Hitting a moving target

1

So I every now and then I get reminded by Haiku, check out the web page and think “This looks like a pretty nice project”. Ever since the demise of Amiga some 10 years ago I’ve been searching for the holy grail of OS:es. Well this time there was even a downloadable Alpha R1 (congrats on that, btw!) so I gave it a try in VirtualBox. And though it did look nice in several ways, I couldn’t help feeling more nostalgic than hopeful for the future.

Now don’t get me wrong here, I’d really want a project like Haiku to succeed so I’m just sharing my opinion :slight_smile:

Obviously Haiku is in alpha stage, but it got me thinking about what you’re aiming for. BeOS was probably great 10 years ago (never tried it because, well I had an Amiga :)) when it was competing against Windows 98, MacOS 9 and AmigaOS 3.5. Let’s say you work hard and manage to release alpha 2 and then the finished R1 late 2010. It will have been a very impressive feat, but the end result is - if I’m not mistaken - a clone of a 15 years old OS.

In the meantime there’s been quite a bit progress in the OS field and hardware has become 10-fold more powerful. I think OSX is a good example of how designers strived to reinvent the way you work with a computer, instead of just polishing and reskinning previous versions. But even Microsoft has begun working harder and their Windows has sucked less with (almost) every release. From what I hear Win7 almost doesn’t suck at all :slight_smile:

So my point is that I think you should set your aim in the future. Try to imagine what BeOS would’ve been in 2012, if it had continued to evolve, instead of what it was in 2002!

Once again, don’t get me wrong, I think it’s awesome that Haiku is efficient and can run on 15-year-old hardware, but for me to use an OS that looks 10 years old it will have to compensate a lot by being really great in all other areas. If your intended target group is anything else than only “people who are willing to write their own OS” I think you need to plan for more advanced visual styles and effects. Not entirely replacing the current looks, but make sure Haiku could “easily” be extended to utilize modern hardware.

But I’m not talking about just visual appearance and effects, how people work with computers have evolved too. I’m going to write another post, that nobody will probably every read, with some suggestions regarding this. I suggest everybody who hasn’t done so already takes a look at the Firefox plugin Ubiquity and use it for a few days. Linux users could also get Gnome-Do for similar functionality on an OS level. Neither of these are perfect or complete, but they give a hint of what powers could be at your fingertips.

2

It will have been a very impressive feat, but the end result is -
if I’m not mistaken - a clone of a 15 years old OS.

You are mistaken. :slight_smile: That Haiku aims at being a retro-OS to be run on old is a common misconception. It is not a coincidence that we don’t call Haiku a BeOS clone.

3

You’re wrong… Nor Vista (7) or MacOSX is light yeeearrs away from us. 3d accelerated GUI demands some time,. and after? Because I Don’t belive that an OS shipped with Webkit, Wi-FI, and latest NVidia compatible Drivers won’t succed

4

You are correct that there is a lot of new hardware available that BeOS does not (yet) support but you should really use BeOS and learn what capabilities it has and why it is still considered an advanced operating system, even 15 years later, before suggesting that a clone of a 15 year old BeOS will be less advanced or relevant than current incarnations of either MacOS X, Windows, or Linux. It seems like a logical deduction if you make the assumption that all operating systems 10 or 15 years ago were equally advanced but you really have to examine what the operating system itself is capable of to determine whether or not it is more or less advanced than new operating systems. There are so many things that make BeOS relevant even today and so many different paradigms it introduced that make working on a computer easier and more pleasant… in some regards Haiku will be more advanced than “The Big Three” even if Haiku was content with being a BeOS clone (which it isn’t, it just wants to have binary compatibility with the last official release of BeOS and then incorporate newer features). I’m not going to lay out each and every capability the BeOS has (and that Haiku will eventually have) since you could look that up yourself and just read it in Wikipedia or you could even download a copy of BeOS PE and run it on an old computer or in VirtualBox. The thing you really have to dig into is what’s happening under the hood and why it is a significantly better design. The way an operating system looks can be changed, that’s window dressing. What the operating system does at the very core is really what’s important and it doesn’t even make any sense to make an OS look nicer unless it’s actually well designed underneath that shiny exterior.

5

15 years old, and … 20 years ahead !

tickless kernel, latency, filesystem capabilities - You can compare BeOS to Microsoft, Linux or Apple.

6

Well I don’t mind being wrong in this particular question :slight_smile: Though I think we’re looking at it from different angles. You see it primarily from a low level point of view, which is quite natural if you’re part of the dev team, while I’m seeing it as a hobbyist interaction designer. I’m sure all the technological bits of Haiku are incredibly well designed. What I was concerned about was the other part where Haiku communicates with the user. And that’s not just a shiny exterior I’m referring to, though I think at least a certain amount of shine is necessary if Haiku is eventually going toe to toe with the other desktop OS:es. What’s even more important is to work with how people use their computer, but I have a separate rant in the suggestion forum about that :stuck_out_tongue:

I see it as a potential risk that you finish Haiku in a few years and then you’ll have 10 years of advances in human computer interaction to catch up with. This might eventually lead to that you have to choose between keeping to bolt on additional layers to a legacy kit, or totally replacing some old stuff which breaks previous apps or at least makes them look like crap compared to the new UI. This has happened several times to the other OS:es. For example KDE3 apps look like crap in a KDE4 environment and are generally incompatible with global settings, but I’m sure the devs had a really good reason for doing so.

It could also be a little prejudice on my side. No offence intended, but I think hardcore programmers (the kind that would write their own OS) often lack the interaction design theory. I wouldn’t mind being proved wrong here, but I think there’s reason Linux is such a mess - from an interaction point of view.

7

You are wrong again: I am not a developer, let alone a hardcore programmer. IMHO, even after all these years, Haiku is still way better from a user interaction POV than other OSes, particuarly Linux on the desktop. But I may be biased… :slight_smile:

8

You told it is awesome that Haiku can run on 15-year old Hardware…
but thats not true… it wont work on my 133Mhz nor my 200Mhz Hardware…
it even doesnt support dial-up connection at the moment…

Haiku is still too slow to run on that old machines… but after some time after optimizations it might run on these old machines…

btw. Beos R5 run on these machines… Beos Max R5 did but very,very slow…

9

If it makes you feel any better, not all of the Haiku programmers are hardcore low-level programmers. Like me for example. I have a lot of interest in interaction design and for example I’ve been studying A LOT of research in the latest web browser design so when I’m writing the new web browser for Haiku I can incorporate the best recent ideas. With no need to worry about previous users of my browser (since it is brand new) I can more easily do things that will take other browsers longer to do. Of course I’m not going to go crazy and my browser will still look and act like a browser, just with some nice tweaks.

I plan to apply the same philosophy to the rest of Haiku in whatever areas I touch. I’m not alone either. We also have Stephan, Stefano, Rene, Alexandre and other developers who focus more on UI development. Even our “hardcore” programmers like Axel and Ingo can work on the UI just as well as the kernel. They both started off as BeOS application programmers and just learned to be kernel developers.

In addition we have a very nice base for interaction design in BeOS, and despite small advances the basics of human computer interaction have not changed in the last ten years. Ubiquity and Enzo which inspired it are pretty neat, but I don’t think they are widely used. And does anyone else consider it ironic that these supposed advances in interaction design are nothing but glorified command-lines? :wink: Nonetheless they are not terribly difficult to implement and BeOS even had at least one app that did something like they do. It would not take long for someone to write something like Enzo or Quicksilver for Haiku.

Either way feel free to test Haiku and make suggestions for improvements in the interaction design if you feel it is lacking.

10

Sounds awesome leavengood! :slight_smile: Is it the NetPositive browser you’re working on? Looking forward to seeing a beta then.

Well actually I think it sounds very logical… The command line is incredibly powerful but it had drawbacks such as it’d difficult to learn all commands and their syntaxes and that it isn’t forgiving against even the smallest errors. So Ubiquity et al tries to keep the power and stuff it in a more user friendly package.

Another approach would be to create a graphical interface and try to improve it to a similar level of power but it’s quite difficult because 1) I’m not sure someone’s figured out how to do yet and 2) A lot of people are very reluctant to change.

But I really like your analogy…

A regular command line is application centric, ie you type the name of an application and then instruct it what you want it to do.

Ubiquity etc is task centric, ie you tell it what to do in a fairly consistent and somewhat natural language syntax.

And graphically, a regular OS is application centric. You start an application and click it’s buttons etc to do stuff.

A task centric OS would kick ass but we aren’t there yet. Maybe Haiku could get there as it appears to have some nice mechanics, like possibilities to combine programs in new ways that weren’t explicitly planned by their developers, data exchange between programs etc. But I don’t think it will happen by itself, just because the mechanics are there. Both Haiku Inc and independent developers are going to have to work towards this goal (that I hope you guys share with me, or will eventually come around to :))

I’m working on a Quicksilver thingy for Haiku btw. At first it’ll only start applications (like Humdinger wants) but with plugins it will hopefully become quite powerful. After that I’ll try to make it a graphical equivalent that you control using only the mouse (Another philosophy of mine is that you ought to be able to control the computer using keyboard only or mouse only as far as possible. Forcing users to swap between typing and steer using the mouse should be avoided)

11

I don’t think haiku is standing still at all. Look at the stack & tile which is now in the trunk. It’s the biggest change to the GUI concept I’ve ever seen in any OS.
http://www.cs.auckland.ac.nz/~lutteroth/videos/stack-and-tile.html

The file system is a pretty big deal. A long time ago Microsoft talked about winFS. That didn’t go anywhere after all these years.

OOP is another concept that changes the software relationship. In other OS’s you get more lockin because every single app has to reinvent the wheel translating formats. Haiku has datatypes/translators so app designers can work on more interesting things and users won’t have to worry if their favorite word processor can handle the formats they need.

12

I think the op may have missed the point entirely, With the first official alpha release Haiku has achieved what no other OS can boast today, and that is a robust intelligently designed foundation. With the calibre of developers that can create a fully threaded operating system, I don’t see how ‘keeping up’ can be a problem. The foundation is already ahead of everything else :slight_smile: Now it’s just a matter of adding filler: office suite, one-stop-multimedia-shop, accelerated graphics, UI touch-up etc…

but as I said, I have the utmost confidence in the developers and look forward to a 1.0 release.

13

[quote=Reyenweald]I think the op may have missed the point entirely, With the first official alpha release Haiku has achieved what no other OS can boast today, and that is a robust intelligently designed foundation. With the calibre of developers that can create a fully threaded operating system, I don’t see how ‘keeping up’ can be a problem. The foundation is already ahead of everything else :slight_smile: Now it’s just a matter of adding filler: office suite, one-stop-multimedia-shop, accelerated graphics, UI touch-up etc…

but as I said, I have the utmost confidence in the developers and look forward to a 1.0 release.[/quote]

The difference is that you see the OS as a developer and talk about low level stuff, while I was looking at it from a user’s point of view. A user sees things that are on the screen, not the foundation :slight_smile:

My point is that when the awesome core of Haiku is finished, they’ll have to catch up with the other OS:es user iterface-wise (and I wasn’t talking just about eye-candy) and if they look at the current OS:es and take the best parts it will still be old when it’s finished. You have to aim into the future to have a chance. Revising the user interface usually isn’t something done ad lib so I was wondering if Haiku was prepared for the inevitable changes ahead.

Stack and tile looks like a nice feature when you’re working with several windows but I wanted you to think the bigger questions like “when I’m using an OS in the future, will I even have windows in the traditional sense?”.

30 years ago someone might have set out to write the best text based OS there ever was, but would they still have a chance when everyone else moved on to graphical user interfaces? (Small niche markets excluded :))

14

For me one of the biggest WOW!-moments in recent couple of years was when I discovered that in Google Chrome You can drag tabs out of the browser. That was a BIG thing. And that’s kind of pathetic, if You think about it… (Ok, Compiz and Beryl were also BIG WOW, but they were further back in time… :wink: Accidentally BeOS had tabs integrated in OS over 10 years ago. Add small(ish) update to this basic tabbing (called stack&tile) and You get the same functionality that took all this time for other OS-s to develop (and even that in only one application!). Point being - good foundation counts! And also - You don’t have to invent everything Yourself.

15

@fhein

Text-based OSes never went away. Setting out to build the best command line OS wouldn’t have been futile in the least, if we’d have ended up with something better than the command line tools people are still using on Linux, Unix (incl. OS X) every day. People (Microsoft included) are working on this now, in the 21st century.

The spread of GUIs was facilitated by hardware development. While we still have 2D monitors, we’ll still have GUIs that are refinements of what’s gone before. There have been endless proof-of-concept interfaces that don’t work as well as the current desktops. Unless you have anything concrete to back up your belief in blue-sky thinking, then you shouldn’t expect a brilliant new paradigm to emerge here more than anywhere else.

Also, the computing revolution is happening on the web. Nobody knows how many users will end up using a web-client OS like ChromeOS. The web is in the hands of a million web-sites rather than any OS project. Still, some users will always want their desktop, and Haiku fulfills that need. If the developers didn’t fit into that category, they wouldn’t be working on Haiku.

For those reasons, I think Haiku should carry on seeking to be the very best desktop OS that has ever existed. They’re doing pretty well in that aim so far, and I personally find the whole thing very exciting.

16

With the first official alpha release Haiku has achieved what no other OS can boast today, and that is a robust intelligently designed foundation. With the calibre of developers that can create a fully threaded operating system, I don’t see how ‘keeping up’ can be a problem. The foundation is already ahead of everything else :slight_smile: <<

With all due respect to the Haiku developers, I don’t think the claims above can be taken seriously, nor can the rest of the “Haiku is tougher than Chuck Norris” comments on these forums. They are useful though, in that I’m starting to understand how I must have sounded to my Windows-using friends when I became a GNU/Linux evangelist, and it could no no wrong :wink: I still use GNU/Linux in preference to Windows, as much for political reasons as technical ones, and I’m willing to admit that it isn’t perfect… yet :slight_smile:

Stack and Tile is a nice feature, but it doesn’t hold a candle to Unity or Enlightenment desktops on GNU/Linux that can switch between multiple desktops (how many? you decide!), with a simple key combo. It’s not really much flasher than what Windows 7 can do; drag the Window to the screen edge, it resizes to take up half the screen, do that with the other window, and you’ve solved the problem described in that video just as quickly.

Now it’s just a matter of adding filler: office suite, one-stop-multimedia-shop, accelerated graphics, UI touch-up etc… <<

What your saying is something like “we’ve got the wheels, now it’s just a matter of adding filler and we’ll have a car”. An office suite is not “filler”, it’s a highly complex suite of applications, with many thousands of lines of code. So is a media player. Without key user applications like these an OS isn’t useful to anyone but hobbyists, regardless of how bleeding edge the foundations might be.

As it happens, RC1 already has a working media player, a web browser, and a number of other user applications, so I’m not sure what OS Reyenweald has been running. RC1 is a good start, but fhein is right, there’s still a lot of work to do before it’s ready to compete with any of the current OS. Keep up the good work devs!

17

You know, there are some things missing in Haiku, but the OP did miss an important point. Haiku is not targeted at older machines, but at replicating the BeOS interface and its advanced features, like the Be filing system. Alpha 4 is already an improvement on BeOS in many ways; it runs great on my newer computer–it’s fast and stable, and while it can’t quite do everything, it does a heck of a lot of useful stuff for me. It installs fast, it boots fast, it runs fast–faster than any *Nix or Windows could possibly run or boot. And BeOS/Haiku is easy to use.

Oh, and btw, BeOS and Haiku both have multiple desktops, too. Up to 32 workspaces, if you want. And each one can be set up with its own background, color depth, and screen resolution–does Unity or Enlightenment do that?

18

Okay, just to be fair, Enlightenment does have this:

Enlightenment allows you to have a grid of workspaces called virtual desktops. Switching between them is achieved by hurling the mouse cursor to the edge of the screen, at which the desktop appears to slide across to reveal the next. The maximum grid size is currently 8 by 8 desktops, and you can have 32 grids (each with a different background), making 2048 total possible desktop spaces. (Users can enable a map of the desktops, in case they get lost, which is called the pager.)

But my point still stands: BeOS/Haiku does have multiple desktops, just not 2048 of 'em!

19

You don’t need 200+ virtual desktops / workspaces and if you feel you do, then I suggest going back to Enlightenment or every hip and cool desktop environment, that Linux has. Haiku is what it is and the current desktop experience isn’t going anywhere, at least not until R2 happens.

Haiku is simpler, better foundation for Multimedia Desktop OS than GNU/Linux without loads of legacy stuff (save for ridiculous decision of keeping binary compatibility with BeOS GCC2 apps). With some serious love, care and attention Haiku can end up successful as the main open source Desktop OS you would wanna use, especially on older or less powerful computers. Smaller, quieter HaikuBox, that can handle all kinds of multimedia tasks is the future. Tablets are generally made for content consumers only and some people actually wanna use computers to create things like graphic art and music. If you want to see good example of that “small multimedia box” concept, just look at modern Mac Mini. I just think with enough hardware support (like GPGPU, OpenCL acceleration) Haiku should do better than OS X and run more apps at the same time on machines like that or even less fancy.

20

I wouldn’t try to tell someone how many desktops/workspaces they need, but if you have so many that you get lost, then I would suggest that you have too many!

:wink: