On monday I'm presenting a paper on solar.lowtechmagazine.com during the 'Computing within #Limits' conference.
In the paper I'm describing the design choices we made, such as designing for unavailability and how they were based on reducing a few key metrics such as data transferred, calculations per request and use of third party services. This opens up a way of thinking about #degrowth and ICT. We should be looking in to that waaaay more.
The paper also tries to implicitly answer the 'how do we do sustainable web design?' question and argues aiming for reduction (of bytes transferred, computation and infra required) as useful rule of thumbs guiding a design. That is to say, solar protocols, 'low tech look', 3.5MB dithered images, green CDNs etc ain't it.
However, it also self-criticizes solar.lowtechmagazine.com for the fact that it is so unreproducible and a very limited and 'easy' use case. So yeah no answer but thinking through degrowth should definitely be considered as the direction of the answer..
@l03s is also presenting about the terminology around practices of making alternative computing infrastructures which think about #energy and #ecology. What have they been called historically, what have they been called in academia and more interestingly, what have they been called in grass roots technology practices?
Her paper "has the goal of making the many counter narratives to the one of green capitalism visible, show the differences and similarities between them, create bridges between different fields and contexts and demonstrate the wealth this diversity of practices, and the thinking that informs them, brings"
She used a poll on mastodon and the subsequent discussion (https://post.lurk.org/@l03s/105713332161647316) as a way to surface the terms and she discusses some of them: #permacomputing #smalltechnology #lowtech and appropriate technology.
The paper really centers the diversity of approaches and the existence of alternatives not just by mentioning but also engaging with them. @viznut, @neauoire, and @calcifer are mentioned as are #gemini #LTM and #solarpunk.
The paper is useful for those that want an overview of current discussions on the topic because it points to the many places and ways these discussions take place.
@rra Will be interested to read. Hope it enlightens me as to why the magazine stuck with awful looking dithered images rather than regular old scaled/compressed JPEGs. I'm pretty sure it's an aesthetic decision, but it's a weird one if so. Because with better images the whole site could have been converted to "solar" whereas right now that's the main detraction.
@seachaint the paper doesn't discuss that but points to where it has been discussed instead.
Just ran some test with one 800x600 picture, comparing the jpg from the non-solar version of the site to the dithered png, and to the output of
> pngquant 32 [name].jpg
original jpeg: 49.3 kb
dithered monochrome: 34.8kb
quantized 32 colour png: 34.4kb, and barely distinguishable from the jpeg. I bet it would look even better if generated from a lossless image, i.e. without having to encode the jpeg artifacts.
Anyway: pngquant is a cool thing and very helpful to reduce many an image file. In many cases significantly better than jpeg.
@Mr_Teatime @rra I imagine the differences would be really varied between PNG and JPG depending on the source of the image: photos vs. digitally generated etc. Posterised/quantised PNGs would be ideal for some kinds of photos though, I like that idea. What's your pipeline for quantisation? I'm bad at this and couldn't figure it out last time I tried. :)
I just use #pngquant. Works tons better than e.g. GIMP.
> pngquant [n] <inputfile>
generates a png file quantised to n colors, using Floyd-Steinberg dithering, and compresses better than most others. You can turn dithering off with the --nofs option.
To get line graphs in svg or eps format into presentations, I export to png with 600dpi (no anti-alias!), then quantize if needed. The result is often smaller than the EPS, and always smaller and better-looking than jpeg.
That's not going to work equally well with any photo -- usually photos do compress better in jpeg, but in this case it works very well.
@rra I may be missing something - is it not possible to watch the papers being presented? Or is the presentation format private/“internal”? Couldn’t find any information about this on their website 😓
@amatecha Honestly, I don't know. I don't see any public invitations on their site either. It could be it is only for those that submitted something because it is a workshop? Perhaps you can drop them a mail?
It definitely looks interesting. :D
I, too, couldn't find the links for tickets, or attending.
Could you ask the organisers if the talks will be recorded for later watching please?
-- Also, have you come across the Solar Protocol approach towards solar-powered websites?
Distributed web-hosting where the website is displayed from the solar-powered web-server with the most power.
Combine with co-hosting other websites and you've got better uptime for everyone.
@rra lovely read thank you for writing this
Welcome to post.lurk.org, an instance for discussions around cultural freedom, experimental, new media art, net and computational culture, and things like that.