TBA Software

After a short 12 year gap away from everything RISC OS (!) TBA Software has returned. The original 3 man team (now based in Bristol, North Yorkshire, and Singapore!) are back together again in a very 2011 kind of way.

Games including AXIS, FTT, Cyber Ape, Cobalt Seed, and BHP, were all popular back in the 1990s. Indeed I’ve found and dusted off some original stock left over from 1997 which I’ll be putting up on e-bay for all the collectors out there.

As the former owner and lead developer, I have dug out the PRMs, fired up and upgraded the RISC PC, purchased a BB XM, and got stuck into making all those nasty pc mangling 26bit instructions work in the 32bit world.

Some info is available at www.tbasoftware.co.uk and progress will be reported there.

After a week or so of going at it, TBAFS is running, TAG (the 3D engine) is making progress, and plans are a foot for making the whole 3D API much more accessible. All of our release code is pure assembler (compiled with the BASIC assembler) and first impressions are that things run like a rocket on the BB XM!

I have to say I’m very impressed with the huge amount of effort that’s been put in to getting this open version of RISC OS 5 running on the BB XM (and other modern systems). Hopefully we will be able to contribute something useful in the weeks and months to come. Naturally we have a particular interest in gaming and I’ll start another thread regarding Open GL and hardware acceleration in due course.

Once I’ve completed some prep work, we plan to open up all of the source code to our products and make things as accessible as possible to other developers out there.

Many Thanks!
Alan

Excellent news, and welcome back :)

Games on RISC OS are a subject I’ve got a lot of interest in, I know not everyone is interested but it’d be great if we could encourage a bit more game development.

As soon as I get myself a website sorted out I’m going to release some of my own offerings to the world.

Really fantastic news! Welcome back – looking forward to seeing what comes out of your renewed work.

Rik – looking forward to your website coming on line, too!

Excellent news indeed!

I’ll start another thread regarding Open GL and hardware acceleration in due course.

Perhaps Simon Storr will pick up Exile GL again!

Hi Alan! Glad to see you back.

I’d like to speak/email you about a couple of things, but couldn’t find an email address on the blog (I may be blind). What’s the best contact details for you? If you want to keep this off the list, I’m at rci@rcomp.co.uk

couldn’t find an email address

The legal note at the bottom links to http://www.applettechnology.co.uk/, which does have a contact page.

For the benefit of anyone not following the above blog:

"Some simple tests show the BB XM is comfortable rendering in 640×512 in software at 60FPS[…]"

That’s a shame for those TVs with low frequency displays!

Alan, one of the things the blog says is:

Longer term there are many things RISC OS requires (decent web browser, threading, BASIC enhancements) to make it attractive for more users that we hope to play a part with.

Is TBA considering picking up Rozilla or working on NetSurf? (Some Rozilla c.s.a.* refs: 1 2 3 4 for anyone who’s forgotten.)

Wouldn’t Firefox be more relevant?

http://www.riscos.info/index.php/Mozilla_Firefox

(Would Chrome be an option? Since Google wish to make it into a cloud computing client, that would sort out a massive amount of missing functionality of RISC OS apps)

Wouldn’t Firefox be more relevant?

You’re right but I wonder if any Rozilla code could be incorporated (in either Firefox or NetSurf). We’ll have to wait and see how much time TBA have had to think about this.

If TBA do have time to spare for browser activity, hopefully this will be directed towards a javascript-capable browser such as Firefox. NetSurf is excellent and I use a recent development build daily as my default RO browser, but there are many sites which require JS in order to interact with, as opposed merely to view. Rather than start from scratch with Chrome or Opera, why not take advantage of the work already done on the Firefox port?

If TBA do have time to spare for browser activity, hopefully this will be directed towards a javascript-capable browser such as Firefox. NetSurf is excellent and I use a recent development build daily as my default RO browser, but there are many sites which require JS in order to interact with, as opposed merely to view. Rather than start from scratch with Chrome or Opera, why not take advantage of the work already done on the Firefox port?

Or how about focusing development on http://www.netsurf-browser.org/projects/libdom

Rozilla is incredibly ancient by browser standards. The only useful bit was the threading code, which was imported into Unixlib and now powers Firefox and NetSurf. So there’s nothing to be gained by digging up Rozilla again.

Opera is closed source, and they’ve previously said they need an audience of 1 million (or something) before they’ll do a port. So not much chance there.

Chrome runs each tab in a separate process, which isn’t at all friendly to RISC OS. Performance would be dire.

Firefox’s port needs some reworking: FF3 dropped the Xlib backend that RO FF uses, so a new port would need to port Pango and Cairo for rendering, and integrate them with the RISC OS font system. I think a little work on this has already happened, but it’s a fairly big job. Sticking with FF2 isn’t a great plan for security reasons.

On the other-browsers front, I think the best bet is a WebKit browser. There’s GTK WebKit, Epiphany (in C), and Origyn (an embedded browser):
http://en.wikipedia.org/wiki/Origyn_Web_Browser
But again the question comes up of how RISC OS integrated it would be.

Opera is closed source, and they’ve previously said they need an audience of 1 million (or something) before they’ll do a port. So not much chance there.

This is probably still the case. But it should be noted that the person who quoted that audience hasn’t actually worked there since 2005.

“On the other-browsers front, I think the best bet is a WebKit browser…
But again the question comes up of how RISC OS integrated it would be.”
IMHO, RISC OS integration should come low on the order of priority: 1. we already have a well-integrated browser (NetSurf); 2. if that doesn’t do the job, most of us switch to a non-RISC OS platform and launch a browser that will, so implicitly accepting a non-RISC OS GUI. If we can’t have both functionality and integration, my preference is for functionality every time.

But the Opera person did actually work for them at the time (back in 2001), so that’s as close to the horse’s mouth as we’ll get. Anyway, they’re much more focused on the mobile space these days, rather than desktops.

Of the WebKit browsers, Origyn looks interesting. It has an SDL backend, for which RISC OS libraries already exist. That would mean it’d operate full screen, like a game (or possibly in a window that thinks it’s full screen). So it might not be such a difficult port. Alas the Origyn website is broken at the moment.

[…] did actually work for them at the time (back in 2001) […]

Yes, I saw the date – but just meant that the policy/direction may have changed since the person left.

[…] much more focused on the mobile space these days, rather than desktops.

So the policy/direction has indeed changed – but unfortunately not in favour of RISC OS. Thanks for the update.

Origyn […] SDL backend […] it’d operate full screen, like a game (or possibly in a window that thinks it’s full screen).

Interesting, but not great for saving downloaded files (not sure if in a window would help that – correct me if I’m wrong). And sorry for dragging this thread off topic, back up there!

I mailed Origyn to point out their site is broken. Response:

Subject: Pleyo has run out of business Re: OWB Trac broken
Pleyo company is discontinued and does not exist anymore. We apologize for the inconvinience.

So there goes that idea. I’m now doing a recursive wget on their websvn just in case the source code ever comes in handy, as it’ll probably die soon.

Shame :-(

"Pleyo has run out of business"

doing a recursive wget on their websvn

Good idea! Should I see if their offices are abandoned and left unlocked this summer? Maybe there’ll be some unreleased dev code lying around. (For any concerned Pleyo staff reading this, please take this in the humourous manner intended.)

Wow it seems this thread on TBA returning has headed off into a discussion on web browsers!

A modern web-browser resembles a mini operating system – some potentially have more code than large parts of RISC OS – so it’s not an easy task to port something across. Before this is even attempted there would need to be some major work on RISC OS to put some of the building blocks in place. To start with there is threading, memory protection, libraries for certificates/encryption, Java… the list goes on and on. Then there are plug-ins… Chrome is the ideal browser but would be the mother of all projects to even start to try and port.

I quite like Netsurf and it has potential. For now the certificate errors when using RISC OS 5.17 need fixing, as do the crashes when alignment errors are turned on. Java is probably the biggest thing required for this project. I have no idea what ever happened to the Acorn version of Java?

My particular interests relating to the operating system are:

File systems
TBAFS has potential as a new disc format. With the addition journalling and double copy of the FAT it would be robust enough. It already supports unlimited files per directory, long file names, a fully indexed FAT, and re-entrant caching code with interrupt based lazy writing. (A 32 bit version of the current image file system will be on our website for free shortly – after I’ve tested it a little more!)

BASIC
After 10 years full time using VB.NET, plenty is missing in BBC BASIC. While porting .NET would be pretty impossible (!) a newer BASIC language with similar syntax and all the useful features might not be. Being able to develop class libraries and modules in a higher level language would be great. And a proper debugger….

Threading
Setting up a pre-emptive core similar to the process management in WIN32 wouldn’t be that hard – but re-coding the OS to use it would be! I can remember getting pre-emptive multi-tasking working in a demo on the BBC B so nothing is impossible.

Open GL
I’ve taken a quick look at the old Mesa port and it is pretty straightforward. Challenges would exist in looking at the later Mesa code base. Performance isn’t great because the drawing code isn’t well optimised, and as usual there are issues with how well the C compiler can use hardware floating point. Hardware Open GL for PowerVR/SGX remains a dream as it is closed source. Using the Linux binary might be possible if someone with a lot of experience using it was available to work on a wrapper – unfortunately this is not available within TBA.

More than enough for now. A lot more discussions to have in due course!

Wow it seems this thread on TBA returning has headed off into a discussion on web browsers!

Presumably due to Rozilla.

Chrome is the ideal browser but would be the mother of all projects to even start to try and port.

So would it actually be easier to graft RISC OS onto ChromeOS as a subsystem?

what ever happened to the Acorn version of Java?

Didn’t the license expire? (But are the terms still the same?)

TBAFS

It sounds interesting, but as an image filesystem it would be stuck to 2GB partitions. Can it be made into a full filesystem and would it be available for other OSes?

Threading

https://www.riscosopen.org/forum/forums/5/topics/428

> For now the certificate errors when using RISC OS 5.17 need fixing,

We look forward to your bug report :)

> as do the crashes when alignment errors are turned on.

Make sure you’re not using Tinct for rendering images; use OS functions.

> Java is probably the biggest thing required for this project.

When was the last time you came across a website that needed Java?

> I have no idea what ever happened to the Acorn version of Java?

It stagnated. Hardly anything runs in Java 1.2.

“A modern web-browser resembles a mini operating system – some potentially have more code than large parts of RISC OS "

The Slitaz linux has just released a javascript and flash capable browser from webkit that is only a 24 KBb binary.
http://twitter.com/slitaz/status/62091478582837248

As Theo mentioned, we need to use the sdl library, I noticed that ffplay uses sdl so
it may not be an unreasonable suggestion.

Ron M.

Slitaz

• Code, including current README and docfile

TBAFS […] on our website for free shortly […]

TBAFS 32bit Beta Released