- Mapcomplete hack week? (Mid-july in Ghent?): T -> probably not, away 'till 13th
## What has been done?
- P: More refactoring on the maplibre branch
- P: Experimentation with MapRoulette to import _big_ datasets (openbenches.org)
## Open Questions
- P: styleguide! The MapLibre port is getting to a point that styling is needed to finish it up. T: ticket is in the works, Tobias will ask the designer to plan tomorrow - will probably have something to talk about in two weeks, maybe even sooner
- P: style: how to show the 'currently selected'-element on the map? Previously a somewhat ugly red 'blob', but better options must surely exist
-> Add a border to the builtin icons (pin, square, circle, teardrop)
-> Use a good fallback if no default icon is used
- P: some smaller, hard bugs - any hints? NO
## Planned todos
T: information architecture spreadsheet
# 2023-04-05
## What was done?
- P: Refactoring with MapLibre&Svelte in progress, small demo
- T: More design work
## Design updates
### Add new item
# 2023-03-22
## What was done?
- Svelte is in production!
- some cleanup of the themes, improvement of the documentation
- review of 'setting up the development-environment', tested on a fresh machine, some tweaks fixed
- user census results are published
- Advertisement-theme is merged and deployed
- Work on MapLibre has started (T: feel free to ask questions)
- T: contact with OpenCage
## OpenCage advertisement
Tobias did a follow-up on this
- Target someone with a big community and let them do the advertising
- "MapComplete is a great tool for organisation and topics if they want to work with OSM and want to keep data up to date -> Use MC as a tool for your audience"
## Geocoding by OpenCage
- Cities and neighbourhoods only, too limited
## Wireframes and next steps
- Via screensharee, see https://www.figma.com/file/j0Dxgbl7yx5RVqAWQbf5zH/Wireframes-MapComplete?node-id=242-7735&t=LdSserLloT9X19G4-0
P: user survey result in draft; option to show _all_ questions at once; personas
R: one Svelte-component (all_tags-panel)
T: Wireframes together with Heiko
## Planned
P: publishing user survey results, then Svelte
Is there a usecase already for someone visiting via an external website
## Wireframes by Tobias
Tobias and Heiko looked to one user scenario and discussed broader changes
Which parts are good, which parts need changing? Look to the bigger picture
How can MapComplete be adapted to various contexts (e.g. embedded in the website of a local organisation; or embedded in the website of a global NGO about e.g. drinking water?)
Attempt to create a (mini) frontend style guide
### General flow
The wireframes questioned about how the general flow should be (e.g. being greeted by an explanation + go to your location via search/gps sensor)
### Menu system
Tobias proposes to:
- Have one (labeled, thus with text) button to go to the theme introduction
- Have one menu-button, which allows to go to user settings, theme switcher, privacy policy, copyright, community index, buttons with various tools ...
Pieter agrees that this needs an big overhaul
### Add a new point
Tobias proposes to add a pin at the bottom left to start the flow to create a new item
Pieter refutes this, as:
- it is not very discoverable
- you'd still have to click the map for a rough location
- the pin can be mistaken for a data point that just happens to be there
- the pin (with a plus symbol) can be mistaken for the 'zoom in'-button
Such a label should at least have a label
### FeatureInfoBox
The infobox might be placed on the right, some tweaks to the image element (and some general restructuring) could be helpfull
Buttons to close/mark too hard/... have appeared on develop: https://pietervdvn.github.io/mc/develop/maproulette?z=15&lat=51.21127&lon=3.219745&language=nl#59417173
## Svelte
Deployed on https://pietervdvn.github.io/mc/feature/svelte/
Issues with scss:
classes with colons (e.g. md:w-full)
Feature branch feature/svelte should be mergeable
! Check a fresh install based on the docs - this will change
Update docs about the framework as well
Update docs/architecture
Use 'context' to pass around state instead of explicit passing around?
But: explicit passing of parameters is boilerplate, but prevents 'forgotten' pieces of state and makes testing easier
Tobias: not a fan of persona's as they tend to be a bit useless (and a waste of time). You want to get clarity what the product is about: the core audience
See them as 'core context' and 'core usage scenario' with a goal. Don't 'personificate' them to much, but writing everything down as a list of 'core features' and 'non-features' are important.
- Beginner friendly
- Does also complex tagging in a friendly way
- No JOSM/Vespucci where you can shoot yourself in the foot
The core idea is well established, but to execute it is the trick. It takes an insane amount of love and work to get all the details just right. The polishing part is important (sidenote: a frontend framework that is familiar would help the polishing).
What is the timeline? Persona's done in February, some user tests in Framework and some in March.
Tobias will send some recommendations: "rocket surgery made easy"
Do a UI-overhaul for a more consistent UI (with a designer) once the framework is in place. An expert review will yield the same results as user tests; maybe first a cleanup phase.
A11y: more then just for screenreaders and blind people (e.g. high contrast). Low hanging fruit: follow the web standards; start with defining a target.
T: will talk to Heiko what and how much contributions; Tobias might do an expert review of MC (if feasible)
165 answers, 82% male, 10% female (hugely underrepresented), 7% other genders (according to [this article](https://www.washingtonpost.com/dc-md-va/2021/06/22/first-population-estimate-lgbtq-non-binary-adults-us-is-out-heres-why-that-matters/) and 330 million inhabitants in the U.S., there are on average 0.36% genderqueer people - so hugely overrepresented)
Ages: normal distribution around 40-50
Some conclusions at first sight:
1. MapComplete isn't very well known, some confusion with StreetComplete
2. Search sucks
3. A few feature requests or requests for tools that already exists
Better review will come in a few days.
What are good questions to ask next year? How to improve the survey?
T: guideline: the more specific the question; the better the answer. We try to stay away from general surveys; not a lot of actionable input. What are you gonna do based on these answers? Already broad in what it can do, lot's of invisible features that are unpolished. Tagging is so diverse and is hard to cover the edge cases, lot's of work in the details. Rather focus on what is already there and make it more visible (e.g. maproulette integration, german guys with a contract somewhere: for example: guided imports as feature will not be surfaced by a survey and will not serve as product guidance ).
