TC-wide discussion on symbology and annotation

[ogcscotts]

No description provided.

[jyutzler]

I'm not entirely sure what this is getting at or where it comes from. What I can do is give you my perspective of where I think we are at, based on my work on the Portrayal Concept Development Study.

• This is a high priority throughout the membership and other stakeholders
• OGC is committed to solutions that can be applied across the OGC baseline (W*S, GeoPackage, Vector Tiles, etc.)
• The current versions of SLD and SE have some flaws that make them unsuitable for long-term use but an effort to revise them is in progress
• Document-based approaches (SLD, etc.) are limited in utility but semantic-based approaches (being developed in OGC Testbeds and other activities) have the potential to mitigate the complexity of multiple sets of styling rules, multiple data sources, multiple possible map products (including thematic maps), etc.

[prushforth]

OGC should take a look at CSS for feature styling. CSS is used by SVG today, perhaps it is not too far from being a useful vocabulary for map features? If it is not suitable for geographic feature styling, why not? Can we address those issues with the W3C community? I think it's important to recognize overlapping standards from other domains and not 'reinvent the wheel' where possible, and CSS has global adoption already (maybe that makes it impossible to change, but we should assess that).

[jyutzler]

This is not a new idea. In fact, in this decade alone MapBox has developed and abandoned CartoCSS, a CSS-based language. https://blog.mapbox.com/the-end-of-cartocss-da2d7427cf1

[prushforth]

Not all old ideas are bad ideas :-)

I believe CartoCSS is still in use by Carto. Anyway, Tom identifies two issues with CartoCSS:

1. Non-standard CSS syntax was a mistake
2. The need to draw features more than once, whereas CSS which works with HTML, only needs to draw stuff once. I don't know how SVG gets around that, but it must do somehow.

I'm sure it doesn't work for vector tiles, but vector tiles are an evolving thing so targeting that could be tricky from a standards POV

[PeterParslow]

Some (UK MOD, for example) find ISO 19117 Portrayal a good way to manage a set of styling 'rules'.

Any group should also look at MapML, which has been presented to the OGC/W3C Spatial Data on the Web Interest Group. Terse notes of meetings at https://www.w3.org/2017/sdwig/, MapML usually discussed in their 'plenary meetings'; MapML itself at https://github.com/Maps4HTML/MapML.

I haven't personally looked at it much, but more active members of SDWIG have.

[ogcscotts]

merge with #47

[luciocola]

Sorry to insist and something like this to start ? It gives a practical solution and maybe we can build from there ...
https://github.com/spatialillusions/milsymbol