pouët.net

On booting Ubuntu on my SSD (about 2 seconds), I have no time to watch a nice bootup animation... So who cares nowadays ? ;)

1: Linux tends to go tits up often enough so i do prefer to know when it goes wrong. (Not linux fault per-se since it's mostly hardware being fucked up and depending on specific windows drivers, but linux also has a fair share of faults)

2: I tried linux for a while (even ported 2 demos to it) but found it overall to be underwhelming (Bad hardware support, immature software,etc) and stubborn communities. (I'm a coder and tried helping out but unless you're "in the circle" or something like that patches however simple they might be just falls through). So when neither hardware works or software is good enough, how can one be serious about it?

3: Partially to point 2, if someone did a demo or similiar it probably would not be accepted by the general grub or linux kernels. Specific ports like ubuntu might think that it's an good idea but then you'd be stuck in support hell updating the crap for every new version of grub and/or kernel, especially considering that loader/kernel/ui handoffs would need to be customized. (Look at win-xp f.ex., first you have the text loader, then the logo and finally loading while looking at the login-screen)

4: Even if your stuff was bug-free you'd probably still be blamed for boot-crashes just because there are gazillions of different hardware combinations that can cause problems during boot.

5: C64 cracks,etc was usually put on fastloaders when on tapes so that improved loading times. On the Amiga remember there being trainers and stuff appended that gave some extra value.

6: Cracks,etc was for GAMES that you got for FREE because of these cracks. There wasn't any loader on the workbench that would be considered a real "os".
Also on the amiga i think you had much of the "system" already in the kick rom (similiar "system" stuff becomes available in userland much "later" on linux kernels after most of the HW is initialized)

whizzter, linux isn't so bad if you stick to nvidia cards (using proprietary drivers).

just don't expect graphics stuff to work on intel/ati cards...

still this only further makes the idea of a boot screen more troublesome :)

jaw: You still need to have one of the "blessed" wifi-chipsets etc to get a stable system. It's a mess.

hm, my stationary computers don't have wifi, only intel server nics..

whizzter:



No one would force you to watch a demo. If you'd rather watch text scroll by, no one's going to stop you.

Demos would be an option, for those people who like them.

There are plenty of demo fans out there, and even more non-technical users who don't give a rat's tail about Linux boot messages. Those are just the type to be most impressed by demos, too. So seeing a beautiful demo on boot would be good publicity for Linux.



Every operating system has its warts. Don't even get me started on the deficiencies of Windows. Let's not turn this in to an OS war.

My suggestion was an appeal to Linux fans. If you don't like Linux, and would rather use Windows, go ahead.



There are hundreds of Linux distros. Ubuntu has even explicitly requested demos. I think the reception by at least some Linux users (like me, a fan of both Linux and demos) would be very positive.

In any case, you won't really know what kind of reception it will get until you do it.

As for support, it's like any software you write. If you don't want to support it, don't. Even if you don't, at least the demo will run on the system you wrote it for, and maybe even on some others, and some people will enjoy it. Hopefully it will encourage more people to write more demos for Linux.



If there's any blame, it should be easy enough to find out if your demo is causing the problem: simply boot without it and see if the problem still occurs.

Anyway, even when writing regular demos you face the same risk of blame if the computer happens to crash while the user is watching your demo. How many demo writers lose much sleep over the possibility?



Trainers on the Amiga were a very small part of the scene. The overwhelming majority of boot demos did nothing but play graphics and sound (very cool graphics and sound, but nothing else except credits, as usual).



These are good points. It's clear that writing boot demos for the Amiga was probably a lot easier than writing it for the x86 platform. But since when has the scene been afraid of a challenge?

I once was happy that I discovered a 486 era motherboard with a graphical bios (non text based). I was telling this to my friend and he joked saying "So what else would you like to see? A BIOS menu with plasma effect in the background?" =)))

Anyway, he explained that a BIOS has a primary function, there is no need for fancy stuff in such a critical place. Or something like that.

And not to be misunderstood, you are talking about booting not bios, I just remembered this story that is not the same thing but similar.

Optimus: I think it would be great if BIOS screens had some style. Just because something is primarily functional doesn't mean it has to be ugly.

Unfortunately, commercial BIOS manufacturers have made it perfectly clear that they're not interested in making them look good, but only in getting them readable enough to work (if that). I think the only hope on this front comes from projects like OpenBIOS.

But you're right, we are primarily talking about the boot process and not the BIOS config screens. The objection (and response) is still the same, though. There's really no good reason to keep the boot process ugly.

Speaking as a fan of demos, I don't get why I'd want one running during 20-30 seconds of boot time. If I want to watch a demo, I can just start it once the OS has booted...?

There are already plenty of possibilities to add some visual bling to the Linux boot process if that's your kind of thing. Personally I think the standard boot process is much better than the one in for example Windows - much more simple and informative.

muse: Point 2-4 was to illustrate problems of doing something like that in the context of a community with diverget needs.

I made a small patch to get a graphics card working in linux, it took a good while for it to be integrated and during that while people with the same hardware would also have trouble.

Having your code accepted by the real packages is important if you're doing anything serious and don't want to keep fixing stuff over and over, and "non-necessary" stuff will have a hard time getting accepted (Just compared this to the Kon Colivas interactivity patches and the mess surrounding that). And if you don't get it into the real package you're left with 2 basic situations.

1: You write something and don't support it. If someone really likes your stuff the patch might be fixed to support new versions but if there isn't sufficient interest the code will stop working due to lack of interest.

2: You spend time maintaining a patch-set whenever new releases comes out that breaks it.

Now i don't know about you but i would not be doing that and taking care of other support shit without getting paid, if you had been a developer yourself you might prolly have seen why the issues of support matter. So for this to have a real chance it'd basically amount to someone here getting a full time job at ubuntu or being a zealot with infinite time.

And sure, there is some people that might do this and even keep it up for a time, but if they'd land another job i'm pretty sure that the linux support would be even more flaky over time.

The thing with shuttleworth talking at a demoparty was probably not him offering a job, he's a businessman and likes linux because it enables him to get stuff for free. If he was SERIOUS about this he should just contact some demosceners and offer them a decently paying job.

Just setting up an OpenGL window, that switches from fullscreen and windowed is an hell using x11lib, let alone a full blown demo. It's a documentation mess linux.

Xlib is a mess, but it has little to do with Linux, since X is way older anyway. Why wouldn't you use SDL in the first place? Pure GDI isn't exactly fun, either.

I use glfw for window handling...

How are you planning on using any of that fancy stuff during boot? :D
Anyways, I agree with the "modernists" here; if you have time to watch some pretties during bootup, your OS/hardware sucks.

Just to let you know muse, I support your cause. Don't feel so downhearted.

so, considering only doing vesa/vga stuff, (fading in a picture for start) I guess plymouth provides a working environment for this.

Just for fun I'll look into it, how the framework for coding in it works etc.

I'm tired of waiting for my food to cook. Has anyone thought of putting a demo in an oven door?

It's not supposed to be fun, it's supposed to have documentation, plus I don't need GDI to create an OpenGL window on Windows.
And SDL is just another dependency. The last thing I'll want on linux. Just try to compile and run an old Linux demo that depends on an old version of a library.