Slit-tear visualization. A simple idea for a powerful tool to resume and navigate a video. From 2008!
Watch the video to see several usecases, worth the 6 minutes
Remote
Sylvain Lesage
@severo@mastodon.social
mastodon
4.7.0-beta.1
Senior frontend developer
0
Followers
0
Following
17
Posts
Joined October 28, 2022
website:
huggingface:
github:
openstreetmap:
Posts
Real time public transportation anywhere in Europe.
Eg: next bus of line 8 at Esshaverstraße stop, Salzburg in 4 minutes
Replying to
@frontenddogma@mas.to
@frontenddogma@mas.to cc @ced@mapstodon.space on est beaucoup à ce stade j'ai l'impression...
Replying to
@severo@mastodon.social
I documented some details on my site:
https://rednegra.net/works/pelias-geocoder/
I also released the code in a github repo:
https://github.com/severo/pelias-vagrant
Replying to
@severo@mastodon.social
What I did:
- make it work locally (docker) with the example project for Portland. It requires downloading, preparing and importing 6GB of data to Elasticsearch, which is the engine inside Pelias
- make it work locally for Brazil. It's a whole different dimension, which requires about 75GB and 12 hours of preparation.
- import everything in a VirtualBox machine, using vagrant.
- send the machine (100GB!) to the client.
(...)
I just finished a small gig to create an offline geocoder for Brazil 🇧🇷.
I did it with Pelias Geocoder, an open source geocoder that relies on open data:
- @whosonfirst@mapstodon.space for the country / province / city boundaries
- OpenAddresses for addresses from authoritative (national, local) sources
- @openstreetmap@en.osm.town for addresses and interpolation along streets
(...)
New blog post: About Data Lifetime (2013)
I loved this blog post in 2013, and I still do
Software is trash (today, more than ever), data is the treasure.
Some quotes:
> He wants a one-off development that will do just exactly what he needs for two or three years. Then he will throw it away and get something new. It is always a one-off, never more.
> Your software will go away. Your data is going to stay.
Replying to
@scottish@datasci.social
@scottish indeed, for now it looks like I can bring this added value. Hard to tell if it will last for long though. Thanks for the comment!
Replying to
@severo@mastodon.social
AI also applies pressure on what I'm expected to provide. As agents are able to implement software so quickly, how should I adapt my prices and delays? And am I supposed to use agents exclusively to match some new velocity standard in shipping features?
If you're a freelance developer, how are you adapting your offer? If you're contracting freelancers, what are you primarily looking for? Please comment here or DM me, I'm very interested in your insights! 🙏
3/3
Replying to
@severo@mastodon.social
I've mainly developed data-heavy interactive web apps, which requires expertise with data, web platform, frameworks, software engineering. I feel like clients will soon be able to do as well using AI agents. And even if I'm reactive, the feedback loop is times faster with AI. Why would they contract me instead? For guidance? Temporal team augmentation? Because they trust my result more than using agents directly?
2/3
👋 Hey! Looking for advice. Please comment or DM, and share!
I'm currently taking some weeks off to think about the future of my business (I'm a freelance frontend developer, specialized in dataviz, with many years of experience).
AI has changed how we develop software, and I'm not sure what services I should provide in the next months or years. Why clients will be contracting freelancers like me?
1/3
Replying to
@bacteriano@mstdn.mx
Replying to
@kylebarron@mapstodon.space
@kylebarron@mapstodon.space wow even better! Thanks for the explanation, and let us know when the blog post is published!
Replying to
@kylebarron@mapstodon.space
@kylebarron@mapstodon.space my favourite part is the Arlequin debug mode!