No more big 32-bit cores for RISC OS from 2022
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 ... 19
Clive Semmens (2335) 3276 posts |
I have this horrible suspicion that at 70 I may be not very much older than the median age of this rabble. |
Steve Pampling (1551) 8145 posts |
You, sir would never be considered for a management post – you’re far too quick on the uptake. However, any techie knows that estimating a larger value and coming in under budget is useful because you can spend the excess on a few extras… |
Rick Murray (539) 13790 posts |
Acorn fanatics have yachts? What am I doing wrong…?
;-) It’s definitely 7. Or maybe SEVEN. Anything else would be wrong. That said, if it’s flexible enough to be easily ported to x86, I’m sure we can have long and tedious arguments as to why it would be called RISC OS on a CISC processor… |
Clive Semmens (2335) 3276 posts |
Hmm. I was deputy manager of an ITeC for a whole year (of a two year contract), depute (!) co-ordinator of a TVEI centre for two years (of a three year contract), and Director of Information Technology in a public school for a year. You’re right it’s not my métier, but it was my decision to leave each of these posts. The most recent time I was offered such a job though was at ARM, and I wrote this little story immediately after turning it down: https://www.deviantart.com/coshipi/art/The-project-leader-28849467 To the boss in question’s credit, he took it very well, and we’re still friends, albeit rarely in touch. |
Clive Semmens (2335) 3276 posts |
I think the meaning of RISC must have shifted quite a bit for ARMv7 to be called RISC, never mind ARMv8. |
Steve Pampling (1551) 8145 posts |
Hmmm, “you’re the project leader” “Oh, sorry I missed that email and HR clearly didn’t get the promotion details”1 Sometimes you can’t help saying things – in my case most of the time. 1 I think I have the wording roughly correct. |
Clive Semmens (2335) 3276 posts |
No, the wording in my story is closer to what actually transpired – probably, I don’t remember exactly, fourteen years later. But the boss and I were actually pretty good friends. He wasn’t really any more management material than I am – nor for that matter was his boss, also a good friend. Last time I saw middle boss was at big boss’s untimely funeral. Untimely: this was a few years ago now, and he was a fair bit younger than me. |
Alan Robertson (52) 420 posts |
Just a friendly reminder that conversations not related to the thread subject header should be moved to the Aldershot forum. |
Alan Adams (2486) 1145 posts |
I had the opposite issue – I was working as the “System Manager” of VAX/VMS systems. The term System Manager was used within DEC and all the manuals as meaning the person doing the job I was doing. However the HR department of the company I was in didn’t like the use of “Manager” for the job title since it didn’t involve supervising people. For most of the time the job title was “Senior Scientist”, which didn’t start to describe the job. |
Charlotte Benton (8631) 168 posts |
It might also benefit from being able to port a lot of functionality1 from other free open-source software, instead of having to build everything from the ground up. 1 Once again, assuming the existence of a multimillionaire to pay the team of programmers who would be doing said porting. |
Paolo Fabio Zaino (28) 1821 posts |
I think real Acorn’s fanatics are all probably busy debating how to best use SWP/whatever ARM instruction and how someone could re-implement MEMC in FPGA and break the 4MB RAM limit looool… Acorn was expensive yes, but it was for computer/hardware geeks and hackers and RISC OS 5 is the testimony of what Acorn was about. Maybe Acorn’s management or Acorn’s employee wanted something different (I have heard many times the story of “competing with Apple”). But if they really wanted to compete with Apple then they did it all wrong ;) |
David Boddie (1934) 222 posts |
To be able to judge the amount of resources it would take to update RISC OS, you need to say what it is about RISC OS you would keep and what you would change. For example, some people would like features that would make applications more robust, but other people would disapprove of changes that restrict low-level hardware access. For any big change you make there would always be people who would say that the result is not RISC OS. Maybe someone needs to draw a line under “classic” RISC OS and just tell users to run it in an emulator or in a virtual environment. Either take what they like about the original system and build something new to sit underneath, or just use an existing kernel with good hardware support. Then work on ways to integrate the “classic” environment with the new system. |
Colin Ferris (399) 1806 posts |
Nice – if there was a ARM 64bit diss mode for Zap :-) |
Paolo Fabio Zaino (28) 1821 posts |
Linux was written by a single guy, it sucked at the beginning, but his ideals and ideas convinced others to join for free and now that OS dominates the world… So if we are talking about multimillionaires then we are talking about Windows… which in so many years of development / re-factoring and re-development still struggles to deliver the original idea… Same for macOS until Apple embraced Open Source and restarted from there, yes then they added a lot to it, no doubt, but it still started from a free available source… RISC OS doesn’t necessarily needs investors, it simply need ideas and ideals:
Regardless where each of our vote may go in the list above (or adding more options to it) RISC OS won’t be able to cover all the points, so I think at some point people will have to make some decisions and this is so also for ROOL/RISC OS Developments/Cloverleaf etc. However, with good ideas and good commitments, it is possible to re-write RISC OS so that it will be able to support the past and the future and, through time, also full-fill a lot of the points above as long as the Architecture of the OS will be that flexible and portable and as long as people are willing to make compromises. |
David Feugey (2125) 2709 posts |
And even with that you’ll not be future proof at all. From a back-end point of view, future is ML/cloud platforms. And for user interaction, it’s web and mobile platforms. Oups, sorry, it’s not future ; it’s happening today. The big web browsers are not and will not be compatible with RISC OS. 1/ as back-end on IoT platforms And so the question is not 64bit or not. But… |
Rick Murray (539) 13790 posts |
No, the kernel was originally written by a single guy. All that is Linux (or GNU Linux to the RMS fanboys) is a huge collaborative effort.
I think that’s because Windows doesn’t actually know what it wants to be. Look at the abortive attempt to shoehorn tactile paradigms into a UI based around keyboards and mice.
Darwin, as in the underlying kernel, is based upon open source projects, yes. Because all that low level nitty gritty stuff is hard, so probably best to build upon existing code. The UI isn’t open, however.
We desperately need investors, because ideas don’t get stuff done.
Never. Anybody who thinks that is delusional. It’s a cut throat enough world that compared to the many Android clones and the iPhone, Microsoft largely failed to do anything at all in the mobile arena. Microsoft! Granted, they had no idea what they were going to offer, as a desktop UI simply doesn’t scale to a mobile device (you’ll notice most Android apps are full screen and those that can split screen do so with restrictions; and can iOS multitask properly yet?), however Microsoft had a lot of know-how and a lot of money and still failed epically.
There’s no such thing. “Future proof” is just marketing speak for “not immediately obsolete”.
Yes, indeed. And on future devices?
Yes, I think it really ought to be a priority to get onboard WiFi and Bluetooth working. I think somebody is fiddling around with the maze of twisty passages that is Bluetooth, and the WiFi part might be slightly hobbled by our prehistoric network stack. Hell, we don’t even do IPv6 yet.
This is, actually, a rather interesting idea. I wonder if it might be possible to have a module that can register and run tasks in a polling loop similar to the Wimp, but with no UI components? With the caveat that programs are executed in-situ (none of this &8000 malarkey as it would clash with everything else) and these programs get forcibly time sliced if they don’t hand of back to the coordinator in time. That way, they can sit in the background and just “do stuff”.
Depends upon what you mean by user privacy. The OS doesn’t spy on its users (something I find rather abhorrent), however there is nothing to stop me adding a few lines of code to Manga to quietly rummage around your harddisc looking for things of interest (PhotoDesk or ArtWorks registration key?) to add as a payload during the next check.
No, not until it’s written in a language that people can read.
I concur. The easiest way to watch a system break is to let a bunch of schoolkids loose on it. If there’s an obvious flaw, it will be found. Case in point. My shiny A3000 plugged into the school Econet with a FileStore and a bunch of Beebs/Masters.
Depends upon your definition of fun. I’m quite liking Android since it allows me to watch The Bureau Of Magical Things (that’s a pleasant way to end a day that doesn’t require too much thinking nor runtimes measured in hours).
Personally I think a hybrid approach would be best. CMT by default (there’s nothing wrong with CMT done correctly) with the system able to seize control and poll over to another task if an app stalls without specifically telling the system “this will take time”.
They’re all the rage today. But will they be in the future when people understand better that their data is no longer their own, and their access to it depends upon the whims and management of somebody else? Cloud is easy and cheap. But if your business depends upon cloud, you’re effectively dropping all of that stuff in the hands of somebody else.
Not. Too much of how RISC OS currently works depends upon shuffling pages of memory around. Always has. You might be able to make a hack of something that feels a little like RISC OS that has stuff in fixed (restricted) areas and can only run one “app” (at a time). Might work for embedded devices, I suppose? But you’d probably need to lose 99% of what makes RISC OS what it is. Personally, I’d wonder if making something akin to a BBC Master on acid (or, you know, Arthur) might be a better option here? A separate creation from the RISC OS stable, but not RISC OS but something with similarities that can run on the more restricted hardware?
Yes. In this day and age, an “X, Y, button” is inadequate. But since there was never any actual proper mechanism for adding “pressure” (like tablet stylus tips), I won’t hold my breath for anything exciting like multitouch. ;-) |
David Feugey (2125) 2709 posts |
That’s completely the case today.
I know. But it’s perhaps an easier task than a full 64bit OS :)
MPU does that.
Why not. A limited RISC OS for microcontrollers would be great. |
Steve Pampling (1551) 8145 posts |
I took the view, a number of years ago, that anything I implemented at work was going to last a lifetime1 so I didn’t have to revisit the job. Sometimes that meant producing script/code that was so heavily commented that it was obvious to a five-year-old what needed altering to fit some new circumstance.
I note this: “Raspberry Pi OS (previously called Raspbian)” and wonder is that Linux flavour diverging even more from the Debian? If so is that to aid kiddies getting down into it or to stop them fiddling? 1 For “lifetime” read my working lifetime. You see, the future is someone else so when my stuff finally breaks or can’t cope – N.M.P/S.E.P. I did my bit. |
Paolo Fabio Zaino (28) 1821 posts |
Actually Linux IS the kernel… the entire distribution is Ubuntu or Debian or Fedora… Stallman insists it should be called GNU Linux because the tools used to build Linux and a Linux Distro are GNU.
Investing on WHAT exactly? Apologies, not trying to be mean or something, I am simply struggling to understand. Investors traditionally wants their money back + interests produced by a product/idea/solution built using their investments and that has sold/created a market/produced more money… So, my question to you is: Did you mean we need someone to pay for all the work that has to be done regardless the outcome (for example charity donations) or did you really mean investors in the traditional sense?
That’s exactly my point, it can be kept a hackable OS on future devices, but hackable will unavoidably result into a non stable system for a user that is not careful, hence is it a goal or a side effect? my point was this is a requirement that would collide with the Daily Driver OS requirement for example… makes sense?
Yes of course it is possible! Do you want to start a thread on here or on chatcube and put some ideas together and starting some code for fun?
True, my definition of fun is hacking some code to do something creative, play music and recording it, playing games that I like and that do not cause me stress and infinite learning curves lol have low level access to interface some electronics and do something creative with it, design new programming languages or just play porting old ones… so all stuff RISC OS can do, but you are right this one is subjective.
Sure, but you still need an edge device to access that (or to get your own data stollen lol). I think RISC OS should be perceived as edge computing platform, not cloud. We have cURL now, and HTTPLib, embedTLS and AcornSSL, box access and others, so pretty much getting close to access a lot of cloud stuff and sure we can add more and more protocols for this, but I would recommend we get the multi-core sorted first and the modern browser too.
Agreed and not just that. IoT is an actual market, in other words Start Ups and Companies willing to invest to have a better platform for their own products and ideas. As always just my 0.5c… |
Steve Pampling (1551) 8145 posts |
Stallman is My take on this question:
Simply put, I think investing in RO should be eligible for tax relief like other charitable donations. |
David Feugey (2125) 2709 posts |
Correct. That’s why I suggest a Weboob like platform.
Absolutely. There is (a lot of) money here. And it’s a better target than to believe that RISC OS could compete with Android or Linux on multicore 64bit processors. That does not mean that we should not evolve. I dream of a Titanium like board with a connector for compute module, heavy thermal solution (2.147 GHz) and perfect (PCIe) SATA support :) |
Peter Howkins (211) 236 posts |
Cortex-M cores only support the Thumb1 or Thumb2 instruction set, not the arm32 instruction set. (this would be as big a port as the 64 bit port, but with the added addition of trying to bodge a flat memory map onto risc os, no mmu) |
Rick Murray (539) 13790 posts |
Please tell me that we’ve learned the lesson about not writing it all in assembler…
Just treat it like a supercharged Beeb with some RISC OS like additions. ;-)
Nothing unusual there, then.
What relevance is that? Is everything built using those tools supposed to have a GNU prefix attached? The better explanation is that RMS jumped onto the Linux kernel after seeing what one guy did himself (given the fact that GNU’s own kernel never went anywhere). Thankfully Linus was savvy enough to adopt the current version of GPL but only that version, so the kernel isn’t tainted by the restrictive crap that has found its way into GPL 3.
That’s the question. Though you’d likely be better off directing that sort of question towards people who know a little more about it than, well, me.
Kind of like how shoving a metal spoon into the top of an operating toaster will likely result in fire/electrocution/death. One needs to accept a certain amount of responsibility. If they aren’t up for that, point them towards an iPad… ;-)
What planet are you writing from? IoT is a valid thing, but it’s start ups happy to try to coin it by flogging whatever barely functional “just about does what it says on the box” crap that they can make the cheapest. Allow me to demonstrate: https://heyrick.eu/blog/index.php?diary=20170617
It’s probably supposed to be said like “web oob”, but… |
Steve Pampling (1551) 8145 posts |
So, that’s three years now? The reply should be with you any day now. Yes, Steven, take your medication now…
Yeah, I even tried typing it upside down/back-to-front on a calculator but that W / M has me stumped. |
Paolo Fabio Zaino (28) 1821 posts |
not my problem lol, given that you were trying to be precise I just brought your attention on the fact that I have been extremely precise in the wording chosen on the original comment ;) For me they can debate as much as they want on GNU yes/no, what matters to me is what it can do for me at work or in life… There are hundreds of Open Source OS and/or Kernels out there, Linux (+Kernel, for who still thinks Linux is an entire OS distro) dominates the world because:
Amen on that!
Earth, you?
Isn’t RISC OS already free? Not sure what point you’re trying to make, dare to clarify please? Also a little help for you: Start Ups would:
The above is already a lot compared to what RISC OS has now, don’t you agree? |
Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 ... 19