Meeting minutes
start meeting
scribe
scribe jvanulde
Hi all, difficulties getting zoom started for all, please stand by...
<jvanulde> PeterR: Welcome to Vancouver
<jvanulde> PeterR: Presenting video introducing community group...
<jvanulde> PeterR: Opens floor to questions...
zcorpan: work on web maps research summary
zcorpan: research done for NRCan
zcorpan: httparchive used by browser implementers
zcorpan: static analysis, httparchive 7.5M pages crawled monthly. Data on html,css, js on every page.
zcorpan: no deep pages, only front pages
zcorpan: how many use, how they use maps: 16% of all pages use maps. Video for example is on 4%, Youtube 10%, still less than maps
zcorpan: 22% of maps use features / locations
zcorpan: range between simple and complex use cases
zcorpan: survey of web devs who use maps, make web maps. Promoted widely on Web.
zcorpan: received 78 responses.
zcorpan: key finding is that implementing interactions is the most difficult aspect.
zcorpan: no beginner friendly solutions
zcorpan: map bounds
zcorpan: people want to make maps accessible, but don't have the tools or don't know of them to do so
zcorpan: performance was covered, webgl power use
zcorpan: pan / zoom interactions, ability to define css for maps
zcorpan: from our expert interviews; key finding capability for better accessibility needs more work
zcorpan: switch projections on the fly needed
zcorpan: difficult to integrate maps into larger applications.
zcorpan: takeaways are perf, a11y
zcorpan: also interviewed browser implementers. <model> element is for viewing 3d models on a smaller scale / AR
zcorpan: interactive elements / attributes. Camera, without giving the page access to private info.
zcorpan: overlap of functionality with web maps, interaction may be different.
zcorpan: takeaway from browser implementers: want buy-in from mapping frameworks.
zcorpan: potential solutions: 3 starting points
zcorpan: start with mapml proposal, iterate
zcorpan: start w model element
zcorpan: 3 iterative approach. Bocoup recommends iterative approach
zcorpan: roadmap recommendation, goals and milestones.
zcorpan: a11y, perf, privacy
zcorpan: engage key stakeholders
zcorpan: have started doing some of the research. close gaps in Web map frameworks
zcorpan: polyfill. Open to working with Igalia to implement in browser engines
zcorpan: update mdn documentation
zcorpan: roadmap in more detail - need a link for the record here
zcorpan: reviews detailed road map.
zcorpan: roadmap
bkardell_: astounding amount of web maps; everybody knows that you rarely see the map on a homepage, usually assume it's buried in a site
bkardell_: possible to assume it's even higher than 16%
bkardell_: can't think of anything that is as widely useful. Hope that we can find people interested in solving these problems.
bkardell_: pan/zoom is so valuable in many contexts, not just maps.
bkardell_: also, my comment is that when you present something like this, it seems like something that's not going to get done.
bkardell_: other small things by comparison take a long time. The thing that I like about the iterative approach is that we may be able to find value and deliver it as we go.
zcorpan: very good point; for each capability that we have we might be able to deliver incrementally. Even if that's all we do it will be worth the effort.
bkardell_: html5 took many years, but we got pieces of it along the way. This plan is really good, i agree with it.
eparsons: To be completely honest, Google doesn't have a view on this topic specifically
eparsons: need to reframe what the problem is
eparsons: put a map on a web page, simple store finder
eparsons: Google sees this as a solved problem.
eparsons: so I think we need to re-state the problem.
eparsons: The future of the map is not a map, necessarily. It's about other ways of providing geospatial information.
… prob. not something that looks solely like a cartographic map.
… maps online at the moment aren't good citizens of the web.
… takes a person with geographical training to make them
… from pov of an autonomous agent, what they see in the JavaScipt is invisible i.e. nothing
… not machine-interpretable
… from my perspective, that's the problem that we need to solve
… how do we describe the location semantics in an html web page and make it available to the autonomous agent
… in other words, what is actually displayed on the map.
… should not fall into the trap, that we often do in our industry, of asking the question of the wrong people.
… by talking to GIS people, you'll get a bias in the answer.
… black hole of geospatial content
… putting geospatial on the Web; OGC has struggled with this fundamental problem: Web service standards are opaque to the rest of the internet.
… unless you already understand these services, it's impossible to understand the content
… W3C / OGC worked together to produce SDW Best Practices document, to make geo more webby.
… this has morphed into Cloud native geo content
… vector tile date in cloud native approach.
… from server POV, metadata should be easier to interpret
… need a simple tagging for maps, this is the fundamental problem.
… will help accessibility, valuable geospatial information is largely invisible in its native form
bkardell_: what are you going to render?
eparson: relative geographical terms are important
boaz: alt text for maps
boaz: shares web page with leaflet map
boaz: descriptions of features in "alt text" alternate page
boaz: maps are visualizations
<jvanulde_> PeterR: Regarding the map of Denver... we know it's Denver from the coordinates
<jvanulde_> PeterR: The BBOX coordinates are not available to indexing engines
<jvanulde_> edparsons: we are looking at this at our own level of abstraction, where is Denver? We can't rely simply on coordinates
eparsons: a sighted person can see that the map is a map of Denver, but how is that information exposed to machines?
zcorpan: a11y is a top priority, but it's not the only problem. For example, performance was something that people brought up, as not sufficient for maps.
zcorpan: scrolling is a fundamental feature of how browsers work. panning and zooming widget in the browser is needed.
<jvanulde_> PeterR: I want highlight the work that has been done at a technical level in the community group
<jvanulde_> PeterR: Use Cases and Requirements for Standardizing Web Map document was developed
<jvanulde_> PeterR: A matrix of capabilities was developed, allowing us to show that what has been implemented
<jvanulde_> PeterR: User documentation ongoing including interactive examples for the elements
<jvanulde_> PeterR: Documentation aimed at web developers
<jvanulde_> PeterR: GeoJSON API implementation in progress
<jvanulde_> PeterR: Chrome and Edge extension is available in developer mode
<jvanulde_> PeterR: Extension supports localization according the browser settings
<jvanulde_> PeterR: Extension supports accessibility preferences
<jvanulde_> PeterR: Exploring using links to navigate e.g. pan/zoom
<jvanulde_> PeterR: Geospatial linking to create a web of maps
<jvanulde_> PeterR: Improvements to the custom projections API
<jvanulde_> PeterR: Feature captions for accessibility, tab order is respected, you can navigate into the features in the map by tabbing
<jvanulde_> PeterR: Feature captions presented as a tooltip in the map, features are sorted by distance ascending from centre of the map
<jvanulde_> PeterR: Selecting the feature and hitting enter will provide feature properties in a popup on the map
<jvanulde_> PeterR: Accessible layer control, worked to make it more accessible, keyboard navigable and screen reader compatible
<jvanulde_> PeterR: Feature index, can be enabled in the page, a keyboard navigation technique, allows one to drill down into features
<jvanulde_> PeterR: It's really important to get your organization to join our Community Group
<jvanulde_> PeterR: Closing the meeting, join us on Thursday!