Adrien Destugues sent an email to Haiku developers after the BeGeistert forum, addressing their inability to get R1 out the door, and proposed that they rededicate themselves to getting a beta out ASAP. He then asks a question that hangs above the heads of all developers of alternative and hobbyist OSes: is their goal “to create an operating system that specifically targets personal computing? Or have we evolved to the goal of a fun playground for OS-developers to play around with modern OS concepts?” He concludes … “I do think that the PC-landscape has changed dramatically since the inception of the project, and I also underscore that there is a clear lack of focus when it comes to accomplishing our current mission. I would go so far as to say that the severe lack of interest of developers into finishing R1 is a great indication in that there really hardly seems to be any place for a new (mainstream?) desktop operating system anymore? Even the Linux on the desktop guys seem to have ceased preaching their gospel.” That’s some sober talk that’s important for alternative OS fans to consider.
Is it too late to switch focus to HaikuIA?
I followed Eugena here from BeNews, and so it pains me to say this: the ship has sailed, and short of an intercepting vessel (e.g. beaucoup bucks infusion) there appears to be no way to catch it. Hope I’m wrong.
As great and ahead-of-its-time as BeOS was, it’s now an OS that hasn’t been updated in a decade and a half. That’s several OS lifetimes. GCC 5 or 6 will be mainstream before v1.0 at this rate.
How ironic is it that ‘Internet Appliances’ based on Android and iOS may have robbed this project of it’s last hope for any relevancy?
As frustrated as I was by the change of focus to BeIA in those times, I can’t help but wonder now: What if the FOSS BeOS community had continued looking to embedded mobile (i.e. not a refrigerator) computing?
It really is too bad that BeIA was about 7 years too early for the hardware that would have made it actually useful.
Edited 2014-11-06 07:24 UTC
Sorry to come with the sad analogy of cars but; Just because most drive regular cars (windows) and trucks (linux) there is no interest at all in any other mode of transportation?
I am fine if the Haiku project wants to keep it as their personal playground, they made it after all. But i also am one of those idealists that like something different, perhaps just for fun and perhaps for any other reason under the sun.
I see the personal computing industry moving away from the low-tech home space to a more power-user oriented focus in the desktop/laptop segments and leaving the others to dumbed down touch interfaces on low powered mobile devices and app-enabled TV sets.
“Sorry to come with the sad analogy of cars but; Just because most drive regular cars (windows) and trucks (linux) there is no interest at all in any other mode of transportation?”
I think this speaks to the question of whether Haiku is a toy for devs or an OS for users. See my continuation of the analogy below.
“I see the personal computing industry moving away from the low-tech home space to a more power-user oriented focus in the desktop/laptop segments and leaving the others to dumbed down touch interfaces on low powered mobile devices and app-enabled TV sets.”
I hope you’re right, but even someone like me who has an emotional attachment to BeOS sees very little reason to use a 2017 v1.0 Haiku at this point.
Continuing with your analogy, I have cars for personal use, I have trucks for work, I have scooters for light transport (Android)… wouldn’t that make BeOS something like those funny bicycles with the big wheel in the front and the small one in the back?
I’ll take a ride for nostalgia, but will I ride it to work? Hope so, one day, but doubt it.
Edited 2014-11-06 07:37 UTC
The problem is that without a huge infusion of cash (doubtful) they just can’t keep up with the changes in hardware.
Just look at how far we’ve come in the past 10 years…in 2004 virtually all computing was done on X86 desktops with slow single core CPUs with 512MB of DDR (if not SDRAM) paired with a 40Gb-160Gb IDE and a graphics chip that used 4-64 MB of RAM from the system. Now even the low end desktops are X64 multicores, usually APUs, with 4GB+ of RAM and as much if not more RAM for the GPU than the entire system had in 2004. Even on the mobile side you have multicore with ever expanding memory, systems that make the kind of hardware BeOS was made to run on look like an 8-track recorder compared to a DAW!
Between Windows, Linux, iOS/OSX and Android? There really isn’t any good use case for another OS, between those 4 you have pretty much every base from embedded to servers covered. Since every use case is covered the ods of getting any real cash flow going is next to nil and without the cash? What you have is another ReactOS, a zombie project that keeps shuffling along without ever reaching a useful state. Remember reactOS started out as FreeWin95, a free clone of Windows 95, and after nineteen years they have yet to release anything that can be used on a day to day system.
like it or not it takes a truly staggering amount of money to make a functional OS on today’s hardware, hell Google spends over a billion a year on android development, and unlike android where they could start from scratch with modern hardware support and application frameworks thanks to Linux with haiku you are trying to build an OS on a foundation that hasn’t had mainstream support since the late 90s. I really don’t see it happening, they just don’t have the resources.
> Even the Linux on the desktop guys seem to have ceased preaching their gospel.
That’s more because we’ve realized that most people are tech-retards that can’t handle a basic desktop OS at ALL. Punting them out of the desktop arena entirely to Linux powered solutions like Android, or Darwin in the guise of iOS, works best at this point, leaving the desktop space for programmers, admins, and power users, most of whom can handle GNU/Linux just fine.
I’ve got a very nice Chromebook that says to me that Linux has a place on the desktop. It’s just that Linux ended up bifurcating: highly-customizable workstations for geeks and hobbyists on one side; single-purpose, idiot-proof appliances on the other. I don’t think it’s a bad legacy.
As to why the Linux-on-Desktop guys are no longer proselytizing: They got something arguably better in the form of something linux-like in the pockets of many millions.
It may not be what they dreamed of in 2000, but ironically it is exactly what Be, Inc. was dreaming about with BeIA.
It’s not one PC and one laptop in every home, but it is a phone and a tablet for every member of every household.
Years ago I got my wife an iPad touch, but it was really for me. It didn’t take much more than a day for me to realize it really wasn’t for me.
We can deride Android as an idiot-proof app bin, but you gotta admit: it feels like a computer when you know how to use it, where iOS feels more like an Internet Appliance.
Canonical, for all of their virtues, still produce a desktop OS that is far from idiot proof, and perhaps that’s a problem for them.
Edited 2014-11-06 08:18 UTC
Users seem to have no problem with the desktop on OSs like Windows or OS X (or Haiku).
I think, in general; Unix is bad for “normal user usability” (due to its power user roots and its “half a century” age); and Linux builds on top of that in several ways (e.g. many incompatible choices rather than standardisation) to create the single largest usability disaster that mankind has ever known. 🙂
– Brendan
I suppose that depends on your viewpoint. As someone who often is called to show how to do something or to repair computers I’d personally say many people do, indeed, have quite a lot of problems even with very mundane, basic tasks. This is on Windows, though, I do not know anyone who uses OSX, I do not know if that scene is any better.
This is called “observer bias”. You only see the people who are having trouble, and don’t see the people that don’t have trouble, so you naturally assume everyone has trouble.
You are right though – people do still need to learn how to use Windows/OS X; it’s just that its less confusing, the user interfaces are more consistent, there’s less conflicting information, and changes (and breakage) are much less frequent.
For a very simple example, if I have a CD containing a Windows application and want you to install that application on my Windows machine, do you think you’d have any trouble? What if it was a Linux application on the CD, and I asked you to install it on my Linux machine?
– Brendan
But do you think those users would be better or worse of on Linux? I would argue they would be FAR worse off on Linux as with Windows there is a GUI tool for just about everything, you have system restore, refresh my PC, and if worse comes to worse factory restore built into most OEM units.
This isn’t even bringing up the fact that the Windows driver subsystem is designed to restart itself in case of failure whereas with Linux you get a non functional device, black screen of death in the case of failed graphics drivers for example, and finally in far too many instances with Linux Bash is used like a crutch instead of designing proper GUIs. Causing so much of your OS to depend on a GUI from the age of disco, one that is NOT in any way, shape, or form intuitive or discoverable, is NOT the way to design a user friendly system!
I would argue the only reason why android has gotten anywhere (which just FYI according to their SCC filings Google has lost over a billion a year on android, the odds of android ever making a dime is about the same as me winning the powerball) is that Google took the OS away from the developers and used it to make a locked down black box OS that is GUI only. I would be hard pressed to even call it a “win” for FOSS as Google is taking it ever more proprietary, look up “Google’s iron grip on android” for an excellent article on how more and more critical systems are being tied up behind the “Google playwall” and how they are cutting support for AOSP in favor of their much more black box Android One initiative.
Linux proper is almost exclusively a server OS (the desktop latest numbers show it BARELY beating “other” which is widely accepted to be Win9X/Win2K) and it has been argued that thanks to VMs and prepackaged images server admin is a dead end job, its gonna go the way of the factory worker, replaced by automation. I do NOT see a happy future for Linux desktops and as Windows and OSX get simpler and more hassle free that won’t be getting better in the years to come.
Linux isn’t really for power users though; it’s for geeks who care more about what license the software comes with rather than how it functions. Power users are the ones using apps like Photoshop, Final Cut, etc. Geeks are fine using some half-assed alternative that’s ‘good enough’.
BeOS fan here, who gave up on haiku years ago
Those guys have discussed the release AND project purpose time and time again, and I’m afraid history will repeat itself. They move on to work on whatever interests them.
Haiku would’ve been great 10 years ago (when it was supposed to be ready) since it was relevant back then. Now, they’re replicating an enormously irrelevant desktop os.
I don’t claim the project is pointless – they’re in it for the thrill of making an hobbyist OS. I just wish they called it what it is – a hobbyist OS that’s utterly uninteresting to Joe the Plumber.
Edited 2014-11-06 08:19 UTC
If all these alternative OS’s would all promote good webapps wouldn’t that solve the compatibility problem? A bank could make a webapp and support any OS with a modern webbrowser. That would make it possible to switch while maintaining the abilities of their current OS. When the time comes that they really like Haiku(or any other alternative) they might want to pay to get native apps that really integrate and feel good in the OS.
Can somebody explain why that isn’t happening?
The problem with your suggestion is that you need a decent modern browser environment.
A modern browser engine, such as WebKit, Gecko, or Blink, consists of many hundreds of thousands of lines of code, if not millions. They are incredibly sophisticated pieces of software, but they don’t work in a vacuum – they rely on having a very sophisticated underlying operating system providing many services.
The fact that Haiku managed to get a WebKit port up and running was a good reflection of the maturity of the Haiku OS. One cannot simply hack together a bare-bones OS and get WebKit/Gecko/Blink to run on it – you need a very sophisticated pretty fully featured OS before any of those engines could even compile, let alone run.
Wouldn’t it be more work to get the most popular app developers to port their work to your platform?
I believe it has.
http://www.xkcd.com/1367/
Ignore this comment if it was a joke. I am still waiting for the promise of webapps. My Android bank app is nicer to use than the webpage because the webpage tries to fill the entire screen with useless junk. If webpages made it possible to use it in a border-less window (of for example 800×600 pixels) that would go a big way towards good webapps. Also being able to integrate into the OS. Where is the systemmonitor webapp? Why do I need to be constantly online for webapps to work?
I hope FirefoxOS can someday show the world how good webapps can be because webos didn’t even try.
Edited 2014-11-06 11:15 UTC
Not a joke, and I’ve had precisely the opposite experience. I’ll elaborate at length, but tl;dr – web apps cover most of my casual use cases, and apparently for many others as well. YMMV, as always, but remember – you’re not everybody else.
For most low-end apps such as banking, my experience has been that the native app is typically *less* capable than the website – and I’m not the only one:
http://xkcd.com/1174/
Not sure what mobile device you use, but the mobile web version of most mainstream sites I use don’t “fill the screen with useless junk” – it’s just a more fully featured version than the limited capability app that they endlessly push on you. And there’s always “Request desktop site” if the mobile site is too limited.
Here’s a recent example from my experience – SmartNews on Android, certainly mainstream at 37,000 reviews and over a million downloads. As a news junkie, it sounded like a can’t miss for me, right?
Nope. Reading a news article in the app is often an exercise in frustration. You can’t “open in new tab” any referenced article, and if you click to read the reference and then tap “back”, you go back to the index, not the preceding article. Comments are hit and miss as well, often not showing up or with the login to comment feature non-functional.
Compare to news.google.com, which works perfectly and consistently on every device I use – Android, Kindle, Chrome, Ubuntu, SUSE, and Win 7.
When my dad died a few months ago, I had to choose between a full-featured desktop publishing or a web app that specializes in funeral-related publications. The latter was certainly less flexible – but I got what I needed quickly and with a lot less overhead than the general purpose “heavy” desktop publishing / graphics design app, and was very pleased with the end result.
Cue the “But Photoshop!” crowd – certainly some native applications are more functional than the competing web apps, and a minority of users will actually need the extra features (and be willing to pay $$$ for them). In a few use cases, I switch to my Ubuntu desktop and use the quad-core, 16 GB RAM, dual-screen goodness to it full potential with a native application.
But for most casual use cases, which for many people would cover the majority of the time, web apps are fine and actually better than native in terms of ubiquity, auto updating, and just getting stuff done.
This to the extent that I’m not at all hesitant to buy an Ubuntu or Jolla smartphone. My experience with Android has convinced me that web apps plus the communication bundle that comes with the OS will cover all I need from a phone.
And I think (to drag back around to the topic at hand) that Haiku will be useful for most casual use cases given solid browser support for the same reason.
The opposite of my experience would be webapps that can run offline, in a small borderless window on the desktop. If you have that I would like to see.
Do you use webapps for mail, calculator, agenda, systemmonitor, volume control, search, podcasting, spotify, gpu control, office, video playback, movies, torrenting etc? I see a huge difference between a website like OSnews, Gmail, Docs Newsblur and a webapp. The Verge comes close but I don’t think it will keep working offline and I still see a lot of Firefox chrome around it.
xkcd can be fun but it is not an argument. And your anecdote has just as much value as my anecdote.
I haven’t seen a news website I could call a webapp.
I tried news.google.com just now. It doesn’t remember what you have read. It doesn’t work offline. It doesn’t pre-load articles. It doesn’t show in a borderless window on my desktop. It felt slow on my HTC One M7. I think a good webapp could solve all this.
That is what I am in favor of. I like webapps. I am trying to promote webapps. You make it seem as if I am trying to stop webapps.
It seems you are trying to argue with me that webapps are a great thing. My argument is that webapps could be a great thing if they worked more like apps so: offline, no browser chrome and pre-loading so you don’t have to wait.
I think if you check the Google Chrome store, you’ll see a large and growing list of off-line web apps (click “Offline” on the filter).
If by “borderless” you mean “without the browser controls”, I think ChromeOS does that about as well as Windows, OS X, and Ubuntu as I recall.
[/q]
I use
** mail – gmail (consider it better than Outlook overall, which is required where I work, mostly because it does search so much better)
** calculator – hp15c.com (I use Free42 in Ubuntu and Android)
** agenda – calendar
** search – Google
** podcasting, spotify – not a big music collector, but I mostly stream direct from source
** office – drive (I use LO for complex documents)
** video playback – mostly YouTube or direct from source
** systemmonitor, volume control, gpu control – not sure why you’d want a web app to control local hardware – that’s what a control panel does (including on Chrome)
My point was that enough geeks consider web apps to be more capable than mobile apps that a very popular web comic was based on it. I.e., it’s not just me.
And if people bought 2.1 billion machines last year that can only run mobile apps and web apps, and only 300 million that can run traditional apps, maybe web apps are sufficient for a lot of other peoples’ needs as well.
But an xkcdreference is always worth the electrons just because! 😉
Commenting? Am I the only person here who remembers when you had to have a Usenet Newsgroup client to conduct a public discussion on the Internet?
re: Google News discussion – agree it’s an online-only web app, but works best for me anyway, as I’m virtually always online.
Not sure why you blame News for your browser having too much chrome, though. Maybe you should switch browsers?
Do I? Didn’t intend to do that, sorry. Just discussing whether web apps are sufficiently capable to make an OS without a lot of native apps sufficiently viable for a majority of use cases for a significant number of people.
Enough ChromeOS apps are there now to make my C720P well worth the money, and I think the industry is moving in that direction. I also think that bodes well for niche OS products like Haiku.
Now whether Haiku has any raison d’être is a different question. But for an old OS geek like me, that’s rather a moot point. 😀
The Google Chrome Store has apps for Google Chrome. That makes them activeX/flash webapps.
ChromeOS has apps for ChromeOS. They only work correctly on ChromeOS.
To be clear: to me a webapp runs in any modern webbrowser, can be run 99% of the time offline and can integrate in the OS so you can’t even tell which app is native. So you see that needing a special browser or a special OS like Chrome and ChromeOS does not meet these requirements.
To you the difference between a website and a webapp is at least one interactive element. Each web2.0 website is suddenly a webapp…
Oh, ouch. Google just passed out stone-cold on the floor. 🙂
ActiveX is a Microsoft proprietary technology. Flash is an Adobe sorta-proprietary technology. Neither is even vaguely related to Chrome.
You can learn more about Chrome-specific web apps at https://developer.chrome.com/apps/about_apps.
However, web apps can and usually are written to be cross-platform relative to OS, browser, and display size.
The client side is typically written with JavaScript, HTML 5, and CSS nowadays, and sometimes with a supporting framework such as Angular JS.
The server side is usually hosted on a framework such as Django or Rails.
You can learn more about modern web apps at http://en.wikipedia.org/wiki/Web_application.
You seem to be a big fan of Firefox, so you may be interested to know that FirefoxOS runs only web apps.
No. Read the articles above and we can discuss further.
I think they know about their closed platform and did it on purpose.
I just went to the Google Chrome Store with a modern browser and tried to install an offline webapp. This is what it said: “You will need Google Chrome to install most apps, extensions and themes.
Download Google Chrome”
I don’t have your exact opinion so I must not have read enough about the subject?
Web apps usually act like web apps. Sometimes they only work in one browser and can’t really be called web apps.
You can make a website or a webapp with those tools.
Is there a reason why you think I learned about web apps for the first time today?
Which is why I mentioned it…
To you the difference between a website and a webapp is at least one interactive element.
How can articles not written by you explain _your_ position? If that is not your position you weren’t clear. Do you have the same trouble understanding my comments? Because it seems we are speaking different languages.
Why haven’t you addressed my main points?
Yes. Because you’re saying things that are factually and fundamentally wrong, yet you keep doggedly defending them. I assumed you were just new to web technology in general, and reading up would get us on the same page.
For example:
The tech that Google uses for their web apps isn’t a debatable opinion, it’s an objective fact. And yes, if you disagree with me and insist Google is using Flash and ActiveX, then I can only assume you haven’t read up on the subject. You may as well be arguing that computers run on water pressure.
Or this:
And you’ve been really fixated on chrome (window controls), which are the province of the underlying OS, not the web app itself.
Without agreeing on basic tech and terminology, it’s pretty hard to debate the more interesting question of whether web apps have reach sufficient maturity to enable a niche OS like Haiku to be used productively.
Never got there, but for the record, I think web apps have roughly reached the maturity of Linux apps when I started using Linux way back in 2000. I had to revert to Windows occasionally for a couple of years until all the gaps were filled, but they were filled – and I think web apps will get there as well.
And that’s goodness, because when the personal OS world was “Windows plus debris”, it was boring! 🙂
Anyway, thanks for the discussion, even if we never got off the launch pad. Have a great day!
Really? Really, you are trying to blame the user’s OS?
Have you check what software the banks use? In the tech support forums there are tons of complaints of the banks still using Windows XP for their internal systems or just changing over because Microsoft has finally ended support.
Last year, I could not get my bank to do a money transfer between my bank account (BoNS) and my friend’s bank (CICB) until I switched from Chrome to IE8.
Just because the OS has all the needed tools, it does not mean the banks will set things up right.
Should have used Bitcoin. (/ducks)
I never understand the statement “we are moving away from traditional desktop OS’es”. Are we? Isn’t Windows and MacOS a traditional desktop OS? If they are, they are really doing just fine. Not even noobs wants every OS to look like iOS or Android. It’s just way too restricting. People wants something that is easy enough and flexible enough. Which is probably why Linux isn’t getting anywhere on the desktop OS market: There are TONS of distros and none of them are really inviting. For normal users it’s better to stick with the big players since they are, the year 2014, not that difficult to use.
I think the question is rather: “do we want to be a big player in the desktop OS world?” If yes, then Haiku (and every other hobby OS out there) has a serious problem. They must not just become stable and usuable, they must also become relevant for people who really doesn’t give a shit about how a system works, as long as it does.
I think Haiku should continue as a Hobby OS. Try to recreate what was good with BeOS and expand on that philosophy. Staying on course will probably raise more interest in the project rather than all the half-assed attempts to adjust to what everyone else is doing.
Then again, no one thought Linux would get anywhere at all at the beginning of the 90’s. But it found a place in the server market. Maybe another alternative OS can find a home once it becomes good enough, Maybe not in every computer users home, but at some at least.
When talking about a move away from “traditional desktop OSes”, “they” are probably referring to e.g.,
http://techcrunch.com/2014/07/06/gartner-device-shipments-break-2-4…
and
http://www.theguardian.com/technology/2014/aug/11/google-chromebook…
With 300 million or so traditional desktop PCs being sold, Windows and Macs are certainly doing ok and won’t become irrelevant or niche anytime soon. But they are kinda dwarfed by the 2.1 billion mobile devices being sold in the same time frame.
The problem is that those mobile devices compliment the desktop, not replace them. This fairly simple fact tends to be ignored by almost everyone, especially people and companies who are making lots of money from mobile devices. There are a small number of desktop users at the low end who have replaced their over-powered, over-complicated (for them) desktop with a mobile device, but those are a small subset of the user base.
“Full” desktop & laptop devices will always exist because they’re too damn useful.
You may be right, but do you have a reference for that?
I suspect that a lot of users in developing countries use a smartphone as their only device, but I’m willing to be convinced otherwise if you’re aware of a relevant study.
If you’re right, then it would appear people are buying 7 mobile devices and 1 desktop each. Weird.
As I said. We’re still waiting for the last mainframe to be turned off. 😀
Just because people who previously did not have access to desktops are now buying low end mobile devices, doesn’t mean mobile devices are “replacing” desktops. They’re complimenting desktops. For desktops to be “dying” you have to show that people are actively throwing out their desktops & laptops in favour of mobile devices, and that isn’t what’s happening.
A properly maintained desktop will last a decade; a poorly kept will last 5 years. OTOH, nobody wants a 5 year old phone.
In the USA, most phones last a least 2 years, because that’s the standard contract length / financing term.
A 10 year old computer would be one the first XP machines, and it’s incredulous that most people are using those today, but I’ll accept your 5 year figure -a nice older Win 7 box.
So we should see 2.5x more smartphones being purchased than desktops. That we’re actually seeing 7x more indicates something more than mobile just complementing the PC.
It’s pretty clear what we’re seeing if you look at the market data objectively – huge population centers like India and China use smartphones instead of PCs, as do a large number of young people today. The PC market is propped up mostly by older folks and by businesses that haven’t yet switched to BYOD.
If you have better data, though, I’m all ears.
…but that doesn’t mean they’ve replaced a PC. They never had one to start with.
Sure. If your point is just that PC sales have leveled off while its former explosive growth has now moved to mobile, I’ll concede that. The interesting innovations are happening in mobile, though, so I doubt we’ll see much growth in PC sales going forward. Guess we’ll see.
Exactly. My biggest proof of that is one very well known experiment in transfering the desktop experience into a tablet one:
Windows 8.
I can understand how they where thinking: everyone is talking about the “death of the desktop OS” and Microsoft wanted to be the first to make the transfer from a desktop OS to a tablet like one, making Apple and Google obsolete at the process.
The problem was that everyone HATED the idea. Microsoft got so much bad press and bad rep because of that that they felt they had to change their OS into a more traditional OS with Windows 8.1. Microsoft rarely adjust their business tacticts based on user critizism.
The problem is rather that development is accelerating, making it very difficult for hobby OS’es and their small teams to catch up. Not just Windows and MacOS, even Linux is part of it. 75% of all code written for the kernel comes from paid developers from big companies. Linux still have the image of being made by nerds on their spare time, making other OS development feel slow. Even when development is proporionally fast and steady.
As long as Haiku becomes stable enough for everyday uses, people are going to use it. Some just like to use wierd stuff because it’s fun, or because they think good design is more important than how many apps are available. For me both is true with MorphOS (which I use everyday).
Hobby OS’es will not take over the world. But they will make a dent.
I see no reason for Haiku to stop. I just wished they’d leverage current technologies more. For example, there’s no need to develop your own kernel when there are perfectly usable kernels available for Free, like Linux or kFreeBSD.
This has been discussed recently:
http://www.osnews.com/story/27907/Haiku_debates_kernel_switch_but_i…
The opinions of the current developers is that it makes more sense to have their own kernel than to use a Linux or BSD kernel.
Of course this opinion may change in the future. I’m not a developer and I don’t have an opinion on the subject. This is just FYI.
It seems silly for Ferarri and McLarren to design a new engine for their F1 cars when there are already engines that exist, like those found in the the Citroen 2CV, VW Beetle and Model T Ford. Why can’t they just use those? (-:
Thank you for proving my point: McLaren is actually using Honda engines on their F1 cars. 🙂
I’m not against Haiku implementing their own kernel if they want to, if they have the manpower to do so. If they don’t have it, well, then there is the NIH syndrome, which can be counterproductive.
BTW, regarding hobbyist project, I think the answer “why not?” is valid for most of the questions.
Nevermind on below. I believe I was a bit slow in detecting sarcasm
—–
If your goal is to design a better engine, why would you outsource the design of the engine..
Edited 2014-11-07 16:44 UTC
If their goal isn’t a hobbyist OS then they really need to figure out what the “essence” of Haiku is and offload the other parts by (for example) using the BSD kernel, or standardizing on the Qt framework. If it’s a hobbyist OS, then no problem, just keep fiddling and playing. If the goal is maturation as a usable desktop OS, then everything cannot remain custom. The hobbyist approach will likely win out because it’s the path of least resistance.
AAARGHHHHHHHHHHHHHHHHHHHH!!!!!!!!!!!!!!!!!!!!!
*Head explodes*
The only problem is every single person who says that then goes and sits on their ass waiting for someone else to do all the work.
And moving to a new kernel is a lot of work.
Notice how all the people who do real development work never make the suggestion? That is because from experience they know how much work that will involve.
Edited 2014-11-06 17:15 UTC
Oh really?
http://www.osnews.com/comments/27907
You just linked to an article about an email thread where none of the people who do the real development work on Haiku make the suggestion to switch kernels, because from experience they know how much work that will involve.
An email thread that now basically appears to have been started by a troll, who promised to show up at BG and give a talk, but was ultimately a no-show.
When pressed for real solutions to problems like how one would magically handle the live query and attribute support of BeFS with a Linux VFS system, the troll who started the thread threw out some vague non-responses about creating some kernel modules, etc. Ultimately, the kernel-level support required to manage BeOS/Haiku filesystem magic would require a large amount of rewriting of the Linux VFS anyway, it seems.
Awww, he never showed up? I was really hoping to hear some good technical discussions.
Anyway, thanks for the update.
I was wondering what became of that, but didn’t see anybody talking about it after BeGeistert ended.
The Reiser4 filesystem had this at one point. They had to take it out when trying to get a Linux kernel merge.
If I remember correctly it was done by giving every file a magic “…” subdirectory. Under that directory were files that described the metadata of the file. An example would be “/etc/passwd/…/size”
I forget how query directories were created.
But the Linux VFS / filesystem people didn’t like it. Being able to “chdir” into things that weren’t directories was a big No.
As I see it, if Haiku migrates on top of Linux/*BSD, they always run the risk as becoming yet another Desktop Environment, so basically they will compete with the established alternatives out there.
If they do need a new kernel, something I am not convinced about, the could probably have some success on top of Darwin or L4, which are not that mainstream as Linux and the BSDs but still have interesting features.
That will always be the case, because no user actually cares about the kernel itself, it’s what’s on top that matters, a GUI or even kernel features that lead to operating system features.
Android uses Linux as a kernel, but no one says Android is yet another desktop environment.
If Android ran on the desktop then it would indeed be just another desktop environment.
How can you say that when Android shares absolutely nothing with other Linux distributions, apart from Linux itself? Applications are different, security model is different, runtime is different, etc..
I still think haiku has a chance, even on their current course.
Unfortunately, they don’t seem to know when to release a milestone Alpha or Beta Release
In the old days, Alpha releases were when you had a new feature that was needed before a beta
so, alpha 1 would be bootable code, making sure it works on the existing equipment of the day
alpha 2 would be a gui
and so on.
once you are feature complete then you move into a beta test period to squash the remaining bugs.
Haiku is never going to have anything other than alpha releases if they don’t make milestones.
IMHO we should have moved to alpha 5 a year ago with the inclusion of package management
Alpha 6 should have been Pulkomandy’s stable Webkit 1.x port last month.
Alpha 7 should be the rest of the Debug, networking kit, and the remains of package management building.
Once those are flushed out, we need to fork Glass Elevator out of the main trunk (which PM should have been a part of IMHO) and put Haiku 1.0 on Beta 1 status.
When developers are suggesting people use nightly builds for their semi production environments, it shows it’s stable enough for another alpha, or to finally go with a beta release.
As a developer, The main question the article asks is which direction Haiku is going. I think they are still trying to make BeOS compatibility, but it has grown so much further than that, nothing I had written for BeOS will compile anymore (USB stack is completely different, Memory space is completely different, much closer to that of windows) , so there will never be compatibility for my apps, just a complete rewrite… that’s not a big deal, it’s just an ever changing OS that I want to see a stabilized code base before I start writing for it.
I highly doubt a complex application such as Gobe 2.0.? would compile, let alone Gobe 3.0 (if we ever got the code for it open sourced).
My vote is to get it to alpha 7, then beta 1, 2, 3 and release 1.0… let me do my work at that time to bring my apps up to date and my USB device drivers up to date, then let’s get busy with GE!
Indeed,the slow pace of milestone releases has made me a little less enthusiastic about testing it and using it day to day on my experimental software laptop. Now that package management is in place I hope they release more often.
Do not let the perfect be the enemy of the good!