Improving the the volume of bounty subscriptions

I reproduce here a posting to the Netsurf mailing list by Tim Powys-Lybbe, as he is disinclined to grapple with ‘forums’:

1. By issuing a public report occasionally (yearly? And to csaa?) on what has been achieved and what could do with more funding

There’s not a lot going on with the bounties at the moment. We do post updates to this site – primarily to this forum or the news page – when bounties are opened, claimed, completed or re-opened, etc.

One thing I would say is that while a bounty is in progress, it’s up to the claimant to decide whether they want to post updates. They might, and many do, prefer to work in private, away from pressures of community ‘feedback’ (diversionary or off-topic arguments abound). ROOL doesn’t enforce a policy for public progress updates on bounty progress.

If you want to see historically what the state of the bounty system is, it’s right there on the bounties page – every bounty with its state and how much it has raised.

2. When a Bounty has been fulfilled, write to the subscribers to tell them of this good news and invite them to subscribe to other bounties instead, even including with this a list of priorities for RISC OS.

Again, this is already put onto our site. It seems a bit pointless to start emailing people with the same thing. Never mind the fact we don’t currently actively collate the email addresses of bounty donors in a way that could be used for this (without creating extra work).

As for priorities for RISC OS, there’s our roadmap wiki page and not forgetting the Wish Lists forum.

3. Consider changing some aspects of ROOL to a charity

That’s certainly worth considering but does sound like a lot more admin to a) set up and b) run. I’ve no idea if any of our activities could even qualify as charitable. But yes, worth considering.

That said, how much do you think is needed in the relevant bounty (which?) to encourage a developer to start work?

One can only guess. I’d expect most of the bigger tasks for RISC OS are of the order of man months which would equate to tens of thousands of pounds if you realistically expect people to make them happen (i.e. take time out of their job to do it – they will still have bills to pay).

That’s certainly worth considering but does sound like a lot more admin to a) set up and b) run. I’ve no idea if any of our activities could even qualify as charitable.

My wife has a fair bit of experience of working with charity setups and info on charity law is a bit specialist.
I shall describe the setup and ask her for an opinion – probably amount to a reference to a legal specialist, we know a few of those.

Road map is not so helpful.
First had to log in (which means it cannot be considered publicly accessible), and after logging in using my mobile phone, I got this response:

Access denied. Your IP address, 90.84.146.205, was found on one or more DNSBL blocking list(s).

Why not hide the edit/update links instead of blocking everything?

Road map is not so helpful.

Typo in Steve’s link – he’s linked to a wiki page which doesn’t exist, so then it tries to log you in so you can create it. Not sure why your IP is blocked though!

This is the correct roadmap link: https://www.riscosopen.org/wiki/documentation/show/RISC%20OS%20Roadmap

Apologies for the typo – I’ve fixed the original post.

info on charity law is a bit specialist

Nice guidance on the charity commission web site it seems, and a quick check proved this to be true here

The interesting bits are
1 – the details are likely to change soon to reduce the eligible list.
2 – The line that says “The Charities Act says that a ‘charity’ is an institution which is established for charitable purposes only (see part 2 of this guide)” probably rules out ROOL as the ROOL pages clearly describe the management of the source as just one of the activities with others including commissioned work on ARM based development. The whole organisation needs to be for charitable purposes only.

The whole organisation needs to be for charitable purposes only

Further discussions here at home:
1. In the pdf referenced section 3 f is the bit to aim at, but as stated before the organisation with the charity registration must be set up with all activities for charitable purposes.

2. “If the aim is simply to access additional funds they are looking through the wrong end of the telescope, if you turn it around and look at protecting the future of RISC OS then you are on the right track” quote from C.A.P
see 3 f “the advancement of the arts, culture, heritage or science”

Of course you’d need trustees for the “RISC OS preservation/advancement foundation” which would likely be distinct from ROOL. You will need specialist advice on setting up the charity and framing the description of its purposes. Wrong description of purposes and the registration fails. Once you have the purposes listed you have to stick to all of them.

The thing about running a charity is there’s a lot more scrutiny of the accounts and procedures than a Ltd company. Spinning out the trading side is common – call it RISC OS Open Operations Ltd (ROOOL).

I suspect it might cost more in admin than it makes in Gift Aid though. Especially since it seems time of ROOL folks is a fairly precious commodity. Perhaps more of a distraction than is necessary at this point?

The thing about running a charity is there’s a lot more scrutiny of the accounts and procedures than a Ltd company.

As I was alluding to previously, Christine was pointing out that if this a simple shot at extra funds it is looking at things from the wrong end.
If people have a stated aim of doing something charitable then that aim needs an organisation and the organisation can register as a charity if it fits the charitable purposes requirements. It isn’t a magic key to money with no effort. That is exactly what changes in charity law have been and are continuing to work on eliminating.

Looking at the roadmap:

A9Home port

;-) I’m surprised nobody has tried. Or was the market for these very small?

Add BZ2, LZMA, 7zip archive format support (sponsor SparkFS?)

Having SparkFS on board would be nice, but I worry about support for modern formats (7zip?). What are the licences? We can’t bake in anything GPLised.

Debugger
Allow setting breakpoints in ROM where the ‘ROM’ is actually soft ROM

Mmm, I’m guessing the debugger replaces the instruction at ‘x’ with a break into the debugger? Yeah, I can see how it would be ‘interesting’ to provide write access to the debugger. Maybe simplest to tweak OS_Memory to allow (parts of) the ROM to be mapped as writeable – I’m guessing the OS is the arbitrator of what/where/when here?

Add partition table support to the filing system stack

Uh-hu.

New FileCore format for huge modern discs

PLEASE allow the filesystem to attempt to recover from a map broken…

Larger RAM disc support

Botching the code to allow ~380MB didn’t seem to trigger any untoward effects.

Remove FileCore 4 floppy+4 HD limit

Is this possible? We don’t need a floppy/HD distinction any more, but I ask because I wonder how much stuff would choke on a path like SDFS::10.$

Full media stack (hey, I can wish)

Hehe. Indeed.

Improve scroll wheel support

Well, we all agree that the scroll wheel support is “lacking” but the bounty came and went…

Update from FreeBSD – IPv6 etc

Given the speed of most ISPs, I don’t think we need to be in any hurry to support IPv6. :-(

WiFi support (from BSD?)

Would be nice.

Multi-CPU support and preemptive multitasking (a simple job, clearly)

Yeah, we’ll have that “by Thursday”, right? ;-)

Unicode support more complete

With a fully UTF-8 desktop that has a fallback Latin1 mode for older apps. We’ve been over this before and there are issues regarding transfer of data between old and new, but nothing seems to be insurmountable. Just “a bit fiddly”, but that’s surely better than the current either-or situation.

Animated window and menu transitions

Mmm… iOS has lots of bells and whistles, but I like the fact that RISC OS just does it without playing a fanfare first.

Non-rectangular and transparent windows

Reason? Apart from the ‘rounded corners’ in XP’s Fisher-Price theme, and windows with odd shapes “because they can”, I see remarkably few windows in XP that aren’t rectangular. People just expect that, and to hit them with a lozenge shaped window with an ellipse hole in it, okay, it might look kind of cool the first time you see it, but afterwards you’d be inclined to think that the author is a bit of a dick…

Drag-and-drop in writeable icons

Let’s start with ^A and ^C first. In the OS, that is.

Some additions of my own:

• IIC code that doesn’t run with interrupts disabled. It’s a bit silly that writing a bit of data to an IIC device kills sound playback.
• A simple one, this: The OS (and “important” modules (FileCore, Wimp, etc)) run in SVC32 mode. The rest of the RMA and all user modules that don’t have a [to be defined] signature run in SYS32 mode to give them enough privilege to talk to some hardware without running at OS level. Hardware needs to be “allocated” by a memory request to it. Attempts to enter SVC mode will result in the module being removed from the module chain (killed with extreme prejudice).
• Smaller (half-height) iconbar, now that modern display devices make such a thing possible without looking UGLY. Might need to trap and patch the positions that iconbar menus open at…
  I have put a (poor: simple dithered rescale) mock up here: http://www.heyrick.co.uk/random/pics/codingpic.jpeg

if this a simple shot at extra funds it is looking at things from the wrong end.

Indeed.
You need to learn how to bull**** correctly.

http://www.theregister.co.uk/2014/03/25/how_did_millionaire_supermodel_lily_cole_get_200000_of_taxpayers_cash/

supermodel_lily_cole

Well off-topic:

animated clothes-horse.
Also referred to as “actress” which also seems to stretch things a little.

end.

Well off-topic:

I think my comment in the forum for a follow-up article says it all.
http://forums.theregister.co.uk/forum/containing/2217377

end.

Whaaa? You’re missing a GREAT money-making opportunity here! :-)

Non-rectangular and transparent windows

Reason? Apart from the ‘rounded corners’ in XP’s Fisher-Price theme, and windows with odd shapes “because they can”, I see remarkably few windows in XP that aren’t rectangular. People just expect that, and to hit them with a lozenge shaped window with an ellipse hole in it, okay, it might look kind of cool the first time you see it, but afterwards you’d be inclined to think that the author is a bit of a dick…

Radial menus, anyone? Never used them myself – perhaps it’s a load of pants.

Radial menus, anyone?

Seems an interesting concept, but can easily be faked without the need to use odd shaped windows. Take a snapshot of part of the screen, draw the menu, on menu activity redraw the bit of screen under the menu, and…
As making a mouse “gesture” is what performs the desired action, there is likely to be no support for moving the menu around.

Never used them myself – perhaps it’s a load of pants.

I’m inclined to think “yes, it is ‘pants’”, for the following reasons:

1, The example linked (via Wiki) uses pictures. Look at your software (any software) and ask yourself how much could easily be represented using pictures. The biggies (open, new, safe, print) certainly, but how about “leading” or “this file has DOS style line endings”. Text is easier and clearer.
2, Radial menus wouldn’t seem to lend themselves to more options than there are compass points plus intercardinals (which means eight options). That leaves out several RISC OS applications. Not to mention, menu navigation itself (go back, etc) may claim some of these option points. Not to mention, how do we handle differing numbers of menu items? Varying angles? Fixed location? The latter is more UI consistent, but the former probably looks better on-screen.
3, Support for submenus, writable menu options, dialogues off menus…? Possible, but untidy.
4, If they were that good, Android would be using them instead of the horrid my-Beeb-could-do-better menus that we have. A flicky-pokey interface and flickable-pokable menus would seem to be a perfect match. So where are they? Well, it looks like Google has recently made a patent application (http://www.androidauthority.com/google-patent-android-radial-menu-357704/) – because anything “on a mobile device” is new and innovative, right? So maybe they’ll turn up in a Samsung phone until Apple whine about how they “invented” the idea first (hint, prior art from 1969!).

This isn’t to say radial menus are bad, per se. It’s just I’d more expect them to appear in iOS8 than in RISC OS 5.30…

So maybe they’ll turn up in a Samsung phone until Apple whine about how they “invented” the idea first (hint, prior art from 1969!).

Rotary knob with push selection, built into TV’s and early semi-automatic washers (my mothers 1959 Parnall¹ automatic comes to mind) granddads TV with the circular tube was similar.

It might well be that radial menus appeared elsewhere because context sensitive menu use still isn’t a polished feature elsewhere.

If I wanted that kind of feature I’d help Dave Higton interface a TV/Video remote to the USB input. Since I don’t he can carry on doing bluetooth.

¹ Never heard of them? They went out of business because people tended not to need to buy a second – mum’s was replaced when it finally failed beyond repair in 1985. Didn’t quite make the 30 years use in 3 houses in two cities…

OT: Childhood memories of the an amazing TV (or so it seemed to a nine year old) that we could change channel remotely. IIRC the remote had a long wire and one button. Pressing it and a motor on the rotary channel select on the front of the TV whirled and it changed to the next channel round! I think BBC2 had just started (1964) so we had a whole three channels to choose from. The TV was future proofed in having FOUR possible channels.

Having SparkFS on board would be nice, but I worry about support for modern formats (7zip?). What are the licences? We can’t bake in anything GPLised.

If SparkFS became part of the standard disc image, I think there would not be problems if it contained GPL stuff – after all, that is pretty much the definition of “side by side on the same medium”, so you only need to publish it with source code – the GPL’s “viral nature” is only a problem if you link to it, not if you just distribute it.

Anyway, there are at least LGPL implementations of the afforementioned compression algorithms.

OT

² Unless you’re overseas and without a sneaky proxy workaround, of course.