19 September 2019


satakagi, zcorpan

Meeting minutes

satakagi: kddi corp is cell phone carrier in Japan. I attend SVG wg as member. Interest in standardization for maps; but v difficult
… webis decentralized but services such as huge mappers is not
… we map info such as power lines, and other infrastructure. Linked open data is a use case for our company operation

bkardell: with Igalia and Open JS Foundation

mark baker: audioeye inc. is an accessibility company with particular interest in maps and spatial information. have been interested in how accessibiltiy
… will work in maps and vr / xr. Have heard lots of presentations this week about this. Everyone is interested, but are more or less at the same phase
… of research and development

AmeliaBR: shuffle agenda jump to work we've been doing this year. if more join we can go back to introductions

AmeliaBR: group has been around 5 years. has had a hard time involving browser devs.

AmeliaBR: when I got involved i thought we needed a more explicit statement of what and why

AmeliaBR: existing tools largely solve the same problem: this is a cow path in webstandards terms

<AmeliaBR> https://‌maps4html.github.io/‌HTML-Map-Element-UseCases-Requirements/

AmeliaBR: this is the use cases and requirements see link

AmeliaBR: we wanted to have a complete draft ready for tpac, this turned out to be too much work to get done in the time we thought

AmeliaBR: this is a work in progress, and we welcome more contributions to, either use cases or other contribs

AmeliaBR: applications of the use case: web site author expanding on use case, interact and progressively enhance the base widget

AmeliaBR: capabilities is a word we're using to describe the individual functions that add up to the total map viewer (not "feature" because overloaded)

AmeliaBR: looking at existing script implementations, do they each have this, how are the ui configured, are there commonalities

AmeliaBR: can we use keyboard, mouse, touch etc

AmeliaBR: once we've narrowed the capability down to a low level feature, we ask does this have application outside geospatial that we can
… collaborate on with other groups or standards and so not duplicate effort

AmeliaBR: are capabilities nice to have, critical or antipatterns.

AmeliaBR: idea is that eventually we'll put this all together as a list of things a native map viewer should be able to do, with a ranking of importance


AmeliaBR: we have an existing proposal that (Peter) put togther

prushforth: clarify this proposal is from the OGC community not just me.

AmeliaBR: what hasn't happend yet is contributions from the web side

AmeliaBR: looking to get more engagement from web standards and browser devs as this project moves forward

AmeliaBR: do a comparison of MapML proposal and SVGMap from satakagi/kddi

satakagi: https://‌www.google.com/‌url?q=https://‌github.com/‌satakagi/‌TPAC2019-docs/‌blob/‌master/‌MAP.md&sa=D&source=hangouts&ust=1569024126498000&usg=AFQjCNHec1V6FVVf2CQGvHKDtaifoMWsXQ

satakagi: updated the agenda with link to work contribution to maps for html

satakagi: do we see any change towards decentralized web mapping?
… gartner : decentralized tech is midway up the hype cycle
… companies and govs are responsible for social infrastructure
… discont between geo and web industry resulting in tech mismatch
… no business model for decentralized web mapping, esp for current web browser vendors
… kddi submission and dejure standardization regarding web mapping based on browsers native graphics
… tiling and layering module for svg 1.2 tiny
… dejure standardization, JIS is on svg 1.2 tiny
… updated features of svg based mapping
… decentralized web mapping aplication
… maps with large numbers of layers are controlled by complex javascript, but also centralized.
… js logic can be encapsulated in layers if organized as decentralized mapping applications
… the tiling schem being developed within the geo community is still developing and looks unstabvle.
… there are many issues
… tiling schemes are combined by the application
… more efficient tiling scheme example
… pyramid of quad tree tiling

AmeliaBR: thank you for a great overview of some of the high level features of what we need, such as layers, lining up based on geo coords

AmeliaBR: need a way to tile within a single layer,

AmeliaBR: need a way such as tiles and scale, to load appropriate data for the users view

prushforth: hyperlinks are the key to decentralization, we need permissionless federation of map content, leverage web architecture

bkardell_: question : layers are they independent documents?

all: yes

AmeliaBR: layers across domains, may limit some behaviours for such use cases.

AmeliaBR: everything based on script: you need to ensure yourself that you're not doing dangerous things.

AmeliaBR: have you (satoru) thought about cross origin issues with your solution?

bkardell_: is everything in one document, or different docs?

AmeliaBR: svg use is similar

simon: you want to do cross origin without
… proxy

satoru: iframes are similar security model

prushforth: different rendering model tho

simon: what do you want to do with mashups

AmeliaBR: we are doing exhaustive review of existing tools and common features/ limitations

AmeliaBR: what should we be able to do, rank them

simon: as part of that have you id missing primitives?

AmeliaBR: some things that we've discovered are very map specific

AmeliaBR: there are other much more general things such as being able to pan and zoom that doesn't match the current document rendering model

AmeliaBR: what's unique about maps is that you don't have an exhaustive list of the urls, you need to use templates

simon: have you considered using svg as a fallback?

AmeliaBR: satoru's model is completely based on svg, so may provide a path forward there

simon: realistically the fallback is going to be a 100% javascript map, not an element fallback

bkardell_: like the idea of svg has lots of stuff that you need, what's missing in svg

AmeliaBR: custom elements in svg

bkardell_: would be good but the media elements, where the thing that separates the requesting of data from decoding of data

AmeliaBR: they have a common api, part of the value is that it knows how to find stuff in the files and it is independent
… maybe we don't need a new map data format, maybe we just need to

prushforth: the issue is not the format, its the interaction with servers

AmeliaBR: query strings and form data is how the mapml metadata specs inputs that name parameters to the server

bkardell_: should be decoders instead of hypertext

satoru: mashup tool demonstration

simon: do we want to support projections

AmeliaBR: yes, the default is web mercator, not universally good, but

AmeliaBR: huge database of geometric projects, some are associated to standardized tile sets.

AmeliaBR: mapml uses standardized keywords limited though; may need a separate data format that can define the projection used in a data set

<AmeliaBR> https://‌en.wikipedia.org/‌wiki/‌Well-known_text_representation_of_geometry

simon: if want support diff projections seems like a different primitive

AmeliaBR: only asking for vectors to be transformed

satoru: webgl image textures

<AmeliaBR> More relevant link on well-known text: https://‌www.opengeospatial.org/‌standards/‌wkt-crs

AmeliaBR: once you break it down to those low level primitives

mikes: workshops don't have a fixed format. others have done it in the past where you submit position papers
… montreal june in coordination with the OGC meeting might be a good opportunity

mikes: goal is to produce a report leading to a working group, or at least an interest group
… if that's where you want to take it

mikes: I could coordinate, WEndy is coordinator of strategy. I would take to Wendy, try to raise awareness.

mike: i look for technical merit ; plenty of working groups have been launched over my strong objections.
… part of the goal would be to try to get some companies and orgs involved in bringing use cases and requirements to make sure the work aligns
… that's how we make this work revenue wise
… some enthusiasm around this in Japan, there could be other places and orgs

AmeliaBR: we are developing use cases and requirements hopefully done by next May, or hopefully in advance of when we have to announce the meeting

AmeliaBR: need more stakeholders to talk to us

mikes: too much work going on where even w3c team to articulate what the benefit is to actual end users. Come up with a statement.

mikes: odds of anything getting into html are relatively low.
… successful way to engage and the community around is to come with a problem, not a solution
… the default answer is to use custom elements

simon: need to say what are the missing primitives
… what are the cow paths that we want to pave. show that it would be worthwhile to add a declarative mechanism
… data showing there's demand

open source extension of some mini app

AmeliaBR: mini app goal is to enhance user experience of web apps bringing the web into closer competition with app platforms

mikes: use cases recommendation -> decentralization is a bit abstract. Its good to talk about things in more concrete terms.
… a compelling story is a user can take maps from different providers put them together
… need to put together a pitch

AmeliaBR: lots of things that can be improved, not just paved over

simon: might be worth conducting an mdn type survey. to find pain points, could make convincing people easier

prushforth: difficult to get people to see what the state could be

simon: need to design a survey to get at what the issues are

simon: maybe add a question or two to the mdn survey

mikes: hadley works for uk gov is in community of people who are advocating for more open data

AmeliaBR: good example, exploring open data in my city, some is in GIS and some is in google maps and that data don't talk

yoav: what prevents devs from amp - collection of web components - you could develop a web component

AmeliaBR: custom elements have been developed , but we haven't promoted widely. want / need to iterate before promotion
… as we get more confident that we have the requirements and architecture, promotion is on our road map.
… for the user need to get accessibility

mikes: have discussion about missing primitives with browser developers

mikes: accessibility talk to doug schepers, who's done a lot of svg related accessibility

Minutes manually created (not a transcript), formatted by Bert Bos's scribe.perl version Mon Apr 15 13:11:59 2019 UTC, a reimplementation of David Booth's scribe.perl. See history.


Succeeded: s/thought/though; may need a separate data format that can define the projection used in a data set/

Succeeded: s/google/some is in GIS and some is in google/

No scribenick or scribe found. Guessed: prushforth

Maybe present: all, AmeliaBR, bkardell, bkardell_, mike, mikes, prushforth, satoru, simon, yoav