Cloverleaf Campaign is Live

And Firefox today is nothing like Firefox 15 years ago, so the patches are probably useless for any project today.

Nota: I did not talk of Firefox, but of Goanna based web browsers, where it could perhaps help.
Thanks for the link!

Even so, the codebases have moved on a lot in 15 years.

A general question for Stefan, Andrew, etc. I see statements on the Kickstarter like this:

Risks and challenges
As we are using existing hardware we will not have any risk of technological matter. The only risk is that we cannot keep the planned timeline to adopt the RISC OS hardware access layer (HAL) to the hardware and develope the corresponding drivers in time as anticipated. For example the display output uses different systems on the desktop PC, the laptop or the All-in-one PC. So when it works on the desktop PC then not automatically the laptop or the All-in-one PC will work at the same time.

Can you explain your strategy in terms of achieving the goals according to the timeline? There’s a number of fairly serious pieces of technical work and to me the timeline looks relatively tight. How are you planning to staff those activities? Do you have existing developers lined up? Are you basing your work on software that already exists, in which case can you tell us what it is? Do you have examples of what’s been achieved so far? What skills do you bring to the table?

I’m a bit concerned that the Kickstarter plays down the risk angle more than I would be comfortable with. It is possible that’s because it’s been written by a non-native speaker, but some of the size of the tasks seems to me rather understated. For example:

How hard is it to compile Linux apps for RISC OS?
Simple C/C++ code without GUI can be compiled without any changes. RISC OS programs can be programm on RISC OS but also on Linux and Windows systems by the GCC compiler. When creating applications with user interface then this must be adopted to RISC OS. Another issue might be libraries that are included on open source projects if these are available on RISC OS. In RISC OS th Toolbox and TBX library is used to access the RISC OS WIMP.
http://www.stevefryatt.org.uk/risc-os/wimp-prog
http://gerph.org/riscos/ramble/toolbox.html
Last updated: Sun, November 22 2020 3:55 AM UTC +00:00

Here a list of possible open source software that can be ported to RISC OS:
kicad-pcb.org (2) (PCB layout)
audacityteam.org (digital audio worksation)
libreoffice.org (Office package)(3)
openoffice.org (Office package)(3)
rust-lang.org (programming language)(2)
gnumeric.org (spreadsheet)(1)

[trimming from the list a few of the suggestions I’m less familiar with]

Do you really have a handle on how much work this is? Software like Libreoffice is very complex, and porting GUI software needs a lot more work than command line things. And porting Rust would require porting LLVM, which is another large project.

If you look at a good Kickstarter is that the proponents are clear on identifying and addressing the risks. For example, hardware projects always have supply chain issues, contract manufacturers who aren’t as easy to deal with as you first thought. While this isn’t a hardware project, the primary risks are in terms of staffing and timeline. What is your approach in terms of addressing these risks?

A good Kickstarter in essence says ’we’ve got so far on our own and proven the concept, and now we need your help to scale’. On this one it would be good to have some evidence of the ‘got so far’ point. Maybe ChatCube is a suitable example, so tell us how you developed it and, if the Kickstarter is funded, how the experience you had with it makes you qualified to undertake a bigger project.

Kickstarter is about asking for money, yes, but it’s also about track record to know that the project will be successful. I’m seeing lots of promises on the page but not much detail on the ‘who’ and the ‘how’.

As far as an image editor goes, I’ve been using Photofiltre on Windows for years.
It seems to easy to use (I could never get on with Photodesk.
It’s not open source but maybe the author would be happy to work with?
http://photofiltre.free.fr/frames_en.htm

@Rick Murray

If Jeffrey is finishing his work on accessing and working with the other cores, why is it on your roadmap?

My mission is not only related to why my company does but also in motivating existing RISC OS programmers to join our efforts. This can be improving their programs, releasing their programs as open source so they don’t get lost to the RISC OS community or motivating them to finish their plans. Multi-threading is also 10 years cooking and would be a good time to finish it. It is an important advertisement point for RISC OS that we have multi-core support. So I have ask Jeffrey if he can finalize his code and he agreed to do it. The reason is motivation and when nobody is going to use your stuff then there is no motivation to do it.
This way I could also motivate Jason Tribbeck and Nemo to join our efforts. Jason is doing the new sound system that will lift RISC OS audio system to current demands without breaking the durrent system. Nemo is working on changes of the font manager to enable us to use UTF-8 fonts which is crutial for asian and arabian markets.

That is why it is on the roadmap of Cloverleaf as Cloverleaf not limited to my company. All people in RISC OS community have to join our efforts to achieve what I have planned for RISC OS to become respectable and well know around the globe again. Alone I cannot do it as lack of money and know-how.

If my project can achieve all that what is in the roadmap in 10 years time then I would have achieved more that was achieved in the last 20 years. But I have no plan to wait 10 years!

But if RISC OS cannot keep my company afloat then ….

Will these cores be able to accept and run code written in C?

You should know that a C Compiler turns C code into ARM code like any other compiler also does. So of course it will run on the multi-threading module.

So it’s a bad thing to want to better understand what exactly is going to be done by this project?

Of course not. But the tone matters. I totally agree that not all might be clear and not all can be descibed in detail also.
New users are not interested in such details and in the Kickstarter there is no place anyhow to do it. The text space if used up to 100%.
We need to motivate people to fund our project.

For old RISC OS user the interests and background is different. But they are not the target of the Kickstarter project as the aim is to extent the RISC OS userbase and not to milk more money out of existing RISC OS users.

@andym

I quite like the Photo tools in Xara. Now, if only Xara had some links to RISC OS in some way.

You might not know that Xara has its orgin in RISC OS as Artworks which was converted to Windows as Xara. Sadly Martin Würthner is too busy with other work now and also no financial benefits from sales. So the question is if ever there will be more improvements of Artworks….

@Theo Markettos

Can you explain your strategy in terms of achieving the goals according to the timeline?

The strategy is:
a) Get as much funding as possible by Kickstarter and Indiegogo to fund programming on RISC OS.
b) Get as many achievements as possible out of that funding
c) Get more funding by selling RISC OS hardware & sofwtare around the world
d) Selling services related to RISC OS (programming)
e) Improve RISC OS more and more so it can survive not as RETRO OS but real OS that is an alternative to other OSes as (Home PC, Iot, AI device).

@Theo Markettos

Maybe ChatCube is a suitable example?

Did you install and use it? My programmers use GCC on Linux and Toolbox and a modified TBX libraries. The question is not if someone is qualified but if someone is motivated to finish the job.

A good Kickstarter in essence says….

I think our goals go beyond any Kickstarter as any Kickstarter that I know off is just sellings one “simple” product. Explaining what is really behind all what needs to be done in RISC OS will take month and 10.000 Euro more. For what reason? WiIl we get 1 Euro more funding for it? Unlikely! You “sell” ideas or products on Kickstarter not documentation.

I’m a bit concerned that the Kickstarter plays down the risk angle

What risk you are taling about? That the funder risks 29 Euro? Or that we cannot deliver the hardware we offer?
For the hardware Andrew Rawnsley is responsible and he has enough experience with previous projects. So the risk is near zero that we have to give up because for example we cannot get the documentation what is needed to solve some problems with the adoption.

@DavidS

Why the choice of apps to port. We already have what I would consider as better for some of those: …

Right ! But… There are a number of issues with that.
a) Are these RISC OS software solutions still maintained/upgraded?
b) Is the code already open source?
c) Can they compete with the quality of that open source projects?
d) Can you motivate people who aren’t RISC OS users yet to invest in programs they don’t know off?

The software packages that I have listed are well known so if someone uses them they would like to use them also on RISC OS. No motivation speech needed.

I don’t say that I would not help to upgrade existing RISC OS apps to meet public demands like Word or Excel file support. We must find the most efficient way to get as much achievements as possible.

@fwibbler
Yes I have used Photofiltre also. I am using photo editor just for cropping photos/screenshots and brightness/Gamma control. So I am stuck with using a Paint Shop Pro from year 2000 as it is simple and effective. I have tried to use PhotoDesk like that but failed…

I am using photo editor just for cropping photos/screenshots and brightness/Gamma control.

Why not DPIngScan? For this things it is the best choice. Any times I also use the Win Version ;-)

Nemo is working on changes of the font manager to enable us to use UTF-8 fonts which is crucial for Asian and Arabian markets.

Nemo will probably state that a decent font manager and allied items is crucial for every market not just ones where a non-Latin character set is in use.

I quite like the Photo tools in Xara. Now, if only Xara had some links to RISC OS in some way.

You might not know that Xara has its origin in RISC OS as Artworks which was converted to Windows as Xara.

You recall me mentioning the subtleties of the English language earlier? I think you will find AndyM is using a slightly ironic tone there. He knows Xara is basically Artworks with tweaks and additions

For the hardware Andrew Rawnsley is responsible, and he has enough experience with previous projects.

So people could expect to see Andrew selling 14" and 17" screen laptops with full keyboard (not the cut-down ones on current kit) in the next 8-12 months?

I mention both the screen size and the keyboard because

• My eyesight means I have difficulties with small screens
• Many people have made comments about “missing keys” and that’s an aspect that is easily taken care of with a machine that is wide enough. Typical 15" screen designs in the x86 world even manage to pop in a numeric keypad in the extra width.

Yes, a small 12" screen unit is handy to carry around in the hand, but becomes limiting in use, especially when critical keys are not there (F11, F12 missing on the miniature keyboards)

@Steve Pampling
I would also need a 17" laptop but I guess 2 customers is not enough to produce like that. So that might be a reason for another crowdfunding.. But there the target is 1000+ units.
Same applies to other keyboard layout. A customized design you will not get under 1000 orders. Some companies ask for even 3000 or 10.000 units….
So without a much higher userbase not achievable…

@Steve Pampling except that the 12" ARMbook / Pinebook did have F11, F12 etc keys.

There were some odd elements, granted, but !PineTools found solutions to these, and the feedback on ARMbook and its keyboard have been very positive (the touchpad was more mixed). The main oddity (from a RISC OS perspective) on the ARMbook was the use of a US keyboard layout, which required the supplied PineTools to get the most out of. However, it worked well given the size, with very few compromises.

The good news is that the 14" is even better. It uses an ISO standard keyboard, so basically a UK layout, which is a much more natural fit for a RISC OS user. And a PineTools equivalent will naturally complement this too.

Incidentally, I have also done 4teTools this past week to complement the 4te and allow extra functionality tailored to that.

From other threads, you know there is no pre-existing 17" ARM laptop. There are scant few Intel ones these days (sadly) too. So, unless you know where we can raise the costs of producing a totally new unit (hundreds of thousands), it doesn’t help to keep asking. As I said in the other thread, I’d like one too, but none of us have that kind of money.

17" is more on the specialised end of the market, yes.

A 14" screen is an extremely common small laptop e.g. DELL E64xx series from a few years ago, the more recent 54xx series and there’s a full keyboard there (not numeric keypad, but that’s rarely used anyway)
The same applies to Lenovo models.

15" are becoming the more common model with the 17" still being regarded as the premium product, so as I said above – yes 17" is a “special” product.

I’d hardly classify the keyboard layouts they use as “customised”

except that the 12" ARMbook / Pinebook did have F11, F12 etc keys.

I should have been clearer – Raspberry Pi keyboard.

The main oddity (from a RISC OS perspective) on the ARMbook was the use of a US keyboard layout, which required the supplied PineTools to get the most out of.

US keyboard is an oddity in any OS

The good news is that the 14" is even better. It uses an ISO standard keyboard, so basically a UK layout, which is a much more natural fit for a RISC OS user. And a PineTools equivalent will naturally complement this too.

Incidentally, I have also done 4teTools this past week to complement the 4te and allow extra functionality tailored to that.

I need to pay more attention, I hadn’t noticed you were selling a 14" model.
What’s the price tag on that? I’m going to test visuals on a 14" PC laptop as that might be a worker for me.
Just looked on the ARMbook page – just mentions the ~12" (11.6?). That looks like it needs an update.

From other threads, you know there is no pre-existing 17" ARM laptop. There are scant few Intel ones these days (sadly) too.

Scant few choices or scant few in general? I know the big manufacturer stuff is er, “a bit away from cheap” (there we go, I did that without the obvious words not suitable for mixed company)

BTW the 17" reference was me misremembering the Kickstarter info – it’s 11.6 and 14 not 14 and 17.

Why not DPIngScan?

Seems to have problems with big photos.

My mission is not only related to why my company does but also in motivating existing RISC OS programmers to join our efforts.

Thank you for the explanation.

You should know that a C Compiler turns C code into ARM code like any other compiler also does.

I wasn’t referring to the type of code executed, more asking about the presence of a C library that can be utilised. Can a thread, for instance, malloc() or call printf()? Can it read in data from files? Can it call SWIs?

Seems to have problems with big photos.

I should be very interested to know what those problems are.

I should be very interested to know what those problems are.

Problem 1 – it is extremely slow at loading and displaying anything. Compare with PrivateEye.

Problem 2 – as shown here, images seem to be arbitrarily sized onto a “page”. Notice the scroll bars. That part of the image that is visible is a 1:1 view of the upper left. There is a lot of the image that isn’t visible. Why can’t DPScan size its view to the image (and worry about page sizes later, like when one might want to print something)?

That’s all for now. The first problem is an annoyance. The second is a show stopper. You can’t edit what you can’t get to. :-)

Why can’t DPScan size its view to the image (and worry about page sizes later, like when one might want to print something)?

I believe it does, in the latest version.

The second is a show stopper.

Perhaps not… :)

What version are you using? The latest is version 1.31.

Problem 1 – it is extremely slow at loading and displaying anything. Compare with PrivateEye.

They are two different beasts. PrivateEye is directly displaying the image, using standard OS calls, and does no processing. DPScan loads the image data into an RGB raster in memory, converting the format as necessary. There is a lot more work being done.

Why can’t DPScan size its view to the image

Version 1.31 has options to autozoom the image on loading to fit the screen and auto set the page size if it is too small to include the image at 100%.

I think you should always check the latest version of software to see if the problem still exists.

I should be very interested to know what those problems are.

No problem here. My only concern is that it’s working on a fixed page size and not on the photo size. Of course, all of this because it was (is) a scan tool, not a photo editing software. But it could.

Version 1.31 has options to autozoom the image on loading to fit the screen and auto set the page size if it is too small to include the image at 100%.

Damned. I should check that :)

What version are you using? The latest is version 1.31.

I was using 1.24. When I saw your message I thought maybe you’ve taken it on, checked Store, and there was 1.31 which behaves in the same way.

auto set the page size if it is too small to include the image at 100%.

Wasn’t aware of that, I just wonder if A0 will be big enough. ;-)

loads the image data into an RGB raster in memory, converting the format as necessary.

Fair enough, that makes sense. Until I tell you that it took nearly a minute to redraw that bit of picture that you see. Yeah, the redraw took longer than the loading. [on a Pi2 at 900MHz]
Perhaps it is taking time to work out which of the other 90% of the picture it should chuck away? ;-)

There is a lot more work being done.

Hmm, decoding JPEG on the fly and drawing it, scaled, into a window,

versus picking up raster data from an already converted image, bashing that (what is it, RGB?) into pixels, throwing them at the screen. Seems to me like DPScan ought to be the quicker of the two.

Thanks for the suggestion. I’ll need to go rummage around the settings.

Until I tell you that it took nearly a minute to redraw that bit of picture that you see.

Are you sure you have the RAM usage set up correctly? On a modern machine you can set it to say 512000KB so everything, including keeping a couple of undos, is in RAM. If your max RAM is set to a small value then it may be having to use virtual memory, which means means stuff is being swopped between RAM and disc regularly, (even during a redraw since not all the image is in RAM). On a Pi SD disc is not exactly spritely.

That is interesting because I found DPscan to be much faster than (say) ChangeFSI, even on legacy systems, for both JPEG and PNG images (especially the latter). I have used it a lot recently. I need to check out that auto-scale feature though, as that’s really the only bugbear I have and it sounds like it has been sorted. Yay!

@steve Pampling the 14" unit is currently offered as part of Cloverleaf’s kickstarter. However, it is a project that we’ve been working towards in the background for a while. It isn’t yet complete (by any stretch) which is why it isn’t on the website or anything. It’s a faster SoC and a more expensive chassis (aluminium unibody type thingy). It is rather sexy, although still no MacBook Air.

It isn’t yet complete (by any stretch) which is why it isn’t on the website or anything.

Ah, patience young¹ man, patience.

It’s a faster SoC and a more expensive chassis (aluminium unibody type thingy).

Bonus.

It is rather sexy, although still no MacBook Air.

If I was into Apple products I’d have an iPad from work as well as the iPhone – always passed on that offer. If the kit is going to incompatible with lots of things² then I’d prefer a fun OS.

¹ 18 and a bit.

@ Andrew Rawnsley and Steve Pampling

@steve Pampling the 14" unit is currently offered as part of Cloverleaf’s kickstarter. However, it is a project that we’ve been working towards in the background for a while. It isn’t yet complete (by any stretch) which is why it isn’t on the website or anything. It’s a faster SoC and a more expensive chassis (aluminium unibody type thingy). It is rather sexy, although still no MacBook Air.

Absolutely, the laptop IS super slick and lovely matte black colour:

• it also support NVMe disks as well as the usual eMMC and SD.
• It’s a 6 cores laptop and the screen is good.
• The touchpad is way batter than the ARMBook and with Linux it works nicely, however with RISC OS desktop it’ll be super snappy
• with a 14" screen working on coding on RISC OS or writing emails etc… is way better (IMHO) than on a smaller screen.

The laptop also comes with:

• 1 USB 3, 1 USB-C and 1 USB 2 ports as well as a psu port for recharging it.
• Battery life is way better than the ARMBook
• the performances are better than the ARMBook
• memory is 4 GB RAM
• comes with Stereo Microphone (yes ideal for video conferencing when ChatCube will add support for that)
• Camera is connected to the USB port so it’s easier to make a driver for it than the RPi for example
• It comes with LEDs for power, caps lock and num lock

…AND THERE IS ONE MORE THING ;)

They keyboard layout is very very similar to an Apple MacBook and it a good keyboard for the price with a nice travel

Honestly it’s a really good step forward for a RISC OS laptop, I use it already with Linux and it’s a very nice laptop, stays cool and it’s light.

Absolutely, the laptop IS super slick and lovely matte black colour:

Nice

it also support NVMe disks as well as the usual eMMC and SD.

Fast…

It’s a 6 cores laptop and the screen is good.

That should be interesting with a decent multicore use. There’s also the simultaneous RO & Linux possibility.

The touchpad is way batter than the ARMBook and with Linux it works nicely, however with RISC OS desktop it’ll be super snappy

I’ve never been a fan of the touchpad stuff, but I suppose the damp finger stutter is the worlds way of telling me to go out in the sunshine. “Batter” you say? No, I will be gentle ;)

with a 14" screen working on coding on RISC OS or writing emails etc… is way better (IMHO) than on a smaller screen.

Been testing with a 14" laptop (right now) and it’s OK (wouldn’t, and probably couldn’t go smaller)

The laptop also comes with:
1 USB 3, 1 USB-C and 1 USB 2 ports as well as a PSU port for recharging it.

USB charging? I could plug that into the sofa

They keyboard layout is very very similar to an Apple MacBook and it a good keyboard for the price with a nice travel

I presume you’re working with an ANSI layout¹, whereas I’m used to a UK laptop layout.

`1234567890-=
!"£$%^&*()_+
qwertyuiop[]
QWERTYUIOP{}
asdfghjkl;'#
ASDFGHJKL:@~
\zxcvbnm,./
|ZXCVBNM<>?

Honestly it’s a really good step forward for a RISC OS laptop, I use it already with Linux and it’s a very nice laptop, stays cool and it’s light.

The temperature probably goes with the battery life – less battery energy dumped into the environment.

¹ Although Apple do tend to break any standard just for the lulz

USB charging?

Just checked what I assume is the base machine (Pinebook Pro) which is a barrel connector for the power as well as a USB C (data/power/video out)

I presume you’re working with an ANSI layout1, whereas I’m used to a UK laptop layout.

yup, sorry, too many year of American keyboards everywhere. I actually use US keyboards also for all my RISC OS systems except the Archimedes which comes with it’s own keyboard obviously.

USB charging?

Just checked what I assume is the base machine (Pinebook Pro) which is a barrel connector for the power as well as a USB C (data/power/video out)

Correct on all counts :)

Been testing with a 14" laptop (right now) and it’s OK (wouldn’t, and probably couldn’t go smaller)

Same for me

And yes really can’t wait to have RISC OS running on it! :)