2024/Brighton/pictures: Difference between revisions
(archive etherpad) |
m (→Participants) |
||
(One intermediate revision by one other user not shown) | |||
Line 14: | Line 14: | ||
* {{lazcorp}} | * {{lazcorp}} | ||
* chris | * chris | ||
* murray | * {{murray}} | ||
* adactio | * {{adactio}} | ||
* {{qubyte}} | * {{qubyte}} | ||
* jakob | * jakob | ||
Line 25: | Line 25: | ||
== Notes == | == Notes == | ||
Scout has some open questions to begin with | Scout has some open questions to begin with | ||
* Large amount of photos, | * Large amount of photos, separating them from content | ||
Paul has example of hosting images, as hosts lots of illustrations | Paul has example of hosting images, as hosts lots of illustrations | ||
Line 34: | Line 34: | ||
* 4¢/month! About 1000 users a week | * 4¢/month! About 1000 users a week | ||
Ros asks about concerns hosting with Amazon, Scout asks why can’t be just images in a | Ros asks about concerns hosting with Amazon, Scout asks why can’t be just images in a directory (Amazon doesn’t feel like that) | ||
Mark (qubyte) says hosting video even worse/more challenging | Mark (qubyte) says hosting video even worse/more challenging | ||
Line 47: | Line 47: | ||
Ros asking why put effort into uploading full quality image, when users don’t get to experience that resolution | Ros asking why put effort into uploading full quality image, when users don’t get to experience that resolution | ||
Any tips on | Any tips on labeling structure | ||
* Paul stores filename without suffix in database | * Paul stores filename without suffix in database | ||
* paulrobertlloyd uses a file structure that follows his post URL format (yyyy/ddd/…) | * paulrobertlloyd uses a file structure that follows his post URL format (yyyy/ddd/…) | ||
Line 57: | Line 57: | ||
-*Says doesn’t trust S3 more than any service | -*Says doesn’t trust S3 more than any service | ||
How easy is it to get photos off | How easy is it to get photos off Instagram? | ||
* Murray says not easy, the photo quality isn’t great | * Murray says not easy, the photo quality isn’t great | ||
Latest revision as of 16:12, 12 March 2024
Pictures was a session at IndieWebCamp Brighton 2024.
Notes archived from: https://etherpad.indieweb.org/pictures
IndieWebCamp Brighton 2024
Session: Pictures
When: 2024-03-09 14:00
Participants
- Scout
Paul Robert Lloyd
Paul Watson
- chris
Murray
Jeremy Keith
Mark Everitt
- jakob
- letorey
David Shanske (remote)
Tantek Çelik (remote)
- Add yourself here… (see this for more details)
Notes
Scout has some open questions to begin with
- Large amount of photos, separating them from content
Paul has example of hosting images, as hosts lots of illustrations
- They are fairly low resolution
- Hosts on Amazon S3
- Set cache control set to a year and a day, saves on cost of hosting
- Uses Amazon Cloudfront service, again cuts cost
- 4¢/month! About 1000 users a week
Ros asks about concerns hosting with Amazon, Scout asks why can’t be just images in a directory (Amazon doesn’t feel like that)
Mark (qubyte) says hosting video even worse/more challenging
Ros is putting together a spreadsheet of hosts/costs and plans to share it
Paul is hosting images on GitHub and using a resizing service Mark has come round to resizing images beforehand (posting it from his phone using shortcuts), saves all image transformations to GitHub (JPEG takes up the most size, AVIF etc. formats are smaller)
Chris commits original images to GitHub, creates resizes images on build
Ros asking why put effort into uploading full quality image, when users don’t get to experience that resolution
Any tips on labeling structure
- Paul stores filename without suffix in database
- paulrobertlloyd uses a file structure that follows his post URL format (yyyy/ddd/…)
- why don’t people use YYYY-MM-DD!!
- overlap between keeping personal archive/backup versus sharing photos with others
- GitLFS is a solution – but it’s a horrible thing! Has a number of pain points
Jeremy says there’s a tool that creates a mirror of Flickr (Flickr Mirror?), could put that on a sub-domain -*Says doesn’t trust S3 more than any service
How easy is it to get photos off Instagram?
- Murray says not easy, the photo quality isn’t great
What is Amazon CloudFront
- S3 is the storage
- CloudFront is edge CDN (so if somebody in Australia, don’t need to get file from London)
- There’s a cache between S3 and CloudFront (so only needs to be fetched once)
- csswizardry has a great article about caching
- CloudFront allows you to invalidate caching
There’s no one-click solution for storing/caching/delivering images - barrier to entry