Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

StarLite 12.5-inch Linux tablet

us.starlabs.systems

Post reply on HN
Print view View original

311–320 of 347 posts

Re: StarLite 12.5-inch Linux tablet

#311

Post by LeonM »

snvzz wrote: post #301
LeonM wrote: post #225

Earlier quoted context omitted.

> ... on Legacy ISA. Since when is x86 legacy? > I'd rather a RISC-V Pinetab-V, based on the same JH7110 as VisionFive 2. Have you actually used one of those? They are excruciatingly slow. Less performance than most Raspberry Pi. Maybe it works for your niche workflow, but for StarLite to succeed they'll have to build a _usable_ product. Also competitive ARM based alternatives (such as Snapdragon) require many closed…

>Since when is x86 legacy? Since the first commercial RISC chips obsoleted x86 in the 80s. >Have you actually used one of those? Yes, JH7110 on VisionFive 2. >They are excruciatingly slow. Less performance than most Raspberry Pi. Not my experience, owning VisionFive 2 and all raspberry pi B boards (1,2,3,3+ and 4), all of which have been used heavily. The CPU is much faster than RPi3b+'s while using way less power. O…

> Only a little slower than RPi4

That was my point. It is slower than a RPI4, which is already too slow to use for anything serious. You can't expect StarLite to develop a tablet that is 'a little slower than RPi4', because nobody but you will buy it. And yes, your workflow may be possible on a low powered machine, but that is not how the majority of customers use a tablet. Until there is a serious competitor to modern x86 processors in terms of performance (which the JH7110 definitely isn't), RISC-V is not going to be offered in any commercial grade product anytime soon.

> Here I just do not understand what point you're trying to make.

The point I was trying to make, it that the other alternative, ARM, is also not suitable for StarLite's mission due to the closed blobs. So really, x86 is the only viable option they have.

Re: StarLite 12.5-inch Linux tablet

#312

Post by predictabl3 »

ryantse wrote: post #111
eternityforest wrote: post #15

Why did they have to use a micro HDMI port? Couldn't they have done dual USB-C and made one the DisplayPort output? MicroHDMI is not a very nice connector, it's a lot of pins under mechanical stress from the heavy, rigid HDMI people generally attach to it.

From https://us.starlabs.systems/pages/starlite-specification >: Micro HDMI USB Type C 3.2 with Power Delivery 3.0 USB Type C 3.2 with Power Delivery 3.0 Micro SD Memory Card Reader 3.5mm Headphone Jack HDMI version: 2.0 USB-C Interface: Display Port (DP Alt Mode) USB version: 3.2 Gen 2 (up to 10 Gbps) Maybe I am missing something, but it seems to have Display Port for their USB-C?

It's very odd that their 12-port dock also shows a DP port in the image, but it's not listed in the specs.

But the 2x HDMI are visible + listed. :(

Re: StarLite 12.5-inch Linux tablet

#313

Post by forgotmypw17 »

TheDong wrote: post #100
forgotmypw17 wrote: post #80

Earlier quoted context omitted.

Another option is to create more progressive enhancement websites which work with such devices. Giving older devices new life is part of the reason I started developing my framework. For example, the websites linked in my profile are backwards compatible all the way to Netscape 3.0, while still supporting more modern features like in-place updating vote counters and adding dialogs to a page without reloading it. And…

I agree that these devices should be supported, but the only entity that can actually support them is apple. It's not enough for sites to work, the underlying software also has to be secure enough to handle the internet. That can come from Apple providing software updates to the device to keep it secure, or Apple providing a supported mechanism for someone to install Linux, or some other operating system that can be…

For orphaned devices, this is impossible, because Apple has already decided to abandon them. I think they can be still be useful for non-critical information browsing within a closed network of safe websites.

I think the accessibility necessity of supporting older devices is often underrated severely, while the need for security is overstated. New, currently supported devices people use day-to-day are also exploitable.

Re: StarLite 12.5-inch Linux tablet

#314

Post by dheera »

lproven wrote: post #192
dheera wrote: post #150

Earlier quoted context omitted.

USB-C is equally bad. I really wish they would have the rubber housing of the USB-C go inside the case like an IEC power connector. I break about a USB-C cable every week. Broke one today just by putting it in my backpack while connected to a power brick. Power bricks of the 1990s didn't have this problem. The barrel connectors were almost indestructible. You could drop bricks on VGA connectors and they'd still work.

Wow. I have never yet broken a USB-C cable. Not one. When I used an iPhone for a couple of years, I went through approximately a Lightning cable per month, sometimes more. At one point I took a carrier bag of broken Lightning cables to the electronics recycling. You must treat equipment exceptionally roughly.

No, I treat equipment as any consumer would.

Like if my phone is low on battery I plug it into a portable battery and shove the phone and portable battery (connected by a USB-C cable) into my pocket, like any consumer would. I then go hiking, snowshoeing, biking, sleeping, like any consumer would. Cables have bent and broken in my pocket in these scenarios, among thousands of others.

My parent comment was yesterday; today's USB-C cable broke when it got stuck in an office chair's wheel caster, bending the sheet metal housing. Thi doesn't happen with ANY connector of the 1990s. BNC, RCA, DB-9, 1/8" headphone jacks, 5.5x2.5 barrel connectors, they ALL withstood office chair crushing and were all probably designed with consumer abuse like office chairs in mind.

I've even run over DB-9 and BNC connectors with heavy duty carts, they were fine.

Re: StarLite 12.5-inch Linux tablet

#315

Post by dheera »

jwells89 wrote: post #245
kennydude wrote: post #215

Earlier quoted context omitted.

Barrel connectors I found always fail within 18 months for a lot of laptops I've known people buy (mostly from cheap brands such as Acer). While not saying USB-C is a great choice instead. Personally I just wish everyone could have generic magnetic Magsafe style connectors.

Barrel connectors are easily one of my most disliked. Even if they don’t fail outright, they grow loose and finicky over time even if you’re trying to take care of them. On top of that, there’s numerous different sizes and pole configurations so even if an adapter is physically compatible, you have to look closely to make sure you’re not going to fry your device. Just terrible overall.

So just agree on 5.5x2.5 and federally ban 5.5x2.1. The biggest source of finickiness is 5.5x2.5 plugged into a 5.5x2.1 socket.

Inventing a new standard with a weak sheet metal connector and 19 pins to deliver + and - isn't a solution to the above.

Re: StarLite 12.5-inch Linux tablet

#317

Post by jauntywundrkind »

zlg_codes wrote: post #54
jauntywundrkind wrote: post #24

Earlier quoted context omitted.

I have a tablet with a i5-7200u (2016) with a small-ish ~34WHr (originally, well worn since!) battery that gets 7 hours usage easy. How is this "power hungry"? For an ancient quad core & inefficient old storage drive. As with 90% of everything, people take a bunch of highly visible indicators to decide their opinions & then shit on anything & everything that doesn't meet their set conception. People hate Electron bec…

I tried VSCode on a Librem 14 by Purism. It has a tenth gen Core i7 with 12 threads. With only very basic essential extensions like syntax, Git integration, and a Vim mode, I could type faster than the view updated. Electron may be great for people who want to write Web apps, but browsers are RAM and CPU hungry beasts compared to the GUI toolkits they're trying to replace. In short, which environments are running VSC…

Your poor performance seems unexpected in extreme. Folks at work are on m1's with quite a lot of extensions & everything feels snappy.

Having a huge git repo could poyentially hurt; would be interesting to test. Who knows maybe vim mode is partly to blame? I personally still mainly use vim, not vscode, so I don't know what usual suspects there are. It'd be awesome if there's good profiling that can show what plugins are taking significant time in the core loop. We have brutal eslint to apply but that for example all happens async. Once again, we're back where we started, of one bad situation - some random plugin somewhere - creating a lightning rod for disdain.

Re: StarLite 12.5-inch Linux tablet

#318

Post by jauntywundrkind »

RetroTechie wrote: post #153
contrarian1234 wrote: post #110

Earlier quoted context omitted.

My Lenovo Y700 does passthrough power. It also has a mode where it only starts charging if the battery is below 40% and it'll stop at 60% Unfortunately its not a phone. Its a mini tablet. But I find the size ideal for daily use - browsing，reading pdfs, small sketches . (no SIM slot though)

> It also has a mode where it only starts charging if the battery is below 40% and it'll stop at 60% Wish every battery powered device would have a setting for this. As opposed to always charging when charger is plugged in, and always charging up to 100% (non-configurable).

My sample size isn't big, but on Linux every laptop I've seen has amazing battery reporting information galore. Oh sure battery level. And various assessments of wear. Things like realtime charge or discharge rates.

But more notably, I think around half also have charge control. It's just been on/off. But it would be a pretty basic bash script to make this happen.

Re: StarLite 12.5-inch Linux tablet

#319

Post by jauntywundrkind »

RF_Savage wrote: post #109
mcny wrote: post #78

Earlier quoted context omitted.

We had USB 3 with the Samsung Galaxy Note 3 in September 2013. It is now almost a decade. If rumors are true, Samsung will announce a USB 4 (thunderbolt 3) phone before the end of 2024. I agree with the grandparent. While most people don't care too much for eGPU (?), I'm sure if we let people innovate, good things will come. My dream is much smaller. I just want whatever hardware circuitry is required to accomplish t…

One of the tragedies of USB-C is how it can be anything from dumb charging only to one's capable of 40GB/s data transfer among everything else.

USB4 mandates 40GBps. It mandates DisplayPort. These would be pretty helpful baselines to expect, reasons for consumers to want USB4: they know it will be fairly featureful.

PCIe transport ("Thunderbolt") and Power Delivery are both optional though, I think.

Re: StarLite 12.5-inch Linux tablet

#320

Post by snvzz »

LeonM wrote: post #311
snvzz wrote: post #301

Earlier quoted context omitted.

>Since when is x86 legacy? Since the first commercial RISC chips obsoleted x86 in the 80s. >Have you actually used one of those? Yes, JH7110 on VisionFive 2. >They are excruciatingly slow. Less performance than most Raspberry Pi. Not my experience, owning VisionFive 2 and all raspberry pi B boards (1,2,3,3+ and 4), all of which have been used heavily. The CPU is much faster than RPi3b+'s while using way less power. O…

> Only a little slower than RPi4 That was my point. It is slower than a RPI4, which is already too slow to use for anything serious. You can't expect StarLite to develop a tablet that is 'a little slower than RPi4', because nobody but you will buy it. And yes, your workflow may be possible on a low powered machine, but that is not how the majority of customers use a tablet. Until there is a serious competitor to mode…

>RISC-V is not going to be offered in any commercial grade product anytime soon.

That sentence isn't gonna age well.

>That was my point. It is slower than a RPI4, which is already too slow to use for anything serious.

I am still using my RPi1, but I guess my uses would not meet this arbitrary "something serious" tablet requirement.

Joking aside, the RPi4 is actually very fast, and its shortcomings are elsewhere than the CPU.

E.g. its GPU is grossly inadequate. Its video codec block can't encode. It lacks cryptography acceleration, so LUKS and SSL need to be done in the cpu with non-specialized instructions.

These are issues JH7110 does not have. The experience is much better with JH7110.

>The point I was trying to make, it that the other alternative, ARM, is also not suitable for StarLite's mission due to the closed blobs.

Ah. They buy the FSF nonsense where closed blobs need to be burned into a ROM for the machine to quality as "open".

Unfortunate, but I understand.

RISC-V is inevitable.

Post reply on HN