Platform Platform Platform

I don't think one can compare Canonical to Haiku. I mean - Canonical was started by a zillionaire willing to put a lot of money into the project. If Haiku had zillions, they could have a ports list as long as NetBSD!

it isn't a generalization -- an entire industry (the vertical monopoly hardware/software vendor) doesn't exist anymore outside of apple, and emulating it is just not a good move. the appliance computer, much as it's been foisted upon us since the 1970s, hasn't caught on.

yes, i would say smartphones and tablets aren't appliances but simply embedded personal computers (we use them for very general tasks, and if you've handled a low end smartphone it becomes evident the "phone" part of a smartphone is nearly an afterthought), and whatever i might consider a games console, haiku isn't it.

ARM chips are nearly ubiquitous, it's true -- but largely in embedded and industrial applications such as, say, the onboard diagnostics system and subsystems in a modern car. where they exist in phones and tablets, they are, again, embedded systems onto which you cannot install your own software. you can't buy a single arm processor at retail, find a motherboard with a matching socket, install RAM of your choosing, place it in a case with a power supply of your choosing and then upgrade any of those parts at will later on, so already it lacks the flexibility of the x86 platform; manufacturing an ARM system means designing an embedded system and manufacturing boards, which from the outset is prohibitively expensive and that new centralization means no one who finds out about your system online can just download software for free and try it out, instead they must put down money, sight unseen, for a custom embedded system that is still in the prototype stage, all the while building an entry level x86 workstation gets cheaper by the month; the only open ARM systems on the market are the raspberry pi, the beagle board and gumstix, which are all pretty great embedded systems entirely unsuitable to workstation tasks except when clustered in a manner better suited to a server os, at which point you begin to rival the expense of an x86 workstation anyway. working on a single embedded system, then, you would not only need a new kernel for ARM, you would also face new limits in userland, so why even call it haiku anymore at that point?

there's just no gain -- there's no installed userbase (you can't install on a google nexus or on the oxygen sensor in a volkswagen).

the examples in your initial post are x86 machines and again, as stated earlier, nothing's stopping anyone from making touch-based applications.

+1

I have a thinkpad x61 and it works great. Upgrading to an x230 in a couple of weeks, hopefully that will also work well.

If your argument with the HP Envy is simply form factor, I don't think you're going to enthuse many of the Haiku staff to do a touchscreen interface redo (yet). Touchscreens on mobile devices are now seen as (almost) a matter of necessity, albeit the blackberry doesn't do too badly without them. The value of putting touch on a desktop seems less compelling to me. Some of that mindset is just a reflection of the fact that I spend most of my time with a keyboard. A whole new generation of smart phone users has become more comfortable with the touch screen than the keyboard, and in the future they will dictate the desktop (or appliance) form factor. It's coming, surely, but right now it's a fairly small market share. I'll probably never want one, as I don't subscribe to Googlethink, appliances everywhere, and delegating all of my decisions to the machine. My brain needs a reason to exist. Neither will I follow Kurtzwell into the singularity, since I know the difference between my consciousness and an unconnected copy. In the future, the powers that be will use some-such as the new packaging for the coolaid.

Another thing to consider is that developers are glued to keyboards, and develop an inherent dependency on them. They naturally don't relate as easily to a finger pointing environment.

I do like the Haswell processor in the Envy (ten watts of power dissipation compares favorably with the high-end ARM at six or eight watts).

I am just a user but I thought I would chime in just a little. As for me I would love to see Haiku pick some 'brand' for lack of a better word. Maybe Asus or something and if enough money could be gotten to get the devs that kind of hardware at least of we would have something that was supported. I don't want too see Haiku only support one thing and a lot of people will say it supports x86, but the driver issue is something that can't be ignored. It will be almost impossible for Haiku to support all chipsets currently available. So, I would like to see some focus on some subset, maybe even AMD APU as it is a smaller target and seems to be more open source friendly then intel. It is just a headache to try to find hardware that isn't years and years old to run Haiku on.

Just my 2 cents.

David

conflated your arguments with ronald's for some odd reason, sorry, this thread happens roughly every two weeks.

oh sweet, insults.

i already wrote those things you want -- targeted applications (appliances have very specific application, you know), a touch interface -- can already be done, on hardware that is already supported.

why does your want have to preclude mine? as it is, right now, we can both get what we want, and even with a focus on a specific configuration (like, say, mac os), other configurations become supported because everything is standardized. i can go on swapping out parts in desktops and workstations, and you can go on hacking unibody x86 tablets. it's a beautiful thing, really.

My point is that while Haiku works, it doesn't seem to be fully supported/tested on any one particular piece of hardware. That is what I would like to see. It is like rolling the dice, some things will work great, some things will kinda work, some things not at all. I would love to be able to just go and buy a new laptop and know that not only will Haiku work on it but it has been tested and works to the best of it's ability on that hardware.

But, I know resources are limited and I am greatfull for the work already done. It's just my wish and will not ever keep me from trying Haiku.

Lets just suppose that Haiku supported Asus laptops, fully, all divers, power management, etc... I would buy that hardware. Or even just one particular series of hardware, but fully tested and supported. Most of the hardware on that site you linked to has "this issue" or "that issue" where Haiku works but not completely or is many, many years old and not availability anywhere but ebay. Note: this is not to say all of the hardware on that site is this way, but a lot of it is. This is a hard thing to do for the devs to do (support any given hardware completely), so I wish they would all have the same hardware series to develop on. That way we, the users, would have a series of hardware that worked flawlessly under haiku. that is all I am saying. But, I do not take the stand that this is a must do thing. I only think it would be a good idea for a nich OS like Hiku to do.

David

My understanding is that Intel based systems are completely supported by Haiku. AMD/ATI and Nvidia based systems may never get supported due to their closed source drivers. Intel based systems hold the best possibility for future hardware support, including hardware acceleration in the GPU. So far, I have not found an Intel based system which did not work with Haiku.

far as a cpu goes, intel and amd are pretty equivalent, evenly supported. as for graphics support, haiku currently supports only modesetting/switching on any video card, regardless of manufacturer and this includes intel's integrated graphics. gallium3d, at the moment, supports most graphics cards from nvidia and amd, and intel integrated graphics. intel's own open source graphics drivers aren't compatible with haiku. right now, there is no hardware acceleration regardless of graphics hardware manufacturer so you are not particularly better served by any over any other -- and when hardware acceleration is available, our support will come from gallium3d, the status of which can be seen here.

You could always ask the devs on the mailing list if they could make a page on this site that tells the hardware they develop on, the hardware they intend on supporting, etc. Some of this info can be found on the faq(http://www.haiku-os.org/about/faq#7)

Arctos might be on to something here. It would be much easier to attract developers if we had some pieces of hardware we can point to as being more than just barely compatible with Haiku.
Someone in the later posts mentioned Thinkpads as being used by most or many of the current developers. So it appears that programmers like them. (I'm not a programmer, but I do use one to write this.)
Thinkpads are and they have a reputation of being a decent quality hardware. They have a large user base and dedicated fan base. There is plenty of information and advice on the internet to go around (I can remember ThinkWiki, thinkpads.com, Lenovo's forums...) IBM put and now afaik Lenovo puts all Thinkpad user, repair and whatever manuals on the web, for free. So, good hardware and plenty of information about it.
Be Inc had quality hardware - BeBox - designed and built by some cost of time and money. Haiku could use instead an existing series of good quality hardware. I don't mean writing OS exclusively just for Thinkpads, obviously, that would be stupid, that's not the aim. Haiku should be aimed to as wide audience as possible - x86, ARM etc. But at this very moment... There are millions upon millions of older and newer Thinkpads out there. Older models are already quite cheap, so even if a programmer is more talented than wealthy he could probably get one. He would probably write more elegant and effective code on slower hardware anyway... ;) Having a solid piece of hardware that the OS can be run on is not a bad thing.
There are fixed number of generations, series and models of Thinkpads. When not focusing on the more obscure-experimental models, not the most ancient ones (maybe also not on the bleeding-edge ones, at least not right away) we could have quite nice somewhat fixed set of hardware that Haiku could be made run on, and run WELL. Quality OS on quality hardware - who could resist that! And they will come... :)

PS. If this were to happen, at some point, Lenovo... I think Haiku could never have too much support. ;)

True, but before HAIKU can get there, they need to insure that the system platform ( file system layout, packaging tools, etc.) are stable. With the recent PM changes, many developers are taking a "wait and see" approch to Haiku, once bitten, twice shy. They want a stable platform, or at least some serious statements that certain things will not change, before they jump back on to the Haiku bandwagon.

Platform, Platform, Platform!

I spend a lot of time dinking around with hobby, research, and other arcane operating systems - including even RTOS and bare metal setups. Everywhere I look I see the writings of people with favorable "first impressions" of Haiku, and have found that Haiku is regularly listed as the most promisingly viable "alternative/hobby" operating system out there. I think some day it will find its niche, and explode into stardom...

Maybe that "niche" will come about via some standardized platform of a type that takes care of all the hardware angles. Maybe some big company will buy it and launch the "niche" - then all the devs can quit their day jobs...