The part of KDE's UI guidelines that I disagree with in practice (although not necessarily in principle) is that an option is only worth having if it's worth having on by default. Which means that the default experience is having everything turned up to 11 with every edge-case-useful thing enabled. And that goes against providing a sensible non-scary initial experience. So hopefully they'll reconsider that guideline in endeavouring to make things sane.

View this comment - View article - View full comments

Nah.

If they wanted to do it, and they could see there's a business case for doing it, they'd have already done it.

These things are just so that they can say you're suffering because you didn't pray hard enough. It's just bollocks.

View this comment - View article - View full comments

Quoting: F.UltraWell here's to hoping. Wish I was a billionaire so that I could invest tons of money into creating the perfect Linux computer that was as slick as a Chromebook and have it been sold in normal stores. We need a new Canonical anno 2004 to pour resources into moving Linux closer to the customers.

It would need to be someone bigger than Canonical was. They did pour resources into a slick Linux device, worked out software packaging and distribution, took cues from Android about how to avoid the cruft of X11, developed an interface paradigm for convergence with the desktop... and then ran out of money, had to fire a bunch of people, and had to give up. So like Canonical, but with sufficient resources to be able to survive, say, 5 years more burn.

View this comment - View article - View full comments

The timing's interesting. They first showed their native build running on the Steam Deck on 29 September, I think, so that's about 9 weeks between having a build and releasing it. That might be informative for the cadence of other indie dev releases in the future.

View this comment - View article - View full comments

Quoting: tuubi

Quoting: CatKillerWhile AMD should get all the kudos and headstart from releasing their open-source cross-vendor scaling tech first, I hope Nvidia's similar one will also ultimately get integrated for those developers that want to use movement vectors to help their scaling. It would be a shame for that use case to only be covered by DLSS.

Does the open source Nvidia Image Scaling actually make use of motion vectors? As far as I understood, it's just a clever single-pass scaling + sharpening algorithm, much closer to FSR than DLSS. I don't see any mention of motion vectors in the readme on the github page.

Actually, maybe it doesn't. I saw references to directional scaling and assumed they were using motion information for their scaling (like they do with their anti-aliasing), but maybe it's just for different scaling in different directions.

View this comment - View article - View full comments

While AMD should get all the kudos and headstart from releasing their open-source cross-vendor scaling tech first, I hope Nvidia's similar one will also ultimately get integrated for those developers that want to use movement vectors to help their scaling. It would be a shame for that use case to only be covered by DLSS.

View this comment - View article - View full comments

Universe Sandbox is still Early Access, too, I think.

View this comment - View article - View full comments

It's been on my wishlist since GOL first mentioned it. Looking forward to playing it on the Deck.

View this comment - View article - View full comments

Quoting: undeadbydawnI remain disproportionally irritated that discounts are given as '-n% off', which means a price increase

Good news: here it's a hyphen rather than a minus sign - it says "<game> - n% off" rather than "-n% off" - and on Steam it just says -n%. So you don't need to be annoyed.

View this comment - View article - View full comments

Quoting: a0kamiI know DMABUF was quite a big deal for a while on nvidia GPU.

But I ended up changing side for an amd card.
Are amd also concerned by this dmabuf or is it specific to nvidia ? If not specific to nvidia, am I correct to assume the openness of radv helped getting it supported ages ago.

Nvidia had work a decade ago on dma-buf, but kernel developers decided that they wanted that interface to be GPL-only. So then Nvidia spent a couple of years reworking it to use fences, which wasn't GPL-only, and then kernel developers decided that they wanted to make fences GPL-only as well. I'm not sure exactly what happened after that.

View this comment - View article - View full comments