Apps Icon displays folders in filer window.

@Paolo. I dare say that Vince H would be very willing to host a mailing list for you. Invite only.
Anyone for ’Paolo’s OS’ as a name?

Regards.

@ Ron

@Paolo. I dare say that Vince H would be very willing to host a mailing list for you. Invite only.

I perceive some hostility in this remark, maybe just a different cultural background.

Anyone for ’Paolo’s OS’ as a name?

I am sure you do not mean it in the way I perceive this Ron, but in my culture this sounds like mockery. In the remote case it was intended as such, please re-read the madness shared in this thread, before attempting any mockery. Also, if this is an attempt as such, is another confirmation of my thoughts.

And just to be clear, the Apache 2.0 allows rebranding/renaming as well as:

You can freely use, modify, distribute and sell a software licensed under the Apache License without worrying about the use of software: personal, internal or commercial.

This license explicitly grants rights to users that can be applied to both copyrights and patents, unlike other permissive licenses that are applicable only to copyrights and not patents. The rights given are perpetual, worldwide, irrevocable, but also non-exclusive — so you can use the licensed work, and so can anyone else.

Actually is quite clear that the derivative product SHOULD be rebranded because the Apache 2.0 license does NOT grant the right to use the original brands and names ;)

read this please: http://www.apache.org/licenses/LICENSE-2.0#trademarks):

So, again in the remote case it was an attempt to mockery, it ended up quite badly, and yes probably third party distros, unless allowed by ROD/ROOL should not use RISC OS as a name…

Regards.

Hi Paolo,

Obviously Ron can speak for himself but I take it as some advice of where to go if required and nothing more especially given his other positive comments further up the thread.

Anyway as you said yourself you want to do something that you enjoy so as Ron said earlier as well, along with others, just go for it and we look forward to seeing the results and those that don’t want it don’t miss out on anything.

@ Doug

Anyway as you said yourself you wnat to do something that you enjoy so as Ron said earlier as well, along with others, just go for it and we look forward to seeing the results and those that don’t want it don’t miss out on anything.

yup, that was my early conclusion too. so yup that will be the way, thanks :)

Just be grateful that someone is willing to put in some real effort to possibly improve our lot. If you don’t like the result, you are free to not use it. Others may have different opinions.

maybe just a different cultural background.

Ah, yes. I had to explain to an Aussie once that when I said “awfully nice”, I wasn’t insulting her. It’s a British thing.
Was it the “dare say” that you felt was hostile? If so, it basically means “I would imagine this to be the case so I’m going to say it even though I’m sort of volunteering somebody else”.
As you can see, “dare say” is a lot less of a mouthful than all of that.

should not use RISC OS as a name…

It’s not RISC OS™ or RISC OS®, so is it actually a trademark?

that do not like the idea of bring the additional background tasks into the picture.

I think this depends upon what the background tasks in question actually are.

these things are disliked by many that use those OSes

[Citation needed]

system calls to DR-DOS Background Tasks, or those for GEM

Hands up anybody who isn’t DavidS who knows what the hell this even means.

like indexing the filesystem in the background

When would you prefer this to happen? While you sit twiddling your thumbs as a little on screen graphic crawls from one side to the other?
The problem is, again, one of implementation. It’s perfectly fine to index stuff in the background, but it needs to be extremely sensitive to user operations and be willing to throw away partial data (to do it later). The reason why file indexing is annoying on certain systems is because it has the attitude of “I’m almost done, wait a moment”.
No, wrong, shut up and get out of the way.

It’s the same thing with Android that when you restart the device, connectivity doesn’t really work for several minutes as all of the background tasks are given precedence. This is unbelievably broken. What the user is looking at and doing is the most important thing and everything else can fit in the quiet periods (I’m looking at YOU, sodding Google Play Services).

better centralizing the configuration would be a good thing

Uh, did you read his detailed explanation of why that’s just a naff idea?

actually it was just a pretty distraction, giving the impression of something useful. There is a lot of this about in ‘modern systems’. ;-)

Did I mention MS improving things, and then losing the plot?

In fact IIRC RISC OS 6 (I know a dirty word for some)

Confusion labelling, just RO4.3x with a different label (I think VW do (did?) something similar)

Anyone for ’Paolo’s OS’ as a name?

I am sure you do not mean it in the way I perceive this Ron, but in my culture this sounds like mockery.

The joy of the quirks of colloquial English. As Rick pointed out, the misunderstandings of the phrases are endless.
It works inbound too: Chester Thompson said “Man, that was bad” in an early session with Genesis “bad” in that American slang context meaning “really good”

If you’re looking for a name tag, I quite like “PRO”

To repeat myself though, please use Apache. It enables a clear divide between the core of RO and any GPL code.
Andrew R would probably point out that a lot of the interested people he’s talking to about RO are interested because the OS is NOT GPL.

When you have people suggesting things like:

Actually, we could just argue that the ROM is a mere distribution medium, and take advantage of GPL’s ill-defined “mere aggregation”, “volume of a storage” and “distribution medium”. Notice how they failed to make those things any clearer in GPL V3.

you know you’re creating a shade of grey that suits the nastier legal types. A clear gap between GPL and the core OS holds the reptiles at a distance.

Ah, yes. I had to explain to an Aussie once that when I said “awfully nice”,

and

The joy of the quirks of colloquial English. As Rick pointed out, the misunderstandings of the phrases are endless.

Ha, thanks guys and so my apologies to Ron for have misunderstood his words, sorry Ron!

@ Rick

It’s not RISC OS™ or RISC OS®, so is it actually a trademark?

I am totally confused on what RISC OS is, probably best to ask ROD/ROOL, because in the recent past I have found this:

http://www.riscos.com/news/news_items/PR201001.htm

From 2001 which, at the bottom of the page, states that RISC OS is a TRADEMARK of Peace, and if so it should have been inherit from Castle and so now (hopefully) should be either ROOL or ROD. But I have no idea.

So, I was about to send an email to ROOL or Andrew to ask if it’s ok to use “RISC OS Community” on github or if I should change it (still with something that is for everyone to enjoy/use of course).

However if that page is correct then indeed RISC OS is a trademark, not just a copyright.

system calls to DR-DOS Background Tasks, or those for GEM

Hands up anybody who isn’t DavidS who knows what the hell this even means.

That, I think, is typical DavidS style info-cocktail (apologies David if I used that term and may not sound funny to you).

DR-DOS traditionally was based on a kernel called BDOS, which introduced the concept of SysCalls from CP/M-68K (yes BDOS was also the kernel for CP/M). However early BDOS for x86 did not support the concept of SysCalls (that is because both the 8080 and the 8086 only had real mode, hence no CPU privileged mode, basically like we had on our beloved Acorn MOS on the BBC Micro and the BBC Master). Subsequently (from CP/M-68K) BDOS introduced the concept of SysCalls (we had them introduced with the Archie, before we had on every micro the Monitor Subroutine Calls or Service Routine Calls, basically ROM based routines we called via CALL, EXEC or the usual JMP depending on the micro, BASIC or assembly we had).

Long story short, at some point DR-DOS also got the BDOS Kernel with SysCall support, but to use that you needed to have at least a 286 and use protected mode. And from DR-DOS 5 or 6 it also got GEM UI which made use of some of the BDOS SysCalls.

There wasn’t a true concept of “background” processes (please note the term “processes”, not processing) in DR-DOS. They were more advanced TSR (the regular MS-DOS TSRs to be clear). DR-DOS was a stripped down version of Concurrent DOS (or CDOS) which indeed had multi-tasking and multi-user support and CDOS indeed had background processes.

Now, what all this has to do with my notification framework, I have no idea, maybe David had his own concepts-connection moment and may explain.

The notification framework, first of all is configurable by a user, so NO it doesn’t work like Windows or (older)macOS releases. A user is totally in charge of configuring it to ignore, route or display a notification.

Basically it uses INI files per each notification type to know what a user wants to do with them, so if a developer makes an app too “pingy” the user can always configure the notification framework to either ignore the notifications, or part of them etc…

My explanation of background processing (please note the “ing” and not “processes”) in RISC OS, is mostly due the nature of a cooperative multi-tasking scheduler and how events works in RISC OS. Most our apps do is indeed background processing, unless what a WIMP app has to do requires a reasonably amount of time that can be used within a user event like a mouse click on an icon etc.

Using a NULL event, for example, allows even asynchronous programming in BBC BASIC (with exception of I/O operations of course), which is impressive if one thinks about it for the time it was released and the combination of NULL events with Wimp_PollIdle is event more impressive (from a coding paradigm point of view) because it allows a developer to control the latency between occurrences of an asynchronous event. Sure these days this is just nothing compared to the Async API in Rust or Go, but guys we had this stuff back in the 80s and in BASIC! Sorry moment of passion for the old cool stuff in RISC OS :)

Anyway, sorry for the digression, traditionally, a background process is a process that do not have access to the framebuffer, the video terminal or the UI/Console. It can process data interacting over an API, IPC, Network traffic, Timers, IRQs or Events and generally it either outputs actions or messages (that can be received and/or stored in various ways). You can see why I used the analogy with background process for the notification framework and indeed pretty much ALL runnable RISC OS Modules are by definition background processes, although if a module is not per se a process. This is all blah blah we all know, but in case people are on different pages it may help repeating it.

So, I do not think RISC OS is not using background processing, it always has, just not in the canonical form Microsoft/UNIX uses for “Background Processes”, but indeed background processing (note the “ing”) is definitely happening in RISC OS on both WIMP Apps and runnable modules. So I hope the example is more clear for everyone now.

I have never said “RISC OS calls are background tasks”, not sure where you have read it. JFYI…

@ David

That is not saying that “RISC OS calls are background tasks”. Although it may be used in common English I think.

That simply means that the way one handles “long computations” (that for example requires more than 30ns/80ns to complete) as chunks of work (aka pieces, aka split the workload in multiple steps) and executes such steps using the NULL event from either Wimp_Poll or Wimp_PollIdle.

This is almost the original definition of “batch processing” done in the very old time-sharing systems back in the 60s. However, given that such flows are either in a queue (in certain cases) or a single request happened in the past, they do correspond to “background processing” (please note the “ing” and not processes).

The issue I have to use the word Task or Process (or Processes) is simply that such words (in most Software Engineering literature) identify a well defined and confined entity aka a Process (an instance of an program at run time and its resources), while in this particular case the above are literally just “asynchronous flows”, or “unconfined threads” if you want, that belong to a process which is the WIMP application task or that in the entire Kernel if we refer to a runnable module.

Does that helps?

However the comments were not just about hidden files, for example Chris Mahoney expressed dislike for background processes.

Did I? I said RISC OS isn’t full of them (thinking of the type that drag down the system with ‘unnecessary’ processing, although I admit that this can sometimes be a fine line), but I don’t believe that I said I don’t like them. Indeed I’ve written my own!

Because the courts may not decide the way you intended the license to be interpreted.

This. If the licence doesn’t clearly say, it isn’t fit for purpose.

American slang context meaning “really good”

Between the two countries, if a film bombs at the box office, it’s an entirely different interpretation. One side of the Atlantic means “did amazingly” and the other side means “sucked so hard we’re millions out of pocket”.
Oh, and the current modern usage of “drop” to mean releasing something. As in Netflix will drop The Witcher in December…

Anyway, sorry for the digression, traditionally, a background process is a process that do not have access to the framebuffer, the video terminal or the UI/Console.

I run two Wimp tasks that are exactly that.

The first is a little program I wrote that fetches the current weather conditions every five minutes. As the serial protocol is slow (4800 baud) it breaks it into parts and does a little bit each Wimp poll.

The second is WebJames.

Together, they make http://heyrick.ddns.net/ (and it was fridge-cold this morning, now we’re heading into the arse end of the year :( ).

RISC OS Modules are by definition background processes

It’s a little more complicated given that RISC OS is pretty much a single process system (you can cheat and hook stuff to events and timers, but you could do that on a BBC Micro too). So where does one draw the line if the foreground application makes a SWI call to a module to do something on its behalf? Sure, some things (most Wimp calls, SpriteOp, etc) have direct on-screen results, but there are plenty of others (filesystem activity) that are necessary too.

However if that page is correct then indeed RISC OS is a trademark

Hmm.
https://trademarks.ipo.gov.uk/ipo-tmtext

Found nothing for “RISC OS” or “RISCOS”. It only matched a similar sounding (but unrelated) “RISKOS”.
That’s not saying it wasn’t ever a trademark, but the registry has no account of it, and there’s never been any ® or ™ after the name, as would be common if it were the case.

Why index file contents at all?

Faster searches, perhaps?

to run the command when they are away from the computer for hours at a time.

FFS, this isn’t 1987. I think it’s entirely reasonable these days to have the machine do some housekeeping in the background when it appears “idle”.

And no, not on a timer to do in the middle of the night. My big box is powered up for use and turned off afterwards. There’s quite a saving in power by not running it all the time. That’s why a phone, a tablet and my Pi are my “daily drivers”. Minimal power consumption.

You do not search for files by content

You can, and I have, if the feature is available. It’s also useful if you know you wrote something but don’t remember what file format you used at the time…

And apears to be the view of many users of systems that do it in the background,

[Citation needed!]

Almost as bad as having a background task check for updates for software or the OS components, something that traditionaly was left to the user to initiate,

Because the user often didn’t bother to initiate. Plus it’s a pain in the arse if updates are infrequent to keep on looking “just in case”. It’s something that can be quickly done “in the background” to notify the user.
It’s about not wasting user time. Start up the browser, go to your website (assuming they remembered to bookmark it), compare the version that’s available with the one they have…
…or a little notification pops up as if by magic when there’s an update, and they don’t have to check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and check and do you see where I’m going with this?

And how many people get any advantage from file content indexing?

Sounds like the mentality of the software authors responsible for a database used by one of the specialities. Works fine when dealing with a small/moderate number of records. Then they added the MPI¹ feed that was part of the contracted setup.
Sorta struggled when fed all the changes coming down from the National Spine MPI

You do not search for files by content, you search by filename and attributes. Eg in Widows to search for the file PROGMAN.EXE (just an example),

Actually, I tend to search using a utility called Agent Ransack and I search some (or all) files with a regex (or specific) filename for a regex string (or specific content)
I’m sure that kind of search is not unusual² and indexed files are quicker to search

¹ Master Patient Index, millions of changes per day.

@Paolo. There is no need to apologise and I meant no offence.‘Dare say’ was meant as Rick said ’Paolo’s OS’ was to differentiate between variations e.g:- ROD, RComp, ROOL and so on. And as for ‘Invite only’ By that I meant the mailing list to be for people testing and evaluating your stuff. Not for the OS religious zealots wishing to push their points of view.

Regards.

Thx Ron :)

@ Chris Mahoney

Did I? I said RISC OS isn’t full of them (thinking of the type that drag down the system with ‘unnecessary’ processing, although I admit that this can sometimes be a fine line), but I don’t believe that I said I don’t like them. Indeed I’ve written my own!

Sorry, I must have misunderstood you then. In the mist of all the objections received, it sounded like also background processing was not desired.

bq. Hiding these files in the filer would make development directories tidier, and also help prevent content searches including the .git directory which will generate lots of unwanted hits from the object files.

I would be happy to update !Find (i.e. the underlying tool) to optionally skip such RCS dirs. It would improve my life somewhat, especially as our current out-of-date svn port pollutes every subdir with /svn.

Well that was easy! I have builds of the current DDE’s diff and find that skip /svn and /git directories – diff-nrcs and find-nrcs (-nrcs suffix = no RCS files) – which is my particular desire.

If a new command switch could be agreed upon, e.g. --exclude-vcs (as per gnu tar), that could be implemented for both the backend and frontend tools (tick option in Options menu I think). Easy enough to extend to ignore ‘hidden’ files (and dirs) if we agree on using the / prefix for them, e.g. --exclude-hidden-files.

if we agree on using the / prefix for them

Works for me, I’ll make sure to use /DirCfg and /DirIdx for my filer. I don’t think I am using hidden files for anything else right now.

OK, suitable options added to diff and find and frontend updates done too.

Interested parties can take a keek at: https://croftnuisk.co.uk/coltsoft-downloads/other/upd_df_xhv.zip

That contains diff-xhv and find-xhv that can be used in place of the ones supplied in the DDE. If you use the !Diff and !Find frontend tools, there are updated Desc files to support the two new options:

Exclude hidden (-xh on command line) = any dirs or files starting with /

Exclude VCS (-xv on command line)= version control system dirs and files (currently just /svn and /git, but I can take the full list from the gnu tar reference)

If neither option is supplied, the tools behave just as before. If you don’t update the frontend tools, you can supply -xh and/or -xv from the command line.

Very small change set – if people think it sensible, I’ll submit a pull request for them.

Still more off topic, I am minded to add -xa to exclude descending into archives / image files.

I would be happy to update !Find (i.e. the underlying tool) to optionally skip such RCS dirs. It would improve my life somewhat, especially as our current out-of-date svn port pollutes every subdir with /svn.

Not to drag this further off topic, but does anyone know where the source for the current SVN port is? I still use SVN, and I might need something to tackle once I’ve finished with SQLite :)

does anyone know where the source for the current SVN port is?

Is this the current version? If so, the source is here: https://github.com/ajw498/subversion

Nope. I have 1.6.12 (8-Mar-11), although I have no idea where it came from. If I recall correctly, there is/was another one in PackMan but it wouldn’t connect to VisualSVN Server.