Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Picasa could do that too 2014, single threaded. (And is probably still better)


It does feel like we keep going backwards for functionality while resource requirements keep increasing.

Picasa had face recognition, basic editing, and a bunch of sorting features and ran on computers that had something like 1/50th the CPU performance.

Meanwhile modern replacements like DigiKam, Lightroom, ON1, and so on are incredibly bloated with very slow UI response even on a very powerful desktop PC.

Although Immich does at least work well, the webUI is extremely fast and loads images instantly.


which is odd, because nearly every modern cpu has some sort of hardware acceleration available for AI stuff... has the tech stack just become so convoluted that we're actually going backwards?

I do with they had OSS'ed the code for at least the desktop app, but maybe there were licensing issues? or just no one cared.


And is usable for large libraries. Digikam halts everything to a freeze if you use more than a very little amount of pictures. If your program is a photo management software, you need to be able to index/search/manage 1TB+ of images with ease.


> Meanwhile modern replacements like DigiKam, Lightroom, ON1, and so on are incredibly bloated with very slow UI response even on a very powerful desktop PC.

Darktable and ON1 feel decent snappy on my ~8 year old Desktop PC.

Though it was a looong time since I last tried Picasa.


Picasa was veeery fast on my dual core, 4gb ram, spinning rust laptop from 15 years ago. I’m trying to remember if I was using the laptop I got in 2006 which only had 2gb of ram or it was the upgraded one I got a few years later with 4gb, but either way it was incredibly fast on machines we would laugh at today.


We would probably also lough at the image resolution that was common back then.

I just watched a video of Picasa to jog my memory. It seems similar snappy as ON1 with 24 megapixel raw images on my 2015 PC.


Picasa was ok, you needed to train it a bunch if you wanted good results for facial recognition, that being said, I did use it for our high school yearbook to make sure I had enough pics of everyone and once I'd trained it with enough photos it rarely made mistakes.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: