Hacker News Reader

> I think there is something important that we lose when we increase the accuracy of a representation at the expense of the elegance of the representation.

But that's not what's happening.

Botanical art is still highly valued in the botany field because field conditions often do not allow a completely accurate replication of a specimen via a camera lens. In many cases, the only way to properly represent the plant in question, to a colour-accurate, detail-accurate degree, is by a drawing. Source: My mother is a highly qualified botanical artist and a member of the SBA.

Yes, still valued, but isn't the shift what the linked article is about? "Botanical illustrators like Tangerini are rare and becoming as endangered as some of the plants they draw."

It's not about being valued, they're literally indispensible for the documentation of plants, there are no alternatives that match the quality of the documentation that they produce.

As for them being rare, part of that is due to education and the infrastructure around it. When my mother was growing up botanical illustration or scientific illustration was not listed as even a potential career choice, and she wasn't allowed to take art and science at the same time as "that wouldn't be realistic". Obviously things have changed since then but given that so few artists even recognise it as a career choice has an impact on the number of people who decide to go down that path.

It's like how the UK is running out of high quality vellum, used for legal paper and archiving, because nobody realises it's a viable career choice for them to do. There's literally a free, funded apprenticeship to do it, because they need people to do this job.

My local university had these old books that contained maps of the area, going back to the 1800s. One of my favorite parts of them wasn't even the maps, but the intricate calligraphy and lettering, big brazen impact hand lettered fonts. That kind of thing doesn't exist in the modern flat and minimalist design style.

You might appreciate the work of this artist / mapmaker:

https://www.antonthomasart.com/

I'm sure there are (a few) others like him, but I listened to an interview recently where he talked about completing his map of north america.

What illustrations do is focus on the highest information details (e.g. the shape of the leaves) and de-focus the lowest information details (the leaf is probably green, you don’t need to color that in). That way if you’re trying to see what’s unique about the leaf, the illustration is much better.

How far away are car windows with baked-in AR? Highlighting the mapped route ahead of the driver by inserting a sequence of semitransparent arrows will render any non-AR map style obsolete for in-car use by drivers.

Oh I used to like that one where it looked like a pencil sketch will the main arteries and intersections and roads you had to turn in to.

One thing it was missing was landmarks i think.

> I don't need to necessarily know the details of every twist and turn in the road so much as a schematic overview, even when getting turn-by-turn directions

Some cities' street layout aren't so neat that turn-by-turn directions suffice. I often need to zoom-in to distinguish different turns by looking at such details like the curvatures of the streets and their relative placement. Even with the detailed map, I can make a mistake between similar-looking turns.

Abstractions and Generalizations are useful. Principal Components analysis is useful. Dimensionality reduction is useful. Models are useful. In the last, we had to do these things because our imaging technology storage media couldn't hold full reproductions. Now we have better imaging tech and storage media, so we need to consciously remember to do the dimensionality reduction modelling that we took for granted in the past.