Steady Progress towards Alpha 1

News posted by stippi on Sun, 2008-05-18 11:27

This weekend, the Haiku project has seen some nice leaps forward. Two items are especially noteworthy: Ingo Weinhold and Axel Dörfler have finally nailed bug 2059. This bug specifically prevented serious use of Haiku for anything else than testing, since it meant that the kernel could crash at any time, especially when there was heavy disk activity. All that was supposed to be written to disk at the time of the crash was lost. Luckily, due to prior fixes to the file system journaling and log replay, it didn't mean that your entire file system would be corrupted, but at least anything that you were working on at the time would be lost. So the fix for this particular bug is getting us much closer to our goal of a usable self hosting situation in which you can actually use Haiku for development. This is our most important goal to reach before we wanted to release the first alpha of R1.

The second noteworthy achievement is build system support for a mixed GCC4/GCC2 Haiku environment. It has been known for quite some time, thanks to the explorations of Haiku developer Michael Lotz, that it was possible to set up a GCC4 build of Haiku to run GCC2 applications or vice versa, by installing the respective libraries into certain places so that the correct versions are used for linking. What was missing was support in the runtime loader (the system component used to launch applications and link them to the shared libraries that they use) to do this automatically and on a system wide level. Also missing was support in the build system to effortlessly produce such a hybrid Haiku build. Both of these items have now been implemented by Haiku developer Ingo Weinhold. Also related to this, Michael Lotz had researched the stability issues that GCC4 builds of Haiku were suffering from some time ago and tracked them down to a problem in the specific GCC4 version that Haiku is using. They can be avoided by simply turning off a certain compiler optimization feature. All this combined means Haiku can use GCC4 itself while maintaining our stated goal of binary compatibility to the large pool of GCC2 applications in an automated and transparent fashion.

A few serious issues remain before we can release the first alpha. Some concern missing or buggy functionality that affect the self hosting goal with regards to the development tool chain. A completely native port of Subversion is the last item on this list. As far as I know, some bugs in the TCP implementation are preventing it, but progress is being made on this front as you read this. Formal testing is being conducted to make sure the entire tool chain will work correctly and reliably. Axel Dörfler is currently working on the device manager, the system component which manages everything concerning hardware and drivers. There are some issues with regards to hardware interrupts that, when fixed, will hopefully clear up some driver problems that can be experienced on certain hardware.

I want to conclude with a big "Thank You" to everyone who is helping with tracking and reporting issues in our bug tracker and to everyone providing patches and of course to the Haiku developers themselves! Personally, I am very excited about the progress that is being made. Thanks to everyone who is contributing towards this goal!

Comments

Re: Steady Progress towards Alpha 1

The link don't working! Please, fix it :)
Thank You, Thank You for this big work!

Hungarian translated version: HERE.

Bye!
miqlas

Wonderfull

I just donated for the Haiku Code Drive 2008 but I'd just do it al over again just to thank Ingo and Axel.

Guys, I think it is time to start writing your wish list for a very early Christmas! :-)

Re: Steady Progress towards Alpha 1

Thanks a lot Stephan for this status update!

Re: Steady Progress towards Alpha 1

That's excellent news Stippi. I'm especially thrilled at the GCC4 news - it's very cool that Haiku can be built with GCC4, I was seriously concerned using GCC2 as the primary compiler would add weight to the misplaced impression that Haiku is nothing more than an open-source copy of a 9 year old system.

I'm going to get to a quick catch-up update of the last few months in the next couple of days, after which I'll be sending my Haiku emails to an IMAP account so I can keep track of updates from both home and work and hopefully not get caught in these annoying backlogs again.

Simon

Re: Steady Progress towards Alpha 1

Great news, and I hope to be able to donate shortly!

Re: Steady Progress towards Alpha 1

Simon Taylor wrote:

I'm going to get to a quick catch-up update of the last few months in the next couple of days, after which I'll be sending my Haiku emails to an IMAP account so I can keep track of updates from both home and work and hopefully not get caught in these annoying backlogs again.

Simon

You do not by any chance intend to continue your commit log summaries? If so, let's please combine efforts (with Niels and me).

Re: Steady Progress towards Alpha 1

that mean, what now we can have a haiku 64bits version? or what we don't have yet? cause is a reality, 64bit is now.

Re: Steady Progress towards Alpha 1

So will R1 be compiled by default with GCC4 instead now?

Re: Steady Progress towards Alpha 1

No, skarmiglione, it's not that easy to switch to 64bit.
It means that you can use the current compiler version gcc4 to develop the OS and applications, but are still able to use software that was compiled with the old gcc2 compiler. And vice-versa.
So it's binary compatible to BeOS software and at the same time ready for things like the WebKit that demands the features gcc4 offers. All completely transparent. Very nice!

Re: Steady Progress towards Alpha 1

Say does all code have to be manually updated to support 64 bit or could a good regex do the trick? (this was actually meant to be answered by a C,C++ guru)

I'm probably getting it wrong thinking memory gets allocated by programs themselves, intersting nonetheless.

Cheers

Edit: http://en.wikipedia.org/wiki/64_bit I guess it is more work then I thought, device drivers being the most problematic IMHO though

Re: Steady Progress towards Alpha 1

That is terrific news. First, great to hear that major bug 2059 was squashed.

But, really exciting is that Haiku will be GCC4 based + GCC2 compatibility libraries. This is a super move on your part because R1 is probably 3+ years out and by that time we might have GCC5 ( or GCC4.x where x > 6 ). At least with GCC4 we'll see more developers contributing or porting over programs & games. So the OS can really start to attract & gain users right when 1.0 is released ( rather than 2.0 ). Very Well Done Guys.

@ J.Freeman
From my understanding, they will compile Haiku R1 with GCC4 ( system & libraries ) + GCC2 libraries since this makes more sense and brings Haiku up to date. It looks like what they're going for. Answer = yes. This is happening because they are able to create a mixed / hybrid system ( one that works with GCC2 + GCC4 programs & games ).

GCC4 will make it easier to port over software from Linux/BSD.

Re: Steady Progress towards Alpha 1

64bit, Answer = no

See here for further detail:
http://haikunews.org/1296

Many don't realize what 64bit means. The difference with 64 vs 32bit is that you get to use RAM above 3.3GB barrier & you may see a 5-8% "overall" speed improvement. But, with 32bit you support all x86 computers and with 64bit you only support the newer systems. Most users have 1 to 2 GB of RAM installed so no need for 64bit until they go 3.5GB+. And many still own or use a 32bit computer. Only argument to make then is that you gain 5-8% performance and that is a really "weak" benefit.

I'd rather have better SMP support so that I'd get 50-80% performance on CPU intensive programs rather than 64bit. Even SMP doesn't matter much because there are "few" CPU hungry applications.

I still have 3 32bit systems I use. And many others probably still have 32bit systems that they'll use for some years still.

Actually, it'd be cool if R1 had Hardware Accelerated OpenGL ( for newer Nvidia, ATI or Intel cards ). So that we can play & demonstrate 3d games. I believe that would be more impressive to people.

Re: Steady Progress towards Alpha 1

tonestone57 wrote:

This is a super move on your part because R1 is probably 3+ years out...

Oh really? You have a secret Haiku roadmap that nobody knows about? Or maybe a crystal ball?

Re: Steady Progress towards Alpha 1

then 640k...now 4 gigs enough for everyone? I think people will want many gigabytes of memory for video and games. You can have the most compact OS in the world but the video files will still take up space.

Re: Steady Progress towards Alpha 1

"Most users have 1 to 2 GB of RAM installed"

Actually, many new systems have 3 or 4 GB.

Re: Steady Progress towards Alpha 1

Jack Burton wrote:

"Most users have 1 to 2 GB of RAM installed"

Actually, many new systems have 3 or 4 GB.

Well, one statement doesn't preclude the other. :)
Plus, 3 or 4 GB is still in the 32bit range.

Anyway, there are bigger fish to fry, namely Alpha1 and R1. When those are out, the situation should be re-evaluated and a decision made. Maybe we'll have a Code Drive for a 64bit Summer Camp when we ship our kernel hackers to an island where they code until it's done. It's a cold and rainy island of course, so forget the beach, guys! :)

Regards
Humdinger

--
--=-=--=-=--=-=--=-=--=-=--=-=--=-=--=-=--=-=--=-=--=--
Deutsche Haiku News @ http://haiku-gazette.blogspot.com

Re: Steady Progress towards Alpha 1

koki wrote:
tonestone57 wrote:

This is a super move on your part because R1 is probably 3+ years out...

Oh really? You have a secret Haiku roadmap that nobody knows about? Or maybe a crystal ball?

Ok, I'll explain. If you look you'll see I said "is probably ...". Meaning there is a good chance ( using probability ) that it'll take another 3 or so years for Haiku final. It could happen sooner but only if more developers jump on board. Take a look how long it has taken to reach Alpha level which is still not out. Yes, this is more a rough estimate than an actual release date and may be wrong. As of today, this is what it looks like to me. In 6 months or 1 year from now it could change.

Using "is probably" does not mean R1 couldn't come out in less than 3 years but this is less likely of occurring. Haiku will have to go through early Alpha ( almost there ), late Alpha, early Beta, late Beta, Release Candidate & then Final. If all of this can be done in less than 3 years then I'll be jumping for joy. Yes it may even happen but only if more programmers start getting involved.

PS I, and I'm sure others, prefer to have a rough, estimated release date that is revised every so often then not having one at all.

Re: Steady Progress towards Alpha 1

Is there a way to access the extra ram in a system (above 4gb) like they used to do with ram stackers and similar to partitioning a large drive on an older motherboard?

Re: Steady Progress towards Alpha 1

Jack Burton wrote:

"Most users have 1 to 2 GB of RAM installed"

Actually, many new systems have 3 or 4 GB.

Those that have upgraded their computers 1+ years ago will be running with 1 to 2GB. DDR2 RAM has become very inexpensive lately.

Those that are buying new systems in the last 12 months will have a system with 2 to 3GB. Most of the new systems being sold today are coming with 3GB RAM - I just checked. With high end gaming machines having 4GB & starter / basic systems coming with 2GB.

Still, the point is this. Even if you have 4GB installed. Do you believe that Haiku will use up all of that 4GB? Or maybe people use applications that require 4GB? Most of the memory will go unused. I can see people using up 1 ( or 2GB ) on Haiku, but using 4GB then that would be really something. Yes, you could do it, but only if you opened up lots & lots & lots of programs at once and had huge caches and/or ram disks.

PS On Windows Vista you need 2GB to run real good - the only OS today that needs lots of RAM. Many other OSes will run really well with 1GB. Plus the newer, high end 3D Windows games may need another 1GB+ on top of the 2GB for Vista to run great. So, for Vista 3D gamers 64bit may be the way to go but for other OSes, installing 3 or 4GB will be a complete waste as memory will go unused. So, whether Haiku was 32 or 64bit, you'd still not use the 3 or 4GB you have on the system.

And Haiku won't be able to play any real 3D games without hardware accelerated OpenGL. Which is actually more important than going 64bit.

Re: Steady Progress towards Alpha 1

Well said. That's really the point: 1) It isn't very useful right now. 2) There are more important things we need to focus on.

Re: Steady Progress towards Alpha 1

mikedoth wrote:

Is there a way to access the extra ram in a system (above 4gb) like they used to do with ram stackers and similar to partitioning a large drive on an older motherboard?

In Haiku? No. Haiku is limited to the same as regular home Windows XP systems, about 3GB of RAM.

Otherwise yes, so long as the motherboard hardware (mainboard chipset, memory controller if not integrated into the CPU) supports more physical RAM (check your manual) there are essentially two routes...

The obvious route is 64-bit CPU and 64-bit operating system, typically with say 48-bit virtual addresses and 40 bits of physical addresses - which allows you to use the RAM quite naturally, it's available both to the operating system itself and to application developers. Microsoft, Apple and various Linux distros and major Free Unixen all offer 64-bit operating systems. You can buy AV workstations with 16GB of RAM easily right now.

On older CPUs (say a Pentium III) the solution is wider physical addresses mapped via the MMU. Intel calls this feature PAE on their CPUs. Microsoft and the major Linux distributors support this option too. It's more limited in that individual applications don't get an increased address space, but you can f.e. have 3GB of RAM used by a photo editing program, plus 3GB used by a vector graphics package and still have 2GB of OS disk cache with 8GB of physical RAM this way. Usually it's not a good idea to use PAE for more than 16GB of RAM although in theory you can have up to 64GB.

Re: Steady Progress towards Alpha 1

tonestone57 wrote:

And Haiku won't be able to play any real 3D games without hardware accelerated OpenGL. Which is actually more important than going 64bit.

Amen

Re: Steady Progress towards Alpha 1

there are a lot of prereqs for games. I just think there should be a higher priority for looking towards the future than making sure haiku supports all the old hardware out there.

Re: Steady Progress towards Alpha 1

I can't but to agree. My older laptop can run Linux until it dies for all I care. I mainly want to use Haiku to replace my desktop OS so I can game, graphics and audio editing. This means using it on my new desktop machine and newer laptop. If need be let other distro's deal with other architectures or older hardware.

Re: Steady Progress towards Alpha 1

uuuhmm doesn't 32bit linux support 64gb of ram? and say you have povray running on one cpu (making use of haiku's threaded nature) and you are rendering a huge image.... that takes a lot of ram

video editing can be ram intensive too...and wow already at $6599 :-)

Re: Steady Progress towards Alpha 1

i have a 64 bit processor, i supposse like so much people, and i understand more people have a 32 bit proc, but the open source don`t leave anybody out of, and less now what we have this, gcc2 and 4 retrobincompatibility for the 64 bit in my pc on haiku, is good; no more words.

skarcat:
sorry for the english

Re: Steady Progress towards Alpha 1

64-BIT BENEFITS ONLY THOSE WITH:
1) 4GB+ RAM ( & 64-bit CPU )
AND WHO DO ONE OF THE FOLLOWING:
1) Video Editing ( large files ), large databases, scientific computing and/or High End 3D gaming on Vista ( because Vista OS takes lots of RAM too ), ( & maybe complex 3D image editing? ).

Only a very small number of people, out of all the Haiku users, will do one, or more, of these. And for those, having access to 3GB for the program should still satisfy most of them. Just because you have a 64-bit CPU does not mean you'd benefit from using a 64-bit OS. If you have lots of RAM & are not doing one of the things listed above then part of your RAM will be sitting idle - whether you have 2, 3, or 4GB+.

Athlon64 came out in Sept. 2003, 64-bit Intel CPU in 2004 & Intel laptops switched to 64-bit in 2006. 64-bit has not been around for very long - 4 1/2 years from AMD & 4 years on Intel side. 32-bit has been around for many, many years and there are plenty more 32-bit computer users out there than 64-bit. 64-bit computers may not outnumber 32-bit computers for another 5 ( or more ) years. 32-bit computers are the majority for now. Doing a 32-bit OS supports both 32 & 64-bit systems.

Most of the programs you find are 32-bit ONLY. So, you'd have a 64-bit OS to run 32-bit programs? What for? It's better to run 32-bit software on a 32-bit OS, less likely to have compatibility issues. Why aren't many developers also making 64 bit versions of software? Because it'll use up their resources to support both 32 & 64-bit and only certain software actually benefits from a 64-bit version. Only programs that gobble up RAM like crazy. The general ones I listed above.

Haiku uses a small amount of RAM for the OS ( 128MB or less ). Not like Vista that will take 1 ( to 2GB ) of RAM. So, on a 4GB Haiku system, you'd still have just over 3GB of RAM for the program or game to run with. For almost everyone, this should be more than enough. Screen shot of system with 4GB RAM running Haiku:
http://www.biffuz.it/misc/haiku4gb.jpg

So that a very small number of users would actually benefit - Going to 64-bit OS means Haiku would have to support both 32-bit & 64-bit and create a 64-bit version too. This means they'd have to: 1) rewrite the drivers to 64-bit, 2) redo the OS to 64-bit, 3) create a 32-bit compatibility layer to run all the 32-bit software & make sure that GCC2 & GCC4 32-bit software works properly. All of these things would create new bugs & issues that would have to be fixed too. Creating & supporting a 64-bit OS *could* add another 2 to 3+ years to Haiku release to give a benefit to the very few - this is called wasting resources to make a few users happy. They would have to maintain 32 & 64-bit versions which would add lots more work for the few developers there are.

Mainstream OSes - Linux, BSD & Windows are able to do 64-bit because they have lots & lots of programmers. And 64-bit OS benefits servers - which these OSes are geared towards too. Plus, it benefits those users I mentioned at the start of this post - most of who you'll find running either Windows or Linux. The more developers & mainstream Haiku gets after R1 release, the better chance you'll see a 64-bit version with R2 or R3.

From a logical standpoint, pushing for 64-bit OS makes no sense as I outlined above. I believe most are pushing for it because 64-bit sounds hip or cool today ( doesn't sound outdated like 32-bit which has been around for long time ) and also for those users that may or will install 4GB+ RAM who feel cheated to see only 3.3GB available to them to use ( which is more than plenty for almost everyone ). Adding more & more RAM to systems gives benefit up to a point, after which, it'll give no benefit. Just adding 8GB of RAM to a system doesn't mean it'll run faster/better unless your programs actually make use of the RAM! As for performance, 64-bit may give a slight boost but I've seen in some cases where 32-bit was faster. Still, it isn't a strong enough justification to create a 64-bit version too.

I hope this puts 64-bit to rest because if this doesn't convince some users then nothing ever will.

Re: Steady Progress towards Alpha 1

Back on topic... thanks stephan for your update, I really appreciate it.

Re: Steady Progress towards Alpha 1

tonestone57 I do not understand about what with all suddenly also it leads from these 64 bits too. Who have Intel Itanium on board? Because majority then arrangements x86 / 64, supporting both architectures. Problem establishes these 4GB? O my...
Some look for problems not there it where was should.

BTW: If we tell already about hardware 3D acceleration, maybe we will make sensible bounty?

http://www.haikuware.com/forum/bounty-discussion/307-hardware-3d-acceler...

PS. Sorry for english, I'm not native speaker.

Re: Steady Progress towards Alpha 1

AMD is to blame! They started the hype with their Athlon 64 & many people bought into it. Intel had no choice but to follow suit because people were going for AMD's 64-bit CPUs. Also, AMD/Intel got people believing that 32-bit will hold you back - maybe to sell the newer CPUs. 64-bit is the future but that isn't for another few years ( 5+ years? ). When you'll see a bigger shift to 64-bit. Today, only a select few users really gain from 64-bit. The developers are smart guys and will decide for themselves when it is a good time to go with 64-bit Haiku. It won't happen for R1. Maybe for R2? Or R3? Time will tell.

I actually did a post earlier today about Hardware OpenGL. Looks like it is intended for R2. I don't believe a bounty will get it done any faster for R1. Most developers are working on other parts of the OS to make Haiku stable & complete it faster. If a bounty would help get Hardware OpenGL made for R1; then I would pitch in but first I would need reassurance that a developer would work on it and do their best to get something out, with acceptable performance, for R1 before I donate to this bounty. And support a video card that I could easily buy as either AGP or PCI-E. Either Nvidia or ATI.

http://www.haiku-os.org/community/forum/hardware_opengl

Re: Steady Progress towards Alpha 1

I like the bounty idea for 3d acceleration.

Re: Steady Progress towards Alpha 1

but isn't the whole point of haiku to be the "media" os or the OS that can handle very large tasks. Extreme multitasking with multiple cpu cores.

You have to think about those that will want to switch in the future. We're talking about 5 gigs. That's not so far away from what we have now and when r1 is released, 5 gigs will be old news.

Re: Steady Progress towards Alpha 1

arielb wrote:

but isn't the whole point of haiku to be the "media" os or the OS that can handle very large tasks. Extreme multitasking with multiple cpu cores.

Haiku's claim isn't about media, the site says it is "designed from the ground up for desktop computing" although of course in reality Haiku was mostly not designed at all, it's a clone of BeOS R5. The "desktop" claim is basically just a way to excuse swathes of missing features, such as in this case lack of support for modern CPUs.

Any "media" claims absolutely require broad application support which Haiku doesn't have. The situation is that all the obvious first choices have this broad application support already without making overblown claims. You can already DJ with them, compose or arrange an orchestral piece, retouch photos, record and edit your band's demo, produce 3D animations and so on. And yes, more than a few of those things benefit from having plenty of RAM.

Quote:

You have to think about those that will want to switch in the future. We're talking about 5 gigs. That's not so far away from what we have now and when r1 is released, 5 gigs will be old news.

Realistically it's about 2-3GB. That's the region where all things being equal the argument for 64-bit is made. By the time you have 4GB of RAM installed the only reason to be running a 32-bit OS is because you're stuck with it. With 2-3GB of RAM you have to more or less give up split view memory mapping in the OS kernel, which costs you performance. Your application programmers are cramped by this point too, large contiguous allocations become hard to arrange and use of memory mapped files can exhaust the address space altogether. So you are paying a price to stay 32-bit by that point, and for a lot of users that point is today.

When you're running a modern x86-64 laptop or desktop in its 32-bit compatibility mode you're also wasting half of the general purpose registers and half the floating point registers it actually has, along with other features present only for 64-bit code.

Re: Steady Progress towards Alpha 1

All valid points. But which part of "we have more important things to do right now" don't you understand? There have been plenty of arguments given here on why it is not the most important thing out of the list of missing stuff which is surely not short. So why don't you let us focus on stuff that is higher up, instead of expressing your discouragement here once again?

And you know, I do kind of take offense with regards to your "Haiku was mostly not designed at all" remark. Just because Haiku is binary compatible to BeOS and we are implementing an existing API does not mean we didn't have to design stuff, you know? The bulk of the code is the design of the respective developers. Take drivers for example, take stuff like the app_server, many things. With the same ignorant approach, you could say Windows was not "designed", it's just the same ideas from other OSs - a kernel, filesystems, applications, DLLs... nothing new.

So easy to write a discouraging comment on some website. I would love to learn what stuff you accomplished in your life. Did you ever build something and try to realize one of your dreams? I would be quite surprised if you ever did, because I would assume someone who did wouldn't diss other peoples stuff so lightly.

Re: Steady Progress towards Alpha 1

Stephan, you guys are doing an excellent job. Do not let negative comments get to you. You'll get these from hardcore Linux supporters - who tend to be Haiku bashers. They believe Linux is perfect and are against having another open source OS. They believe Linux is the BE ALL END ALL FOR EVERYONE - best OS out there. Since that is the case for them, they should just go play with Linux & leave the Haikuers alone.

Maybe they are afraid that Haiku will take away ( steal ) many users and developers from Linux. Maybe they don't want the competition or afraid that Haiku may, at one point in the future, overtake Linux popularity? Or believe they have to convince Haiku users that Linux is the OS that everyone should use?

Do they realize it took Linux many, many years with lots & lots of developers helping out to get to where it is now? Plus not to mention how corporations got involved and helped things along. They like to compare Linux, a mature product, to Haiku which hasn't even reached 1.0 yet and is more of a hobby OS at this time. I really feel that Haiku will become mainstream with R2 or R3 release. And these hardcore Linux guys realize this too and why they try to bring down Haiku every chance they get.

My opinion, Linux is a Good OS but BeOS was an Excellent OS & Haiku will be too! Let's see how Linux stands up against Haiku R2 or R3 release. Should be interesting. Once again, fantastic job and keep up the good work. Don't let the Linux fanboys get to you Stephan. Regards.

Re: Steady Progress towards Alpha 1

@arielb
Just so you understand. The main reason they're not doing 64-bit is because they don't have enough developers and it'd take considerable time & work for them to do it. My personal stance is that it'll benefit very few users anyways. That said.

Haiku has multi-threaded design ( this will give responsive multi-tasking ) and has SMP support so will take advantage of multiple core CPUs. These are separate and do not depend on 64-bit.

Just a little article you should read. They should have done benchmarks with 3, 4GB on 32-bit vs 3, 4 & 8GB on 64-bit.
http://www.tomshardware.com/reviews/vista-workshop,1775-5.html

The points made:
1) 64-bit increases memory footprint & should be used for systems with over 4GB. ie: 5GB+
2) 32-bit applications will run slightly slower on a 64-bit OS and not able to access over 4GBs of RAM
3) Might have driver issues
4) Tom's Conclusion "Our testing brought us to a clear conclusion: if you often use several memory hungry applications simultaneously, then there's really no way around upgrading your system to 8 GB."

Ok, the key point. "If You Often Use Several Memory Hungry Applications Simultaneously" For instance, running 5+ Virtual Machines at once or playing newer, top end Windows 3D games, etc.

What do the vast majority of users do with their computers anyways? Email, Web Browse, Office Applications, Chat, Bittorrent, read & write to newsgroups or forums, play 2D games, play audio or video files, view photos/images, etc., etc. Nothing memory intense at all. If this is all you do, then why go to 5GB+? I can do all this stuff with a Pentium 3 w/512MB RAM. Whether 5 or 10+ years from today, I still won't need a system with 4GB+ to do this stuff.

The Exception is the younger crowd which may want to play the latest 3D games, but these usually are available for Windows anyways - so they should get 64-bit Windows. Also, super power users will need it too but for the majority of people they don't need lots & lots of RAM.

With 1GB on XP, I would run 2 to 3 Virtual Machines of BeOS, Haiku or Linux, yes, I would have been swapping to disk but still ran good. Does someone really need to run 5+ Virtual Machines, of Vista, or 10+ of BeOS/Haiku all at once? I would do audio & video conversions, run Photoshop ( create 400dpi images ), compile programs, play 3D games ( back in 1997 to 2003 ), Emulation, plus the stuff I mentioned above. It's amazing how I was able to do all this stuff with 1GB RAM on 32-bit OS. I can only imagine what I could have done with 3GB of RAM - probably opened up and ran 3 times the programs at once. If you can only work with 10 programs at one time, why open up 100? Just to say you can? How did people ever manage with 32-bit OS before 64-bit came out?

I'm certain, if we did a poll today, that most users have 3GB or less installed in their systems. And in 5 years from now we probably will have a similar result because I believe 3GB is way more than enough for the average & power users and even for a few *super* power users too, except for those playing latest 3D games which always need more & more - at least this holds true today & maybe for the next 5 years. I ran with 512MB for 6 years before going to 1GB. I'll upgrade to 64-bit system with 3GB RAM later this year. Do I need it? No. I can still do everything fine with 1GB on 32-bit system but it'll be nice to have & speed up a couple things for me.

If I asked you why you need 64-bit OS - guess what you'd say. Just because Haiku is supposed to be Media OS, just because it should compete with Mainstream OSes, just because 64-bit is the future, etc., etc. And in the end you very likely don't have use for 64-bit today or even 5 years from now. Not one user that has pushed for 64-bit OS has even mentioned what program ( not 3D games ) they run on Linux or Windows that requires over 4GB of RAM. And I really don't believe they'll be able to still in 4+ years from now.

Re: Steady Progress towards Alpha 1

I think it is understood and accepted that for R1, Haiku won't rock the boat. That is what kept Haiku afloat all these years. So 64 is out for r1...but for r2 we should have it in mind.

You mention that most people don't do too much with their pc. But they are least likely to switch to a new OS anyway. Those that would consider a new OS are more tech savvy and want to do more with their computers.

Re: Steady Progress towards Alpha 1

Idea killers
1. It doesn't grab me.
2. It's not in our image.
3. It's not in our style.
4. It sounds too simple.
5. It sounds too complicated.
6. We'll never find the time to do it.
7. Sounds crazy to me!
8. We've never done anything like that before.
9. Has anyone ever done anything like that before?
10. It'll turn everybody off.
11. Yuck!
12. That's not consistent with the way we do things here.
13. How in the world did you come up with that?!
14. Let's be realistic...
15. Come on...get serious.
16. Great idea-but not for us.
17. People will say we're silly.
18. People will say we're reckless.
19. What will people say?
20. It'll never work.
21. Do you really think that would work?
22. I don't know...
23. Why bother?
24. Sorry...try again.
25. That's a new one to me.
26. That's very interesting, but...
27. That's fantastic, but...
28. Yes, but...
29. Who's going to do it?
30. (Silence)

Re: Steady Progress towards Alpha 1

This progress really looks promising, thanks especially for the mixed GCC4/GCC2 support!! GCC2 was really giving me a headache.

Few words on the 64Bit topic:

I didn't like the fact Haiku would stay a 32Bit OS for a long time since I'm using a computer with more than 3.2GB of ram. I was using a 64Bit Linux for a while now just to see thousands of problems with programs I wanted to use that were not ported to 64Bit yet.

I was thinking it would be a good opportunity to start Haiku as a 64Bit system compiled with the newest compilers even though that means it would break backwards compatibility. Actually I think BeOS had some really nice programs I loved to use, but we have to look into the future and old programs usually need updates to meet futures needs anyway. Why not make sure they work on 64Bit first and than release Haiku? That way Haiku wouldn't have the problem Linux has, that a big part of the software pool will not be able to run or run faulty on your installed version of the OS. New software will run, old won't. Every software would be 64Bit meaning less problems in the future.

But that is not the point. It's not just the programs that won't work on Haiku. Haiku as a 64Bit OS won't work on "older" hardware. These days you get a 64Bit Computer for less than 300€, but why throw away good "old" hardware if the OS would actually just fly on it? It will take a while until Haiku has enough software to meet other needs than the usual surf and mail station. Nobody with a clear mind will buy a computer to use such an OS. But I'm sure Grandma(tm) might need something to check mails and hey, Haiku might be easy to use, and there is an old dusty machine in the basement... . I'm not even mentioning the driver situation for new hardware.

That's why Haiku needs support for old computers!

As somebody (sorry it's so much read to find it again :-s) mentioned above, it does not make sense to maintain a 64Bit and a 32Bit version of Haiku. This will open space for new bugs and every old one will have to be fixed in two branches. Since Haiku is a project that people drive mostly because they love their system and big parts are even done in their free time (at an immense pace!), it doesn't make much sense to force them into doing 64Bit.

I don't think it's about performance and it is a valid point to want to use 64Bit and 16GB ram if you have it in your machine! It would even be a good idea to do the switch to 64Bit now.
But it would be a great show stopper for Haiku since it would build a barrier between where Haiku could be an interesting alternative for end users and the machines it could run on. That's why I think people are right who say lets make it run first, then make it run real good and when we're bored or big, lets talk about 64Bit.

32Bit will make others laugh at Haiku, because of the lack of 64Bit support, but I think that's better than having to struggle with two branches making every programmer have to support both 32 and 64 Bit programs. It doesn't work well on Linux, why should it work on Haiku? Lets let them laugh at our 32Bit OS and we laugh at their "not runing in every OS-version" programs.

The people who don't post in forums but built lots of code in the mean time did a good job, lets let them decide what they *want* to do and enjoy the results (or get involved ourselves).

Best regards,
Thomas

Re: Steady Progress towards Alpha 1

Is there anyway to get 64bit Haiku to run on an older chip? Stupid question i'm sure.

Re: Steady Progress towards Alpha 1

mikedoth wrote:

Is there anyway to get 64bit Haiku to run on an older chip? Stupid question i'm sure.

If by older you mean 32-bit, then no. A 64-bit operating system requires a 64-bit CPU. You can check whether your Intel or AMD PC has a 64-bit CPU by looking at the CPU ID flags, 64-bit CPUs have the LM (long mode) flag.

Re: Steady Progress towards Alpha 1

mikedoth wrote:

Is there anyway to get 64bit Haiku to run on an older chip? Stupid question i'm sure.

No, there isn't a way to run 64 bit code on older chips, but 64 bit chips can run 32 bit code. That's why it makes perfect sense for Haiku to target x86_32 as the first platform. All PCs, both old and new, should be able to run it (theoretically, as there may be drivers missing for certain hardware).

Of course eventually there will be a 64 bit port, but it really makes no sense to devote resources to it now.

One of the main reasons Haiku is making such solid progress is that they haven't been sidetracked by issues such as 64 bit, which means instead of a Haiku 0.0.0.1 that runs on everything from toasters to supercomputers we might actually have a chance of hitting an R1 that works on the huge majority of machines people are actually likely to use.

Re: Steady Progress towards Alpha 1

stippi wrote:

All valid points. But which part of "we have more important things to do right now" don't you understand?

This is a very strange reaction. Perhaps the trouble is that I am reading this in the context of a forum discussion and apparently you are seeing it as a list of reactions to a post you intended as mere glad-handing.

Quote:

There have been plenty of arguments given here on why it is not the most important thing out of the list of missing stuff which is surely not short. So why don't you let us focus on stuff that is higher up, instead of expressing your discouragement here once again?

I'm sure comments in a forum don't prevent you from focusing. Particularly when the comments don't ask you to focus on or not focus on anything.

Quote:

And you know, I do kind of take offense with regards to your "Haiku was mostly not designed at all" remark. Just because Haiku is binary compatible to BeOS and we are implementing an existing API does not mean we didn't have to design stuff, you know? The bulk of the code is the design of the respective developers. Take drivers for example, take stuff like the app_server, many things. With the same ignorant approach, you could say Windows was not "designed", it's just the same ideas from other OSs - a kernel, filesystems, applications, DLLs... nothing new.

But just about any question I have about the design, Haiku developers can only point at BeOS and say "We copied that" which is not design at all. Why does every BWindow have its own thread? What decides which UI drawing is done by the application directly and which by the app server software? Why is media node processing asynchronous ? These and many other key decisions which impact the desktop user, weren't made by Haiku but by Be Inc.

And indeed you can see that influential Microsoft people such as Raymond Chen and Michael Kaplan have commented that some of modern Win32 isn't really designed, it's just the way it is as a side effect of decisions taken in Win16 or in much earlier Win32 releases. Some recent APIs were never intended for public exposition at all, and were published to meet the requirements of anti-trust rulings. The Wine core developers would probably agree with this sentiment as well.

The phrase "designed from the ground up" to me is misleading here. It would be false to claim for example that Linux was "designed from the ground up" to be a portable operating system kernel, even though today Linux is the most widely ported and among the most portable. Linus set out to build a 386-only system.

So far as I know or can tell most of the assumptions in the BeOS design went unquestioned by OpenBeOS developers, whose actual goal was to clone BeOS R5 not to design anything "from the ground up". The choice of kernel was made on its purported resemblance to that of BeOS for example. The most notable deviation from that original path has been the TCP/IP stack, written, re-written, abandoned and re-written again with a novel design.

Quote:

So easy to write a discouraging comment on some website. I would love to learn what stuff you accomplished in your life. Did you ever build something and try to realize one of your dreams? I would be quite surprised if you ever did, because I would assume someone who did wouldn't diss other peoples stuff so lightly.

Some of my work was recently demonstrated to Semantic World 2008 as well as at this year's World Wide Web Conference. If a point of criticism is real, then the critic is not at fault in pointing it out. If it is not real, how can it sting? I have nothing to prove to you. On the contrary, for almost seven years now Haiku has appealed for and received developers, and money, plus a heck of a lot of free or very cheap exposure, to produce this supposed simple yet powerful system but there is little to show for it. Haiku has everything to prove.

Re: Steady Progress towards Alpha 1

@NoHaikuForMe

You are taking the expression out of context. It is used in a much broader (less strict) sense, where designed from the ground up for desktop computing actually points to the (undeniable) fact that Haiku has a focus on personal computing, as opposed to, say, Linux and the BSDs, which originated and are still focused on the server.

So, you make a good point: by a very strict definition of what to be designed means, you can certainly argue that you are right. However, that does not necessarily mean that the statement that "Haiku is designed from the ground up for desktop computing" is false or even misleading. I will grant you, though, that we could have chosen a better wording, in order to avoid agenda-driven interpretations. ;)

With regards to your point about criticism on the web, this is what I make of it: you either still care about Haiku (in spite of your choice of nickname), which is why you keep coming back to voice your opinion (in the hope that it can be useful somehow?), or you are simply a poor soul that has nothing else useful to do than to try to spoil other people's dreams and goals. I would want to think it's the former, but if that's the case, you have a very strange and destructive way of trying to help... :)

Re: Steady Progress towards Alpha 1

NoHaikuForMe wrote:
Quote:

And you know, I do kind of take offense with regards to your "Haiku was mostly not designed at all" remark. Just because Haiku is binary compatible to BeOS and we are implementing an existing API does not mean we didn't have to design stuff, you know? The bulk of the code is the design of the respective developers. Take drivers for example, take stuff like the app_server, many things. With the same ignorant approach, you could say Windows was not "designed", it's just the same ideas from other OSs - a kernel, filesystems, applications, DLLs... nothing new.

But just about any question I have about the design, Haiku developers can only point at BeOS and say "We copied that" which is not design at all.

de-sign, v.intr.

  1. To make or execute plans.
  2. To have a goal or purpose in mind.
  3. To create designs.

Sounds like Haiku was designed from the ground up to me. Just because it doesn't fit your nicely-packaged notion of what "design" means doesn't mean that it wasn't in fact designed.

Furthermore, if it is indeed misleading talk to someone on the Web site team or post on the mailing list, and maybe it can get revised. Posting a reply to a thread regarding something completely different is not the best channel for improving the matter.

NoHaikuForMe wrote:
Quote:

So easy to write a discouraging comment on some website. I would love to learn what stuff you accomplished in your life. Did you ever build something and try to realize one of your dreams? I would be quite surprised if you ever did, because I would assume someone who did wouldn't diss other peoples stuff so lightly.

Some of my work was recently demonstrated to Semantic World 2008 as well as at this year's World Wide Web Conference. If a point of criticism is real, then the critic is not at fault in pointing it out. If it is not real, how can it sting? I have nothing to prove to you. On the contrary, for almost seven years now Haiku has appealed for and received developers, and money, plus a heck of a lot of free or very cheap exposure, to produce this supposed simple yet powerful system but there is little to show for it. Haiku has everything to prove.

You've been a naysaying troll since I can remember. Time and time again you've come here and demanded Haiku prove its worthwhile. The way I see it is if you're not contributing anything of value to the project, you have no reason (or right) to criticize. "Put up or shut up," as they say.

Maybe your criticisms have some weight. Fine. The project, however, doesn't have unlimited resources. Decisions have been made and a consensus has been reached. The developers are working toward a goal that has been set. Your points, however valid, are completely moot at this point in the project's lifespan.

Re: Steady Progress towards Alpha 1

NoHaikuForMe wrote:
stippi wrote:

All valid points. But which part of "we have more important things to do right now" don't you understand?

This is a very strange reaction. Perhaps the trouble is that I am reading this in the context of a forum discussion and apparently you are seeing it as a list of reactions to a post you intended as mere glad-handing.

Quote:

There have been plenty of arguments given here on why it is not the most important thing out of the list of missing stuff which is surely not short. So why don't you let us focus on stuff that is higher up, instead of expressing your discouragement here once again?

I'm sure comments in a forum don't prevent you from focusing. Particularly when the comments don't ask you to focus on or not focus on anything.

Actually, this isn't a forum post - it's the comments section on a news article (written by Stephan) that was published to the front page of haiku-os.org and recently linked by osnews.com as well. The article is specific to Haiku's progress towards releasing an Alpha 1 - which clearly will not include 64-bit support.

Re: Steady Progress towards Alpha 1

I created a post saying why 64-bit would be ideal choice for R2. Feel free to read & contribute to it.

http://www.haiku-os.org/community/forum/64_bit_for_haiku_r2

Re: Steady Progress towards Alpha 1

NoHaikuForMe wrote:

If a point of criticism is real, then the critic is not at fault in pointing it out. If it is not real, how can it sting?

It doesn't have to be real to sting, it also stings when it is well written, appears to be well informed and is posted on our website where it is read by a lot of people of whom many are likely not deeply informed about the history and especially the code of our project.

Quote:

I have nothing to prove to you. On the contrary, for almost seven years now Haiku has appealed for and received developers, and money, plus a heck of a lot of free or very cheap exposure, to produce this supposed simple yet powerful system but there is little to show for it. Haiku has everything to prove.

How am I supposed to understand that? That we are a bunch of con men who trick people into believing we can actually write an operating system and then feed on their resources?

There is only two ways to look at this: Either you believe or you don't believe that we are working our best to realize Haiku. If you do not want to diss our work, then you have to look elsewhere for reasons on why Haiku doesn't have much to show for it, as you put it. (Hint: The speed of our progress might be related mainly to how many there are of us, and not so much on our skill or motivation.)

I could apply your very same argument to the Linux world: "With that many people working on Linux projects, it should be much more advanced". Note that I am only relating my perception of the resources that Linux projects have to my perception and expectations of what achievement Linux is and should be as a technology. And I would be showing my ignorance of how Open Source works, which I believe is what you are doing with your supposedly founded critique on our project.

To me it looks like you think you have the right to expect "something", at least more than what Haiku has to show for itself from your perception. And then you "back up your expectations" by supposed arguments, which IMHO show only your ignorance with regards to how open source projects function on the one hand and on the other hand you cannot do that without belittling our work. And then if people take offense in that, you think you must have told something very true, or else they wouldn't have reacted. Can you see how that line of thought is flawed?

Re: Steady Progress towards Alpha 1

koki wrote:

@NoHaikuForMe
You are taking the expression out of context. It is used in a much broader (less strict) sense, where designed from the ground up for desktop computing actually points to the (undeniable) fact that Haiku has a focus on personal computing, as opposed to, say, Linux and the BSDs, which originated and are still focused on the server.

The trouble with this popular fairy tale is that none of it's true. Linus Torvalds didn't own a server, he'd bought himself a 386 desktop PC and decided to write an operating system kernel as a hobby. He wasn't interested in creating something big and professional, just a desktop system for himself and hopefully like-minded Minix users.

Meanwhile the USL vs Berkeley court case was slowly coming to a conclusion. Although BSD had been intended for minicomputers it seemed possible to make a modified version run on Intel's new 386 which was available in the form of IBM PC clones like the one Linus had just bought. Once the case was settled this work lead to the creation of early desktop versions of BSD like 386BSD and from there FreeBSD. Some of the people working on 386BSD may have seen it as a way to build servers from cheaper hardware, but most just wanted a desktop Unix.

Quote:

However, that does not necessarily mean that the statement that "Haiku is designed from the ground up for desktop computing" is false or even misleading.

I don't think there's an honest dispute here, the Haiku blurb just doesn't reflect the actual project.

Quote:

With regards to your point about criticism on the web, this is what I make of it: you either still care about Haiku (in spite of your choice of nickname), which is why you keep coming back to voice your opinion (in the hope that it can be useful somehow?), or you are simply a poor soul that has nothing else useful to do than to try to spoil other people's dreams and goals. I would want to think it's the former, but if that's the case, you have a very strange and destructive way of trying to help... :)

A common mistake is that people project something's outcome onto a third party as intention. For example, you see a man throw a stick into a tree, then an apple falls from the tree and the man eats the apple. You could conclude (and you might even be right) that he threw a stick into the tree to dislodge the apple. But it is also quite possible that he threw a stick into the tree without any specific purpose and the fallen apple is an unexpected bonus.<