzsimmons
@mojodna
Feb 24 2016 17:39
good question, sean did the implementation so i’m not sure i ever have
Kate Chapman
@wonderchook
Feb 24 2016 17:40
I've never used it before but you can get there from here: http://fieldpapers-demo.cadasta.org/advanced
Ian Ross
@ian-ross
Feb 24 2016 17:41
OK, yeah, broken. I'd never even looked at that before. I shall FIX IT!
Oh, that looks like an easy one to fix. I'll try to do it now.
Seth Fitzsimmons
@mojodna
Feb 24 2016 17:43
@wonderchook do you feel like you’ve gotten enough feedback yet? we haven’t coordinated on soliciting additional responses (via the front-page of fp.org, for example) yet
Kate Chapman
@wonderchook
Feb 24 2016 17:43
not sure @linzjax would know what feedback we've received
I only sent stuff out to the OSM Talk list and the HOT list
Seth Fitzsimmons
@mojodna
Feb 24 2016 17:48
i hadn’t looked at it in a while
Ian Ross
@ian-ross
Feb 24 2016 17:48
The feedback is a bit sparse, but I was thinking that we might get more and better feedback about the details of the "big map" stuff once the big map is there on fieldpapers.org for people to play with.
Seth Fitzsimmons
@mojodna
Feb 24 2016 17:48
i really like the solution to the accordion / tab problem
Ian Ross
@ian-ross
Feb 24 2016 17:49
The solution of not having accordions or tabs at all?
Seth Fitzsimmons
@mojodna
Feb 24 2016 17:49
exactly!
Ian Ross
@ian-ross
Feb 24 2016 17:49
I've been encouraging Lindsey to just cut stuff out. Simpler is almost always better, it seems.
Lindsey Jacks
@linzjax
Feb 24 2016 17:51
But they were so pretty... sigh
Seth Fitzsimmons
@mojodna
Feb 24 2016 17:51
such is the cost of design ;-)
speaking of, how was your conversation with alec?
Ian Ross
@ian-ross
Feb 24 2016 17:51
Kill your darlings. It's as good advice for programming as it is for writing.
Lindsey Jacks
@linzjax
Feb 24 2016 17:52
It went really well. He also suggested the full-length menu.
Seth Fitzsimmons
@mojodna
Feb 24 2016 17:52
yay! great
he gets a bit busy with other projects, but please continue to bug him when you like
Lindsey Jacks
@linzjax
Feb 24 2016 17:53
He also suggested putting the row/col buttons back on the grid, but I think with the current grid resizing action it would be more frustrating than helpful.
I'm glad he was able to help!
Seth Fitzsimmons
@mojodna
Feb 24 2016 17:56
how did you end up implementing the geocoder?
(i’m using this as an opportunity to spend more time with it and drop some notes in the PR)
ah, using Nominatim, cool
Lindsey Jacks
@linzjax
Feb 24 2016 17:59
yup!
Ian Ross
@ian-ross
Feb 24 2016 17:59
Demo site might go away for a minute right now...
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:00
i ran into the Mapzen Search geocoder widget recently and was pretty impressed. it includes dropdown options, which might be useful for out-of-the-way places with more prominent cousins: https://github.com/mapzen/leaflet-geocoder
Ian Ross
@ian-ross
Feb 24 2016 18:00
OK. Back again. GeoJSON form at least comes up now. Don't know if it works yet.
Lindsey Jacks
@linzjax
Feb 24 2016 18:00
Oh nice, I was wondering if there was a good option that provided a dropdown autocomplete that wasn't google.
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:01
Ian Ross
@ian-ross
Feb 24 2016 18:01
Doesn't. :crying_cat_face:
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:01
it would end up in the stack of buttons near the zoom control
i’m also wondering about moving the grid resize controls to the top left of the viewport
i didn’t notice them at first because most everything else is taking place on the top-left
(and the grey is fairly similar to OSM’s land color. that or my colorblindness is kicking in)
Lindsey Jacks
@linzjax
Feb 24 2016 18:03
Oooh yeah it's pretty much the same color
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:03
if you moved them above the zoom controls it would form an upside-down L shape and increase their prominence
Kate Chapman
@wonderchook
Feb 24 2016 18:04
@ian-ross i started an atlas 30 minutes ago, should it render one on the field papers demo site?
Lindsey Jacks
@linzjax
Feb 24 2016 18:05
Having the search under the zoom would also free up some more space on the menu
which would help with screen size issues
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:08
this stuff is really all polish at this point, so it’d be great to get it in soon and then iterate on the live site, ideally w/ additional feedback
we can put up a blog post (perhaps discussing @linzjax’s Outreachy experience as well) on the FP Tumblr and tweet about it
(@almccon is the social media quartermaster here, so i’ll loop him in)
this is really good work, @linzjax
Ian Ross
@ian-ross
Feb 24 2016 18:10
@wonderchook There's something wrong with the back-end... I broke it a bit earlier when messing around with Rails things.
Lindsey Jacks
@linzjax
Feb 24 2016 18:10
thanks!
Would this be a blog post that I write? I'm currently in the process of writing one that covers both things and would happily pass it along once I'm done.
Kate Chapman
@wonderchook
Feb 24 2016 18:12
@mojodna how was your trip? Any great Field Papers insights?
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:12
@ian-ross uh, GeoJSON may already have been broken: fieldpapers/fieldpapers#210
@linzjax if you’re game! cross-posting is often convenient
@wonderchook we’re leaving in mid-March (though Drishtie et al are in Colombia at the moment and have sent along some useful feedback)
POSM will be going to Ecuador and Liberia in mid-March, so here we’re mostly focused on all of the other pieces. there are a couple of FP bugs that i’m planning to knock off before we leave that are the result of field reports + other observations:
fieldpapers/fieldpapers#216
fieldpapers/fieldpapers#215
fieldpapers/fieldpapers#214
fieldpapers/fieldpapers#213
fieldpapers/fieldpapers#217
(hover over the links to see previews—yay gitter!)
re: fp-tasks, i discovered that we may be rate-limited by OSM at some point but haven’t been able to dive in and talk to someone about why yet
it has the side-effect of backing up the queues, as each page will take 30s before timing out
there’s an undeployed change that sets the User-Agent when fetching tiles to Field Papers, which i’m hoping will mitigate things slightly
Kate Chapman
@wonderchook
Feb 24 2016 18:16
hmmm, I didn't think OSM did rate limiting I thought it was just a straight block for overuse. Are you seeing the tile that says something about the tile use policy?
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:17
no, requests just start taking forever to complete
WB has seen this in Sri Lanka a bunch
FP pulls tiles from the OSU mirror and WB is getting tiles from a different mirror, so that could be part of the problem
Kate Chapman
@wonderchook
Feb 24 2016 18:17
interesting, i suppose if you made a huge atlas that might cause something like that
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:17
it happens on small ones too
usually in areas that are likely to need to be rendered fresh
Kate Chapman
@wonderchook
Feb 24 2016 18:18
yeah maybe that is just taking time to render the tiles
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:18
so it might just be the OSM tile servers being really slow
(openstreetmap-carto is slow to render)
Kate Chapman
@wonderchook
Feb 24 2016 18:18
yeah it has always just supposed to be an example rendering really
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:19
i’ve poked around the utilitization graphs and haven’t seen anything out of the ordinary, so maybe we’re just getting backed up in some renderd queue
Ian Ross
@ian-ross
Feb 24 2016 18:19
@mojodna I think I should be able to fix it. I also need to sort out something in my AWS deployment that's broken the atlas rendering there. (That's why Kate's atlas rendering didn't work. I'll try to do that right now so the demo site is at least usable.)
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:20
thanks!
@wonderchook gravitystorm/openstreetmap-carto#1975 had me thinking about the “example rendering” thing
Kate Chapman
@wonderchook
Feb 24 2016 18:20
@mojodna might not hurt to just send an email to the OSM dev list and ask
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:20
yeah, i’ve been putting off jumping on irc
it’s not really in FP’s remit to go anywhere near cartographic styling, but i’ve been hearing more and more about the desire for purpose-built rendering of specific OSM features
which makes me think that rather than being a kitchen sink (or in addition to), the “OSM style” should actually be a set of styles that highlight different sets of features present in the map
(i digress)
Kate Chapman
@wonderchook
Feb 24 2016 18:22
yeah, I think the style has always just been hard to get agreement hence the kitchen sink approach
I could see how more style options could be helpful for field papers
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:22
we may be able to drive some of that thinking via POSM + rendering from extracts for field purposes
i remain overdue on putting together the Roadmap
Kate Chapman
@wonderchook
Feb 24 2016 18:24
I wonder if another approach is a generic basemap and then some specific renderings. Assuming you would always need roads for example
regarding this issue: fieldpapers/fieldpapers#216 do you have thoughts about making that easier on slow connections
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:25
i know of some pieces that should go in it, but i think i want to share brainspace w/ @wonderchook and sara for a first/second pass. field experience will inform that as well.
Kate Chapman
@wonderchook
Feb 24 2016 18:25
meaning if there were the 40 scans to upload how to deal with that? I was almost picturing someway to use dropbox or something to get things uploaded
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:26
“map ingredients” that can be composited
oh, good idea
let things churn in the background
Kate Chapman
@wonderchook
Feb 24 2016 18:26
yeah, maybe you would just then paste a link to your dropbox folder
Ian Ross
@ian-ross
Feb 24 2016 18:26
Possibly dumb question: is this styling going to be based on CartoCSS or something, so you can have composable styles? (e.g. greyish basemap, highlights for public transport)
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:27
i’ll see how CartoDB handles Dropbox integration and ping @dalekunce / drishtie about how feasible that is
@ian-ross yes
mapstack.stamen.com kinda goes in that general direction
ditto for the philosophy behind toner-background, toner-lines, toner-labels
i’m thinking less generic w/ the ingredients so there’s a raster layer of “transportation features”, “POIs”, etc. (though labeling could be an issue). vector tiles have the potential to be part of the solution
(digressing again and about to be in a meeting, so i’ll be distracted shortly)
Ian Ross
@ian-ross
Feb 24 2016 18:33
OK, Map Stack is quite nice. That just wasted a good five minutes...
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:33
only 5?
it must be rendering quickly ;-)
Ian Ross
@ian-ross
Feb 24 2016 18:34
So far. Bookmarked when needed for further procrastination.
Seth Fitzsimmons
@mojodna
Feb 24 2016 18:34
it needs more ingredients to be more interesting
Kate Chapman
@wonderchook
Feb 24 2016 18:34
I think I could procrastinate with doing weird things to the watercolor rendering for a long time
Ian Ross
@ian-ross
Feb 24 2016 18:37
One thing I've thought about a bit in other contexts is parameterisable models of "painterly" rendering of things, so you could have a knob to turn to go from Renaissance chalk drawings (à la Raphael and Leonardo) to naive Pieter Brueghel style, to colourful Futurist rendering (something like Umberto Boccioni), or anywhere in between. There's been some work done in using neural networks to capture parameterisations of painters' styles, and some people have managed to run those models backwards, to re-render photos in "painterly" style. Maps by van Gogh, anyone?
Atlas rendering is working on the demo site now. I'll try to fix that GeoJSON thing later today or tomorrow.
Kate Chapman
@wonderchook
Feb 24 2016 18:40
sounds cool!
I'll wander around the demo site a bit more
and test it
Ian Ross
@ian-ross
Feb 24 2016 18:40
s/test/break/, probably!
Alan McConchie
@almccon
Feb 24 2016 18:44
:wave: @linzjax, following up on what @mojodna said, if you or anyone else wants to write any blog posts I can help to post them on http://fieldpapers.tumblr.com/ and elsewhere.
Lindsey Jacks
@linzjax
Feb 24 2016 18:49
Great, thanks! I'll let you know once it's finished
Ian Ross
@ian-ross
Feb 24 2016 18:59
Thanks, everyone. I think we have a good list of tasks for the next few days -- all the small changes @mojodna mentioned, the GeoJSON stuff, the UTM grid overlay, better layout of the map and grid controls and possibly switching to the MapZen geocoder. I know Seth said these things are just "polish", but they're all quite small issues that can make a reasonable difference to the look of the thing, so we'll try to get them done!
Seth Fitzsimmons
@mojodna
Feb 24 2016 20:14
yay!