Was 2015 the end of RISC OS?

I’m not sure to wait for 20 years to get the equivalent of Arthur.

When the jpeg support was updated it was 23 years since acorn did the original & that’s just a small part

As I’m the guy who wrote the only dot-gain correction and monitor calibration system RO ever had, AND hacked the PostScript printer driver to support masked deep sprites and CMYK, I can confirm that yes, it was particularly painful and always has been.

Nuts and bolts techie vs. user.

At work, a phrase comes up at intervals from project managers: “printing just works” and the simple fact is that system after system printing does not just work.
Mildly amusing at the moment is the struggles of the p.m. for the “Managed Print” project arising from not listening to the people that sorted out the printing on all the systems currently in place.

User end thinks it “just works” which it does, because we made it work.

If you talk about a “way to make legacy applications working without a kind of emulation layer”, let’s be clear: it’s impossible, due to processor issues.

It would still be an ARM system.

For 26bit support, ADFFS & Aemulor are the way to go. And they will need an almost complete rewrite to work on another kernel.

Why? Do they use undocumented API/ABI?

I’ve said many times that the task switching should all be in TaskManager (with threading), the UI in WindowManager and the rendering of the UI somewhere else, Desktop probably. That would allow interesting things to be done more easily, but that’s just a change to three modules, not a brand new codebase.

Perhaps you’re right. I’ll look over the RO5 code again.

At work, a phrase comes up at intervals from project managers: “printing just works”

I spent many hours at work trying to get Windows apps to print the way I wanted. So many times, I struggled to get the correct orientation on the paper, or scaled to fit the paper. Shrink to 70% doesn’t mean that an A3 image will fit on A4 paper; it means you get an A5 area of the page with print on, and the rest of the page white. It was usual for the printout not to match the preview.

I spent many hours at work trying to get Windows apps to print the way I wanted.

Call the whole team that worked on Windows printing bad names and see what the reaction is :)

I really want to get a dot matrix printer and hook it up to my print server to print out drafts of my fiction. I would probably be better off with a laser one but who says everything has to be a rational decision?

We do have a Canon inkjet attached to one of the Debian servers but the last time I tried to get anything to print from RISC OS nothing happened (that was with the printing stuff that came bundled with NutPi which I don’t have access to anymore so think I’d have less chance of getting it to work now!).

What is the best/most compatible printer for use over a network (CUPS) with RISC OS? I probably should start a new thread for that question. Also I’m only interested in printing text with it—don’t do any DTP nowadays and any photos we need to be printed we’ll get done by a photographer (or one of those machines in Tesco)

What is the best/most compatible printer for use over a network (CUPS) with RISC OS?

I’m a fan on HP as PCL drivers just work, at least for me, or get any printer with postscript support. I’m not a fan of cups, though.

I’m not a fan of cups, though.

Its my only option for the Canon as its USB and has to be shared with the other computers in the house. Works perfectly with them too and they can print/scan just fine with it. I remember from the last time I tried to get it working though that the general consensus was that RISC OS lacked the right drivers.

I’m a fan on HP as PCL drivers just work

Is that a direct USB connection to the RISC OS machine or an HP printer with built in network port and then some network printing software?

It would still be an ARM system.

Modern ARM processors are the issue for 26bit support. You can change the kernel as you want, the issue will still be here. An emulation layer is needed. And IMHO, it’s not important, as old software is… old.

Why? Do they use undocumented API/ABI?

It’s not only API/ABI. These kind of software uses kernel features. So you’ll need to implement ABI + API + HAL calls + SWI approach, etc. Hum, looks like the current RISC OS after all.

I understand what you want to say: with something more mainstream (POSIX, latest GCC, etc.) it would be simpler to get these news features. But let’s face it: at the end of the road, you’ll get a Linux kernel. Now, you can help porting the latest GCC to RISC OS, work on a module for partitions support, rewrite Wimp2 for RISC OS 5, provide a module for WiFi support, etc. None of these are impossible today and the result will be the same with much less work.

Or you can contribute to bounties :)

Is that a direct USB connection to the RISC OS machine or an HP printer with built in network port and then some network printing software?

I use RComp’s netprint and either a ps or pcl driver. I’m not a fan of print servers, so an lpr/jetdirect solution works for me. Never used USB for a printer though. Only parallel a long time ago and network now.

ARM processors are the issue for 26bit support

I’m not talking about 26bit support. I’m referring to 32-bit RO5 applications.

Hum, looks like the current RISC OS after all.

That’s the point.

POSIX, latest GCC, etc.

Although I said I’d prefer to not have POSIX and I don’t care about gcc (and I hate glibc with a passion)

you’ll get a Linux kernel

No you wont. I’ve mentioned my preference for a microkernel, Linux is monolithic. I’m also referring to building from scratch, so no code from Linux will be near it, and if it isn’t POSIX compliant, it not even a *nix system.

None of these are impossible today and the result will be the same with – IMHO – much less work.

Possibly, but I’m not convinced, yet.

Steve said

Call the whole team that worked on Windows printing bad names and see what the reaction is :)

I was one of the industry experts contracted by Microsoft to help them create XPS and the print system in Windows 7. I worked with their programming and design team. They. Knew. NOTHING. The stories I could tell would make you cry. For example:

MS: Our customers tell us that PDFs sometimes don’t have the fonts included, and that causes problems, so XPS files must have fonts embedded.
Me: No, you don’t want to do that.
MS: Yes. We require all fonts to be embedded.
Me: No, you really don’t.
MS: Yes. We have decided.
Me: Have you tried writing ‘hello world’ in Japanese and printing to XPS?
…some days later…
MS: OMG IT’S 17MB!!!!!!
Me: Yes.
MS: So Japanese fonts, right, are really big!!!!!!
Me: Yes.
MS: Who knew?!!!!!!!
Me: It’s your font.

I could tell you twenty other examples, but then they’d have to kill me.

I’m a fan of HP as PCL drivers just work, at least for me,

Depending on the model, I wrote some of the PCL interpreter (fonts, colour, interpretation, hardware acceleration). The reason HP PCL ‘just works’ is that particular HP models are reference implementations – not that the code is available, but they are a defacto standard… this model for colour, that model for black and white, another model for PCLXL. Reverse engineering them was… hard work. The PCL documentation is useless, and the stuff that PCL drivers actually output is pant-wettingly crazy at times.

We used to joke that the “Spec” in “PCL Spec” stands for “speculation”.

or get any printer with postscript support.

Which is the exact opposite. PostScript is the most carefully specced PDL there is. The Red Book is a joy.

Nuts and bolts techie vs. user.

Exactly this. When working for an estate agency, I needed to create spec sheets for houses. Printing to a Kyocera laser in HP2 emulation.

With RISC OS, it was simple. Easy. Even doing stuff like printing an A4 page at A5 size (so can put the paper in the other way up to print the other side). Making diagrams (!Draw), adding images. All pretty effortless. Unlike Windows, that often got it spectacularly wrong at times (as was pointed out above).

Reverse engineering them was… hard work.

Like trying to get info on URF and… the other one. I forget the name. It’s the bitmap image types used with modern WiFi enabled printers.

I forget the name

One of these?

PostScript is the most carefully specced PDL there is. The Red Book is a joy.

Somehow every time I think of Postscript and its documentation I end up thinking of Tolstoy.

I once created a PDF that had the whole of War and Peace attached as metadata. Perhaps unsurprisingly, almost everything loading it exploded.

The PDF Spec is referred to as ‘The Brown Book’. I’ll pass no comment on it.

The PDF Spec is referred to as ‘The Brown Book’. I’ll pass no comment on it.

You don’t like it brown? I guess that is something of a sticking point? Don’t worry, it’s not that big a deal. A gentle squeeze and you’ll push through to realise that it’s the contents that matter more than the colour.

WiFi were present before

Really? o.0

Yep… here

I’m sure there’s a good reason nobody ever talks about this wrt wireless networking on RISC OS. Maybe nobody bought it, or maybe it didn’t work? Wasn’t the NET100, as mentioned, a product co-developed with RComp, who are clearly still around and developing great stuff for RISC OS.

Not cheap. Could buy a whole OS with WiFi for that kind of money.

I guess you have to remember that would have been about 2003/4? It looked to me like the drivers were only a tenner on top of the USB adaptor, but it still means they existed and is therefore confusing that nobody has (admitted to having) looked at them as a starting point for a new stack.

Where’s the code?

That’s gonna be the $64000 question! I’m assuming Stuart Tyrell Developments or Advantage6, if anyone. Maybe someone has a contact with them – I seem to think an (ex-) employee has been active recently?

Do you not think that may be why people haven’t looked at it? Plus there would be licensing concerns you’d need to sort out too.

Tracking the legal copyright owner of the code, check licensing, then modernizing it is going to be more work than just porting net80211 from a BSD system.

WiFi – I use this with my Pi → https://www.amazon.co.uk//dp/B0062H5SK2
Costs “about fifteen quid”. Hook up an inch sized fan if your room gets warm/hot in the summer (as the thing runs hot), and you’re good to go.

Could buy a whole OS with WiFi for that kind of money.

Could buy six and a half of my dongles for that kind of money.

I guess you have to remember that would have been about 2003/4?

RISC OS stuff doesn’t seem to drop price as it ages. MelIDI doesn’t work on new machines, isn’t an active supported product any more, and yet He Who Has It In Stock lists it as a fart shy of a hundred quid…

Where’s the code?

Why do you need/want the code?

Until the OS has WiFi support native, use an add-on. No faffing around. No source code worries, no build issues. It’ll “just do its thing”.
The second generation of them can be configured from any half-decent browser (that probably runs out RISC OS, but I’ll admit I’ve not tried Otter).

then modernizing it is going to be more work

Another thing to keep in mind is the level of integration. Does the USB dongle present itself as a WiFi device, or as a regular Ethernet device? The thing I use, for example, works with WiFi but it appears to the RISC OS machine as an ethernet something-or-other. RISC OS has no knowledge of how it works, only that it can talk to the outside world. It’s possible that the USB dongle is similar?
It says it has a full network stack based upon the NET100. Follow the link and NET100 is a “10/100BaseT network card”…?

Does/did anybody here have one of these?

That’s gonna be the $64000 question!

You do know the sixty four thousand dollar question was a gameshow from 1955 to 1958? When you take monetary changes and inflation into account, $64,000 (1955) is about $590,920 (2018). But that’s much more of a mouthful. :-)

WiFi – I use this with my Pi → https://www.amazon.co.uk//dp/B0062H5SK2

That’s an inelegant solution. I know about wireless bridges but they’re not as stable. Bridges are bodges. Besides that thing is huge.