Is RISC OS a 1980s OS?

I missed Daryl saying

the lack of virtual memory

Except there is (was) virtual memory if you wanted it. I made heavy use of !Virtualise on the RiscPC. Superfluous under emulation or if you have lots of physical RAM, but in the old days it was the difference between being able to do the job, and not.

A superb bit of work by the great Alexander Thoukydides, no longer of this parish (though still in Cambridge I believe), it was released soon after the RiscPC.

Rick accused

archiving a large file using SparkFS on a 900MHz core takes about thirty odd seconds. My 2.4GHz P4 box? Takes about half a second.

Which may be true, but on RO you can navigate about inside a zip file at the command line, creating files, deleting others… try doing that on Windows.

Daryl had complained

lack of modern development tools

Indeed. Where is the RO equivalent of Visual Studio. Hell, something equivalent to Visual Studio 6 would be brilliant, you wouldn’t need IntelliSense as it now is.

lack of applications

Lack of serious development is extremely obvious. We are now in “managed decline” mode, it seems to me. And that is simply down to lack of demand.

Clive reminded me

keyboard drivers

Oh god, I said I’d do something…

Interesting reply, Clive. I was just a bit intrigued by you using LibreOffice and “layout” in the same sentence. I don’t mean this as an insult to LibreOffice, but I always put it (mentally) in the same box as Word in that layout is far from its forte. Word may be many things (its basic WP/styles/tables etc. are OK and its collabrative tools are something RISC OS sorely lacks) but it has never been well suited to layout. LibreOffice always seemed to follow the Word model fairly slavishly (I followed it since StarOffice days).

LibreOffice Draw is a pretty nice addition to the suite that does offer layout, and is sorely missed from MS Office, but equally it doesn’t really integrate except maybe with Impress (the powerpoint bit).

I’m really surprised you don’t do layout tasks with RISC OS apps because they’re really quite good at that, and blend nicely with !Draw.

The sort of layout issues that Clive will be thinking of is such things as gutter (as opposed to margin), the ability to change running headings per chapter, with different header/footer on the left and right pages (so page number is always on the outside, etc).

Google Docs doesn’t handle gutter (yet?). It is not a particularly mature word processor as yet, and extremely peculiar in that the app is less capable than the web version. However it does offer sections which can have different header/footer as well as different left/right headers. It is a bit of a PITA to set up, requiring a mix of app and website (well, the website can do it all but the app is faster). But I’ve got it working, I hope.

The problem with the likes of Impression and Ovation(Pro) are that they can do this sort of thing, but you run into problems when it comes to creating a book. For example, it is easy enough to print to PDF in order to have a print-ready document, but what about exporting as EPUB in order to create electronic formatting for devices like the Kindle? Trust me, PDFs on ebooks is painful.

Docs can output EPUB. It gets it wrong (outputting something that renders a full page at a time) but that ought to be fixable by bodging the CSS so it renders at a regular readable size. I’ve not looked too hard at this part yet. Failing that, messing with BASIC ought to get me a filter to strip it down and rebuild it correctly. ;-)

Yes, Docs stores its data “in the cloud”. It also has the ability to cache documents locally, useful if you’re going to be out of connectivity. It stores on Drive so you can download, and it exports a variety of formats including DOCX. I think it can import DOCX as well.
I like it, even though it is simple and sometimes frustrating, because it can be accessed from anything with a decent browser. I could edit a document on anybody’s machine, and install nothing. I can also edit on my phone or tablet. Actually, I’ve used my tablet a lot. It’s pretty straightforward. Is there even an Android version of LibreOffice?

Rick has my layout needs exactly: all I want to lay out these days are pretty standard books. Following the Word model slavishly, but doing it better and for free, suits my needs perfectly.

For more complicated layouts, in my life as a technical author, I used Framemaker because that’s what ARM use, and it’s very good – but not compatible with anything else other than by producing pdfs. Before that, for academic journals, I used Impression Publisher on RISCOS – absolutely brilliant, but only if you’re producing camera-ready copy, which we were in those days.

Google Docs sounds interesting, but probably not for me. Apart from occasional entering/editing of plain text, I’ve no desire to work on anything that doesn’t have a nice big, high-res screen – which really means my own Mac at home.

I love !Draw, and use it a lot – but convert the Draw files to SVG for export to LibreOffice for my books, or to a text editor on the Mac for web pages.

OK, I’ve never even considered (or even used, really) ePub. But then, I’ve never really appreciated the joy of a Kindle, although others swear blind that they’re fantastic. For me, Imp or OvPro to PDF is perfect really. Easi/Techwriter improve on that by adding PDF-structure or Word export depending on what’s needed. However, being honest, I find Easi/TechWriter far to Word-y for comfort, and prefer the flexibility of Impression/Ovation.

I suspect I’d also have loved TextEase (click anywhere and just start typing), if only they’d put a less school-oriented icon scheme/UI on it!

TextEase knocked my socks off. All the other apps mentioned do the obvious thing in the obvious way, and could have been built by any programmer. But TextEase has such a non-programmer UI design – ergonomic and intuitive, rather than structural or algorithmic – that it really did impress me.

(I only tend to be impressed by things I couldn’t have done – it’s a character flaw.)

It’s a character flaw I share. It makes me very easily impressed.

I only tend to be impressed by things I couldn’t have done – it’s a character flaw.

Nah, it’s just a fact. Your mind says ‘this is easy’ and therefore expects everyone to be able to do the same.¹ When you don’t find it easy and others do that’s interesting.
The flaw element comes if you look down on others for not being able to do what you can and I don’t think I’ve seen any evidence of that.

¹ I’m told I do that, but then I’ve seen similar in others hence my contention that it’s normal.

Your mind says ‘this is easy’ and therefore expects everyone to be able to do the same.

Half of the Dunning-Kruger effect there.

Smart people tend to underestimate their abilities… What do you mean you can’t log into this website? You just need to click here, put your name here, your password there, click that, turn on your phone and tap here to say it’s really you, then click on this and this and this and finally this. It’s so easy!

:-)

The converse is that stupid people tend to overestimate their abilities. What do you mean I can’t shut down the government just because they’re going to disagree with me? It’s also the basis of the Peter Principle where everybody in management gets promoted to their level of incompetence. And I’m quite sure that everybody here is capable of pointing to a total dunce that is in a position where they are not only useless, but actively damaging to the company and/or the people they work with.

I made heavy use of !Virtualise on the RiscPC.

I, on the other hand, made none. Because I never had anything that big to work with.

It does raise an interesting question – many parts of the Wimp appear to support “{SWAPPING}” in some manner. Does this actually work, if enabled? If not, did it ever work?

but on RO you can navigate about inside a zip file at the command line, creating files, deleting others… try doing that on Windows.

Unfair. Windows is stupid. Can you even access non-Microsoft filesystems yet?

Conversely, I’d imagine that Linux would be able to “mount” an archive? (if not, why not?)

Lack of serious development is extremely obvious.

Yes. There’s been no new important application for a while. Perhaps our last was NetSurf? [I’m not counting things ported from other platforms]

That is quite possibly down to the fact that one can’t expect to write a DTP package, vector editor, or whatever and:

• get anything like enough sales to justify the time unless the price is in the three-digit range
  while:
• having to compete with the problem that something likely better exists on another platform, and it is free
  while also:
• having to engage in a number of modern-era problems such as the handling of files from other systems (which may or may not be (correctly) documented).

It’s a double-edged sword. It is nice that I can download and install an entire OS and office suite “for free”, but I can’t help but think that this has damaged the creation of comparable software on minority platforms.

OK, I’ve never even considered (or even used, really) ePub. But then, I’ve never really appreciated the joy of a Kindle, although others swear blind that they’re fantastic.

I don’t like Kindle either – I just mentioned it as it is an ebook reader that everybody is familiar with.

EPUB (which annoyingly is not supported by Kindle) is an open “book” format based upon a simplified version of XHTML.

For me, Imp or OvPro to PDF is perfect really. Easi/Techwriter improve on that by adding PDF-structure or Word export depending on what’s needed.

The difference is that you are laying out an actual document, which is designed to be rendered page by page.
EPUB, on the other hand, defines the content. The actual words, with some amount of formatting for titles and such. It is then up to the reader (Kindle, whatever) to lay out the text for easy reading with user-specified margins, line spacing, and text size.

Also, with the use of e-ink (a sort of electronic etch-a-sketch), it is perfectly readable in sunlight (unlike most video screens), it can also be read in dark if there’s a backlight (mine don’t have that), and it also keeps displaying what was last drawn even in the absence of power. Conversely, ebook readers tend to be powered by some sort of ARM controller that only wakes up when something happens, and only draws the screen as necessary. It can run for weeks/months on a single charge (depends upon use, obviously). It isn’t like a mobile or laptop, you can read a normal sized book (entirely) and the battery will only go down by 5-10%.

I made a quick little video for you to show it in action: https://youtu.be/A4GggSgnKnQ

You may note ghosting and screen flicker. That is because the correct way to draw an eink is to flash the entire display dark, then light, then draw what needs drawn. Which takes power. So a cheap-and-dirty way is to blank out the previous content and redraw the new (which leads to ghosting). As such, my reader is set to trade-off ghosting with battery consumption, and it’ll perform a full redraw every 3 (or is it 5?) pages. It’s user configurable – always/3/5/10.

It is nice, on a warm day, to sit outside with a big mug of tea (or soft drink if it is really hot, like this summer!) and quietly read something interesting.

The problem with Kindle is the same as the Sky Digibox. It is a device made as cheaply as possible (expected profit recouped through sales) which is locked to a specific service. I tend to download Kindle books on my old S5 phone, because I have ripped the encryption keys into Calibre. That can load the Kindle file and translate it to epub for reading on my reader.
I would consider buying a Kindle, if it could handle epub so could work with third party media. But it can’t, so I won’t.

The reader I have was being flogged off in a sale at the local supermarket a few years back. I think this cost me something like 40 euros. It is basic, but it works. And the touchscreen is quite nice. :-)

and prefer the flexibility of Impression/Ovation.

That’s the primary difference between Word Processor and DTP. In a WP, you write text, it gets laid out with styles, drop in some pictures, and so on.
In DTP, you have a lot more control over the layout, with such things as “I want a box right here with some text in it, that is upside down”. Normal WP software will be “uhhhh… what?” whereas DTP will be “yeah, sure, you want a border too? frame colour? how about text run-around so you can drop this right in the middle of some text?”.

The other barrier to new software on RISC OS is a general lethergy in the market. As David Pilling demonstrated with OvPro, you can produce something that is demonstrably better than the alternatives, and people will still prefer the product they know. I know I’m guilty of that.

This is made much worse by the fact that much of the userbase is above retirement age, and has no real desire to learn anything new. Indeed, that’s often the primary appeal – the ability to use the old favourite RISC OS programs rather than having to learn new ones.

This familiarity often makes us extremely productive with our tools – I can dream ways to use Impression or OvPro + lots of little utilities to create stuff that, whilst eminently possible in other packages, would require me to figure a lot of stuff out before I could even begin to dream.

I suppose it is why a lot of writers prefer their pens and type writers to “obviously more capable” computers. If you know your tools so intimately that you don’t need to think about the tool to use it, there’s nothing between you and your creativity.

and people will still prefer the product they know

I guess we all prefer familiarity.
Exactly how many times has Fred tried to tempt me away from Zap? I think I’ve lost count.

and has no real desire to learn anything new

I don’t really have desire unless I can justify it. It’s part of why I know nothing about Linux. When I first saw it, it was a dreadful crashy mess. Well, it was circa 1996 or so and it only worked with a few video cards (not mine!). It’s also why I never bothered looking into writing apps for Android. I don’t have any get-rich-quick delusions, nor do I plan to spend loads of time developing something that will get lost in the morass of dirge and half-finished “good ideas” (pick any app store or repo, they’re there – Ubuntu and Android both suffer horribly from this).
I’ve also stopped caring about Windows post-XP. My box (XP) does what I want. The end.

This familiarity often makes us extremely productive with our tools

It’s also because the tools are pretty good. While OvationPro under RISC OS suffers from the “RestOfTheWorldDoesn’tExist” syndrome and the lack of modern printer support, it does work quite nicely under Windows. And, again, familiarity means that you can do all the tricks you know from RISC OS. But, when I want something done quickly, it’s usually simpler to fire up OvationPro than anything else. Want four photos in a grid printed to a piece of A4 photo paper? It’s a couple of minutes in OvationPro. Using Google Docs? Can it do it? I don’t know. I got two pictures, but there’s no fine control, only dragging with a finger, so I gave up as the pictures were slightly different sizes (unacceptable). OvationPro made it easy.

+ lots of little utilities to create stuff that, whilst eminently possible in other packages, would require me to figure a lot of stuff out before I could even begin to dream.

That’s a very good point. Traditionally the other platforms haven’t been strong at useful script languages. A common one in the Linux world is perl, Windows has (somewhere) VBScript. Now it looks like both might be supplanted by Python. From time to time devices may have a small Lua runtime.
But as you say, it means learning a new language, the API, and figuring out how to do stuff.

Since forever, RISC OS has had BASIC fully integrated. It’s not a particularly powerful language, it’s heritage is the late ’70s and early ’80s, it has zero concept of modern paradigms. But it is simple, always available, and powerful enough to write full applications so can do simple stuff like filtering/converting files too.

a lot of writers prefer their pens and type writers to “obviously more capable” computers.

Some writers like to go off on retreat to the back of beyond to concentrate on their work. Pens and typewriters might be old-fashioned, but they require no power, they don’t crash, and they are dirt cheap to run. The unicorn-tears ruse never made it to typewriter ink ribbons.
There is also something satisfying about having a hunk of metal whack a character onto a sheet of paper. There’s a certain permenance that transcends the ephemeral nature of computer data.

That said, I think it’s only the Big Name Authors that can get away with that sort of thing. Because somebody somewhere is going to have to take the manuscript and turn it into a digital form (for layout, printing, ebooks, etc). Newbie and lower volume authors will just be putting an obstacle between themselves and any potential publisher if they try handing over a couple of hundred pages of typed (or written!) text.

there’s nothing between you and your creativity.

When you get “in the zone”, it doesn’t really matter what you use. I don’t know about you, but when I’m writing I only need the keyboard and, sometimes, a keypress for bold or italic. All other styles (headings and such) are written as brief notes in the text in square brackets. Any more than that would break concentration as I’d be paying attention to the formatting and not the story. The story must come first, as it’s pretty much the entire point. Any complex formatting can happen during revision.

I have often wondered how a RISC OS user would do in a contest between systems producing (Risc OS ,windows or Mac) a document or a vector graphic etc without the use of wizards, who would come out with better end product in the shortest time.
Look at rick go from an idea to a working prototype in a couple of hours.

having to compete with the problem that something likely better exists on another platform, and it is free

This is an alarming problem regardless of platform. In the future everything is the browser, and everyone runs the same four programs.

Windows has (somewhere) VBScript.

And some Windows programs have Visual Basic for Applications. I did some serious work in VBA for Excel. It was quite, quite horrifying.

I have often wondered how a RISC OS user would do in a contest between systems producing (Risc OS, windows or Mac) a document or a vector graphic etc without the use of wizards

Right, here’s the OpenType font the client requires, and here’s a PDF of the semi-transparent logo they want overlaying… with the Darker blend mode, obviously. Text is in Arabic and Traditional Chinese. Off you go…

Yes. There’s been no new important application for a while. Perhaps our last was NetSurf? [I’m not counting things ported from other platforms]

I’m looking forward to the new browser; for many people a decent browser is all they need, so if it works well then it might be enough to entice some new users over. Then with more users, there is a larger market for native apps, etc.

With that said, I think the next stumbling block is going to be video playback. People expect YouTube and the like to work, so we’d need to port some video codecs (going from scratch is likely to be too much work; the H.264 spec is 812 pages). On the Pi we could make use of the hardware decoding, but then we’re in the situation of having “good” and “bad” hardware platforms.

Look at rick go from an idea to a working prototype in a couple of hours.

I’m afraid that’s possible on other platforms, too – and often much more easily. The amount of work that RISC OS expects the developer to do can be quite surprising when one is familiar with different ways of working. I’d imagine that for someone who wasn’t previously familiar with RISC OS and came to us with experience in those different ways, the effort to get something quite simple running on RISC OS could be a significant barrier.

“Where’s the library to…”

This is made much worse by the fact that much of the userbase is above retirement age, and has no real desire to learn anything new.

In my case it is not the desire that is lacking but the ability to concentrate. I did discover recently how to use Python and the EasyGui library to mitigate some of the things which, as a RISC OS user, I miss in Raspbian. But is it worth the effort to take it further? I count myself lucky to have lived through a time when many things were new and up for grabs. But then there are always new things up for grabs; but the low hanging fruit is now on a different tree.

The amount of work that RISC OS expects the developer to do can be quite surprising when one is familiar with different ways of working.

Case in point, VisualBasic (the 5/6 generation, not the new fangled .Net stuff). I made a number of applications for Windows and I don’t know how Windows works internally. I never needed to know. I design a window, put some icons in it, attach events to the icons, write some code to deal with each event.

Want to display a JPEG? That’ll be a picture box, and then you load the JPEG and assign it to the picture box’s “Image” (or was it “Picture”?) property. That’s it.
Changing format, font, style, text of an icon is basically ThisWindow.ThisIcon.ThisPropery = SomeNewValue.
Some icons were smart. List boxes, RTF input box, blah blah. You could hook into events the fart around if necessary, or you could just let them do their thing and await the end result.

One of the reasons why EBook was easy to write, asides from using DeskLib, is because I had already written a program that output plain text to a window, so I simply copy-pasted huge swathes of the redraw code and tweaked it for its new role. And tweaked it again for outline fonts today. I didn’t need to mess around with origins and working out from window position and scroll which line was which as it’s a solved problem from some code I wrote 16 years ago.

Fun fact – the origins of VDU text and outline fonts are different. If you hold Alt during an EBook redraw, you’ll see the background become alternating lines of white and light grey (like the old fanfold printer paper). This was so I could see visually how the font fit into the line spaces, however one could argue that this is an additional complication that the programmer shouldn’t have to worry about. If I want to output a letter at X,Y then the character should go there, and not have to anticipate coding in differing behaviour depending upon the actual method of plotting said letter. But, there you go. You do.

DeskLib helps to automate a lot of things, but it is very far from the likes of VB. I simply specified a “main” routine where the application began. Starting as an app, loading windows, all that guff? That’s handled by the 8-odd megabytes of runtime. Not me.
I would imagine that Windows uses a polling loop like RISC OS in order to pick up events, but at the same time it is preemptively multitasked. So… Key presses, redraws, all that sort of thing, is it queued for the application until it is ready to accept the events? I don’t know, and the nice thing is, I didn’t need to.

Another nice idea was that the development suite was integrated with the debugger. So I could load a program and use it, and when it would crash (like with a divide by zero or whatever) it would report the error and take to the offending line with the program suspended at that point. I could look at all the variables, the call stack, and so on. Fix the problem, then resume execution (which may have meant poking new values into variables and dragging the current execution point up a little to redo some of the code). It was a much faster way to develop stuff than embedding lots of calls to DADebug to write tracing messages.
The editor sucked, it was nowhere near as competent as Zap, and oddly I don’t recall a way of viewing memory (like as a byte or word dump), an odd omission for a development system. But in terms of getting stuff done in a hurry, VB had it made.

My VideoList software took years (1998 to 2003) to write under RISC OS. I don’t begrudge that time – it was a good learning exercise, and a source of frequently borrowed code.

The version I wrote for Windows (just because), I think I got the core done in a few weeks. The Windows version is more capable and took about five or so months. Two was recreating what the RISC OS version did, the rest was just adding lots of extra features. Yup, the creeping featuritis.

That’s an example of a Dev Suite done well. Look at my experiences of ESP32 programming for the polar opposite.

So, back to RISC OS. Back to knowing that, even with DeskLib, I have to start up the event library, point to the resources, load the template file, load each window… And so on.
And when it doesn’t work right, I’ll maybe get a backtrace to help narrow down the cause of the problem or the output on screen won’t look right, which will then mean dropping in numerous calls to DADebug so I can see what’s going on (if it isn’t obvious). Rebuilding, rerunning, redoing, until the problem is resolved.

Note: durations quoted do not imply 24 hour coding sessions, it has to fit around work, life, movies, cats, cups of tea, etc.

As David Pilling demonstrated with OvPro, you can produce something that is demonstrably better than the alternatives, and people will still prefer the product they know.

To the extent that they even demand the removal of perfectly sensible changes to the OS when it breaks Aemulor, on which Impression X depends…

so I simply copy-pasted huge swathes of the redraw code and tweaked it for its new role.

Of course, when it comes to using proportional fonts and custom sizes, it’s a whole different bag of kittens, and one that I can imagine taking a huge amount of time and effort.

The editor sucked, it was nowhere near as competent as Zap, and oddly I don’t recall a way of viewing memory (like as a byte or word dump), an odd omission for a development system. But in terms of getting stuff done in a hurry, VB had it made.

The thing is, even back then, Visual [whatever] development was a lot higher level than the RISC OS equivalents. You didn’t poke about in memory; you used the higher level data structures provided for you.

In probably fifteen years of using VB6 and now C# in its various incarnations to drive lab and test equipment, which often seems to involve using ugly manufacturer-specific libraries of varying quality, I don’t think I’ve ever had to do anything with memory that’s more advanced than pointing the compiler at a locating and saying “that holds a string/32-bit integer/whatever”. I’ve certainly never had to disentangle blocks of memory. I’ve had to declare code as unsafe in C# exactly twice.

Compare that to doing the same in BB4W (and yes, I have plenty of experience of doing the same in that too), and the difference is really clear. Doing anything much in BB4W requires extensive use of indirection operators on memory blocks, and can be incredibly arcane (I still have memories of getting ODBC database drivers working in BB4W, which was a world away from the same in C#).

And yes, VB6’s editor was properly nasty. That said, modern Visual Studio is pretty good — and I’m fast coming to like Visual Studio Code as well.

Of course, when it comes to using proportional fonts and custom sizes, it’s a whole different bag of kittens, and one that I can imagine taking a huge amount of time and effort.

Maybe; maybe not. Until I re-wrote the report code for CashBook 1.40, I’d always assumed that doing page layout and printing “properly” was far too complex to consider. It turned out to not be quite as scary as I’d thought – so long as you start by working out what the Wimp’s redraw and Printer Driver’s page generation code actually require, and then build the structures up from there.

In fact, if you’ve ever tried to do custom font display on Windows (such as plotting syntax-coloured, indented text to a window “by hand”), it’s not too dissimilar to what RISC OS requires. I’ve never tried printing from Windows¹, though.

¹ Beyond opening a data connection and squirting printer-specific commands through, 1980s style — which is what a lot of “manufacturing-type” printers still seem to require.

I would imagine that Windows uses a polling loop like RISC OS in order to pick up events,

It does.

but at the same time it is preemptively multitasked.

Windows programs are multithreaded. The thread that collects events is the UI thread, and if that gets preempted (which would be unusual) you’ll get the “<programname> is not responding” bit in the window’s titlebar. There’s (at least) one other thread that does window redrawing – but WPF (I assume you were using the framework) hides that one away from you: Your UI thread deals with an abstraction of the GUI, and the background thread updates it when it can.

So… Key presses, redraws, all that sort of thing, is it queued for the application until it is ready to accept the events?

Yup.

I don’t know, and the nice thing is, I didn’t need to.

Well as soon as you do anything big, you do need to.