Supporting a new WASM-based application runtime would be no different than supporting Java apps, Qt apps, or SDL apps in Haiku. It doesn’t mean Haiku gives up on its native libraries and applications, it just provides a new source of potentially useful applications that also have a good security model around them, as described in the article linked in the OP of this discussion.
I appreciate the simplicity of Haiku and how it is possible for a single person to understand almost all of it, and I also would not want to see that lost. I don’t think it will add any more complication to Haiku to experiment with WASM apps in Haiku, or support new WASM runtimes than the complication added with supporting Qt or Java apps or SDL games.
The matter is not about technical complexity, but about user experience. The main selling point of Haiku is the tight integration of native apps, in some way, the idea of UNIX pipes brought into the GUI age.
In Haiku you can easily combine applications and make them work together. Paladin + Koder is an example of this, as well as all the demos we do with creating a query and converting it into a MediaPlayer playlist, using different mail clients with the same contacts database and email data, etc.
With ported apps, a large part of that is simply lost. So, they are nice as a stopgap, to get Haiku into a basic workable platform, but at the same time, if they do the job well enough, will people bother to write native apps? Will all the effort in making the app integration possible go to waste because no app bothers to use it?
This is where Haiku could turn into basically just yet another OS to run the same apps, and one that doesn’t get as much attention and support from app developers.
So that’s my view on it: yes, probably someone will make it happen. But I think the core Haiku team should focus first on the needs of native apps.
I have a similar vision as you do PulkoMandy.
WebAssembly is a promising technology. For web applications which can be compiled locally and run faster on different target (SAAS market).
Also for the cloud server market, this is just perfect once WASI is fully specified. You can then develop your app and run it natively on any kubernetes managed service.
The best option for Haiku is to be a development platfrom for anyone willing to develop such apps. But Linux is an easier candidate.
The problem is that native apps in haiku only run on haiku. The most people want to develop on Portable things. A Company too, because they can make more money to support other platforms.
Yes, but we will always be “second class” for portable apps. People writing these will focus on other OS where there are a lot of users, and so these apps may run on Haiku, but there will be less official support from the app developers.
So, it creates an unfair comparison for users: portable apps are easily compared from one OS to another, and it seems Haiku is not as good as other systems. Not because it is, but because the developers of such apps did not put a lot of effort into integrating them with the OS. And, on the other side, if there are no native apps, all the places where Haiku can bring things further are not exposed. What’s the point of all the cool tech (queries, etc) if no one uses them?
Not all do because supporting other platforms means higher support costs any many companies do not want higher costs even if they might make more money (or rather, majority of companies consider the costs to be higher than the gains). Stuff like WASM makes it even worse on the support cost side.
Compared to other OS, Haiku has advantages even if native software is not used:
It is lightweight. It not use a lot of RAM and don’t run hundreds of processes. More resources are available for user software.
It is simpler to manage, it don’t require manual config files editing and complex command line configurations. Install process is fast and simple. Unlike Windows and Linux, Haiku can be easily moved to another PC by moving disk or copying disk contents. Linux require Grub reconfiguration and I don’t and won’t know how to do that. Unlike Linux, Haiku can run even if video card drivers are not available.
Unlike Linux, Haiku has common stable ABI. Application compiled for Haiku have larger chances to be runnable after many years.
Unlike Linux, Haiku has common GUI, clipboard, drag-drop, audio API. It allows to run applications together and allows communication between applications.
Unlike Linux, Haiku is complete OS. It has all OS components including GUI. Linux is only a kernel and it has a lot of incompatible GUI (X11, wayland; Qt, Gtk+), audio (alsa, jack, OSS, gstreamer, portaudio), launch (init scripts, systemd, upstart) etc. systems. Making commercial application that run on all Linux distributions is difficult.
Haiku is really free. It is licensed under MIT license. It allows customization for application domain (ATM terminals, industrial controlling interface etc.).
Haiku has desktop oriented GUI. It has small GUI elements and small spacings allowing to place a lot of information on screen. Modern giant GUI elements are terrible.
There are versa drivers for other os too, i does not Think that this is a big spezial
Why it use so low ram and not many processes, because the available apps does not have so much grafical Elements and processes running. If we ever would get games like in Windows or Big apps like gimp, etc. running on haiku, this point goes away.
Yes you are right, but Who will create apps only running on haiku at the moment? The most one are porting and does not use the Power of haiku. How to make this attractive is the biggest question.
So long simple things (in the eyes of users) not minimum 80% are supported you would not get much users. A great software for a special work can be help, but on wich market?
For people like us, using a os to develop and playing around, Internet and mail it is a very nice system and i love it. But be clear, show the current generation looking for, playing games, streaming videos and music, cloud using, messanger…
It may not work by default and require special configuration not available for non expert users. At least I experienced failure of starting X11 when trying Linux some years ago while Haiku was OK. Haiku run and show GUI almost on any PC I tried.
I think that Haiku can be good alternative of Windows. Windows 10 became very bloated and slow, has spyware impossible to disable by documented methods, many things become broken (for example printing is not working at all for me including printing to PDF/XPS on Windows 10, reinstalling printers/drivers not help), it force to use applications with new closed (impossible to develop apps without permission from Microsoft) API and closed app store. Linux is not suited for commercial applications.
About WebAssembly: it has bad memory model, all memory must be in 1 continuous block. So if application allocate a lot of memory and then free most of it, memory will be still allocated if small heap object are still allocated at end of heap. It also don’t have good exception support yet.
When I went to buy a new laptop a few years ago I bought Haiku on an USB stick to test first. I was looking at an HP that ran the Haiku great.
Bought it and took it home. My god, it just crawled when I booted into Windows 10, it looked good at the store but turned into a pig when I ran any power Windows software. Yet on Haiku the software to handle gigabyte files worked in seconds, Windows 10 just wastes CPU power.
Linux isn’t much faster than Win10 and the biggest, most bloated GUI toolkit from the Commodore Amiga (Magic User Interface) coupled with hosted AROS (a hosted environment like WINE but imitating Amiga) yields a smaller memory footprint than any X11 based GUI.
The point I’m making is that ALL the mainstream OS’s are bloated beyond reason. Haiku and AROS aren’t and it shows in the performance. If we could just get people to leave those bandwagons of compatibility, the tech sector and the world would be a better place!