On 27/10/2020 9:03 pm, Theo wrote:
> Johann 'Myrkraverk' Oskarsson <joh...@myrkraverk.invalid> wrote:
>> What is a good way to get started with programming for RISC OS? I used
>> to have an emulator with RISC OS about a decade or more ago, and back
>> then I played a bit with GCC, but never got really far, and I've pretty
>> much forgot /everything/.
>>
>> For the record, I'm already proficient with C, C++, and other languages
>> and operating systems, including Posix. ARM and RISC OS are new to me.
>
> Where does your interest and skill set lie?
All over the place. I have a history of contributing to a wide variety
of projects, both as part of my work, and in my free time with open
source projects. My project history includes device driver code, though
what I can show would be software only projects [on my blog] for OS/2
and Illumos [formerly OpenSolaris]; database code, server plugins for
Postgres, SQL clients, and such things; some graphical and desktop
applications; and a whole bunch of other stuff.
That would make me a Jack of all trades, so to speak.
> There's various points in the stack you might start:
>
> Desktop apps: C, C++, BASIC are popular. Other languages like Python and Lua are
> a bit more niche (and more likely support is a bit more patchy)
A desktop app is something I might do as practice.
> (User-facing) Modules: filesystems, system services, drivers. That's mostly C (or
> assembly, for the masochists).
This would also be high on my list of interesting things.
> Operating system: modifying deeper level stuff, porting to new hardware: C
> or assembly.
This might also interest me, though it would probably depend on the
scope of any individual project.
> The DDE is needed for actual OS development, but its C++ is ancient and the
> tools can be a bit quirky (there's a Make but it's not GNU Make, for
> example). GCC has good C++, shared library and cross-build support but
> isn't supported for building the OS.
I tend to prefer C over C++ these days, and I try to avoid GCC when I
can, for my own reasons. That's why I went for the DDE to begin with.
>> I guess I'm going to need some hand holding with day to day operations,
>> and read the user's guide, and that's ok. I'll try to keep user
>> questions to c.s.a.misc though.
>
> If you don't get answers here, the ROOL forum is also worth posting to.
> (many only read one but not the other)
Thanks. Right now I'm one of them [posting here, not reading the
other.]
I think the bounties and wish lists are something I'll look at more when
I've gotten some RISC OS experience.
> There is typically a skills disconnect between people who can, say, debug
> the USB stack, and developers who work on application code in BASIC. This
> is why many bounties are slow to be taken up.
>
>> What are people willing to /pay/ for? If I can get my employer on board
>> and sell some software, then we'd be discussing a lot less time to
>> delivery than if I only work in my free time.
>
> The usual problem is the market size. People sometimes wave cash around
> ('I'll pay £50 for X') but if only three of them want it, it's not cost
> effective. I'm not sure any RISC OS project is cost effective in terms of
> commercial rates.
>
> If any of the bounties work for you, that could be a place to start.
The bounties /may/ work, but I'll not start out with them.
> You could also ask if there are starter projects you could try.
> (many open source projects have them, but ROOL hasn't traditionally)
What do you mean with a starter project? I'm not familiar with the
term.
> If you have time you could commit, it could be worth talking to someone like
> R-Comp to see what they see as pressing needs. They probably have more
> contact with real customers than most.
Thanks, I'll keep it in mind when I have some experience to show.