BeIA was just a web browser… it’s not like there is much to recreate. The app embedding and the stuff Binder was used for is the main part you would want, and that would need to be completely reimplemented (embedding) and ported (OpenBinder)…
The clever part of BeIA was the file system and exe compression. These days, that is not as useful as storage is a lot cheaper.
I think having a proper API with actual UI controls would be far superior.
Not to mention Haiku packagefs already has some compression (I don’t know how they compare too, are there ideas we could borrow from BeIA in that area still?)
While I do enjoy this constructive discussion on Haiku on phones, I would recommend staying away from the “Maple” name: it is a registered trademark for a maths software from Maplesoft.
Something to remember if someone starts coding for the “Haiku on phone” project and we start having downloads/redistributables/logos/etc…
@Yann64
Agreed. Since the name is already taken, I’ve modified the title of this thread to just “Haiku on phones general”. We need to avoid infringing trademarks always.
@thread
Since all discussion has centered around the user-interface so far and could maybe extend to BFS as defining characteristics, would it be appropriate to suggest that we look at using another modular OS like Genode at least as a starting point? They already have a 2 year head start on the PinePhone port. I don’t know if their SculptOS GUI will be good but replacing just the parts that don’t work for us would still save time and effort.
AFAIK Genode uses Linux drivers on the PinePhone port. However, how much of it would be left and how much from Haiku would have to be added in?
The core premise of the Defintely-not-Maple idea is to have a mobile OS that can be used as either a basic phone OS or as a Haiku desktop when connected to external displays and peripherals (e.g. mouse and keyboard).
The point I was trying to make toward using Genode as a starting point would be that we could replace bits and pieces as we go. Yes, Genode is AGPL and uses GPL drivers from Linux. I don’t like the licenses but the possibility of offering a secure OS on a phone other than Linux sounds appealing to me. I’ve heard of the 5 Eyes governments as using tainted drivers as a means of spying on people so having completely open-source drivers and making them use micro-kernel user-spaces for their memory spaces sounds like it would be one means of hardening the PinePhone against invasion.
Isn’t one of the reasons Haiku exists to avoid the entire “mobile operating system” fallacy that GNOME, MS Windows, and many other environments have fallen prey to? Interestingly enough, Android has also fallen into the exact opposite fallacy, and is trying to become a viable desktop operating system.
None of them make for good mobile computing experiences, and Android is not a good desktop either. So I can’t see why people want Haiku to go down a well-trodden path to failure.
In addition, there is also a rule applied to naval construction, which is to avoid “gold-plating”, especially that of unfinished products. In the words of former Secretary of the (US) Navy John Lehman:
Gold-plating and a change order culture were ended, which enabled fixed-price contracts and annual production competition. Navy shipbuilding actually had a net cost underrun of $8 billion during the Reagan years, the first and only time in history. Congress saw that we kept our word and did what we said we would, and gave us its support year after year.
Working on the mobile version of a desktop operating system (which will turn out to be an entirely different operating system) when the desktop is not even complete definitely fits the definition of “gold-plating” and “change order culture.”
And what makes you think change-order culture and a tendency towards gold plating will spare an operating system developed by hobbyists, if it drove even the US Navy to what was in effect the bottom of the ocean in the 1970s?
Agreed. And I fear that development on a Haiku Phone project would:
Fail, its a ton of work that deals with very low-level details. How long has it taken to get ARM working? Not a slight against @davidkaroly and others, but it’s really difficult & slow work, and there’s only a few developers who understand how that stuff works. Who’s going to write the modem drivers? The GPU drivers? The bootloader? Implement the GSM protocols? Get the code to work on whatever various hardware? It’s nice to sit around and brainstorm UIs and how it looks, but that’s the tip of the iceberg, and the stuff underneath is a ton of work most people don’t do unless they’re paid to do it.
Take away from important work that would benefit Haiku, such as VPNs, webcam, and other nice things that people often complain about. Leaving an unfinished project to take a chance on a phone OS (which seems to have an extremely low chance of success) means a risk that nothing overall gets done.
Full mobile functionality including communication and apps
Windows 10 apps scale to phone screen size
The separation of desktop and mobile modes including applications used is to avoid running into the pitfalls of attempting convergent applications encountered by most mobile OSes (including WP).
“Mobile and desktop interfaces with separate apps” is exactly what Microsoft tried with Windows Phone. It wasn’t remotely close to successful in the market, so Microsoft replaced it with “convergence” and Windows 10, and judging by how the stalls are still full of Android and iOS devices, Windows 10 is also not very successful in the mobile market.
Also see what I said above about excessive gold plating.