log in | register | forums
Show:
Go:
Forums
Username:

Password:

User accounts
Register new account
Forgot password
Forum stats
List of members
Search the forums

Advanced search
Recent discussions
- WROCC Newsletter Volume 41:11 reviewed (News:)
- WROCC March 2024 meeting o... Hughes and Peter Richmond (News:1)
- Rougol March 2024 meeting on monday with Bernard Boase (News:)
- Drag'n'Drop 13i2 edition reviewed (News:)
- South-West Show 2024 talks (News:4)
- February 2024 News Summary (News:1)
- Next developer fireside chat (News:)
- DDE31d released (News:)
- South-West Show 2024 Report (News:)
- South-West Show 2024 in pictures (News:)
Related articles
- RISC OS 2002 show
- RISC OS 2001 show
- Wakefield 2001 show report
- Wakefield 2003 - the preview
- June quickies
- Quick bits #4
- Wakefield 2000 show report
- Wakefield 2006 show report
- Wakey Wakey, it's show time again!
- RISC OS - 24 bits
Latest postings RSS Feeds
RSS 2.0 | 1.0 | 0.9
Atom 0.3
Misc RDF | CDF
 
View on Mastodon
@www.iconbar.com@rss-parrot.net
Site Search
 
Article archives
The Icon Bar: News and features: Post show quickies
 

Post show quickies

Posted by Richard Goodwin on 10:34, 26/10/2001 | , , , , , ,
 
Time to catch up on all the news we've missed since the show (or were too lazy to put up before the show!)

Some of you have already noticed that the MicroDigital website has been updated, and talks about the Omega as a current product. Are we going to have some good news in the next few days?

Budget games meisters Paradise have also revamped their website, with new pricing, a print-your-own order form, and cool graphics tablet and Playstation joypad interface hardware.

The RISC OS Banner Network is setting up a system to help advertise RISC OS related websites. The service is free to both commercial and non-commercial interests, all you need to do is add a little code to your website.

Digital Photography Now is a website all about digital cameras, scanners, photo printers, accessories, software and related services. It's a little PC-centric, in that some of the pages don't yet work with RISC OS browsers properly, but at least someone's flying the flag - our old friend Paul Vigay is the RISC OS editor for the site, and has an article up extolling the virtues of digital photography on RISC OS machines. And here's me without a new embarrassing picture of him to put up...

In the spirit of his previous PGP work, Privacy Guru Programmer (see what I did there? :) ) Nat Queen has produced !Spook, a program that produces random phrases containing "trigger words" aimed at defeating Echelon, the NSA-run system vacuuming up emails, 'phone calls and faxes with scant regard for individual privacy. Get !Spook here. This news item would have been even more topical if it'd been posted before Jam Echelon Day, but sadly that was the day after the show so I had other things on my mind - sorry! As has been pointed out elsewhere, another way to jam echelon is to encrypt every message with the strongest encryption you can get your hands on, even that thank-you email to Granny, to make the spooks work harder.

And finally, it's semi-traditional to have at least one item about either ArtWorks or Vantage in these updates, so it would be rude of me not to mention that Martin's upgraded ArtWorks to use 128 levels of antialiasing, up from the usual 16. Watch his website for the upgrade, or join the new ArtWorks smartgroup.
 
  Post show quickies
  (10:36 26/10/2001)
  Ann onymous (11:19 26/10/2001)
    William Black (11:41 26/10/2001)
      Lee Johnston (12:00 26/10/2001)
        Rob Kendrick (13:32 26/10/2001)
          MicroAnalogue (13:53 26/10/2001)
            Anonymous (13:54 26/10/2001)
              Jim (14:19 26/10/2001)
                Whinge-a-lot (14:40 26/10/2001)
                  Guy Inchbald (16:28 26/10/2001)
                    Andrea Lloyd (19:45 26/10/2001)
                      Michael Stubbs (21:35 26/10/2001)
                        Rob Kendrick (14:01 27/10/2001)
                          Jon Hall (14:18 27/10/2001)
                            Annraoi (15:26 27/10/2001)
                              Rob Kendrick (15:31 27/10/2001)
                                Lee Johnston (16:29 27/10/2001)
                                  Rob Kendrick (18:27 27/10/2001)
                                    Guy Inchbald (23:08 27/10/2001)
                                      Annraoi (16:51 28/10/2001)
                                        Anonymouse (19:36 28/10/2001)
                                          Lee Johnston (10:59 29/10/2001)
                                            Guy Inchbald (15:11 29/10/2001)
                                              Annraoi (15:52 29/10/2001)
 
Richard Goodwin Message #89310, posted at 10:36, 26/10/2001
Unregistered user Usual afterthought - disregarding the styleguide to use !Spook instead of just Spook, I though it looked like a "Not-Spook" kind of joke :)
  ^[ Log in to reply ]
 
Ann onymous Message #89311, posted at 11:19, 26/10/2001, in reply to message #89310
Unregistered user Well, If their website is anything to go by, the Omega will be an awfully engineered bag of crap. I sincerely hope that they shoot whichever goon made the site, and quickly.

As for what's actually said about Omega, when you piece together the total disgrace of a website, it actually looks quite good - IF it all works (*mutter about sound and USB on Mico*)
  ^[ Log in to reply ]
 
William Black Message #89312, posted at 11:41, 26/10/2001, in reply to message #89311
Unregistered user Someone who should know about these things informs me that it's probably going to be slower than a Kinetic due to the methods used in bodging the hardware to make RISC OS work. Omega: big pile of rubbish.
  ^[ Log in to reply ]
 
Lee Johnston Message #89313, posted at 12:00, 26/10/2001, in reply to message #89312
Unregistered user That's a bit harsh really. How about

RISC OS: OS based around archaic engineering principles that prevent it from moving easily to other hardware.

Your comments deride MD, my comments deride the original RISC OS designers. Which is correct? At the end of the day the problem we face (and have been facing for three years) is a software problem, and the software is RISC OS.

Note it is not my intention to deride the RISC OS developers. All of them are far more talented than I in OS design and construction but it doesn't remove the fact that the kind of tricks the Omega will have to pull in order to work shouldn't be necessary.
  ^[ Log in to reply ]
 
Rob Kendrick Message #89314, posted at 13:32, 26/10/2001, in reply to message #89313
Unregistered user RISC OS is also very old, but even by its age, it's an incredibly shoddy bit of code. It isn't MicroDigital's fault that Omega bodges things, and may be slow, but it is their fault for trying. Lots of people have been saying an open source replacement for RISC OS is needed, and might be interesting to write - I'm beginning to agree. And this time round, write it in a higher level language...
  ^[ Log in to reply ]
 
MicroAnalogue Message #89315, posted at 13:53, 26/10/2001, in reply to message #89314
Unregistered user Whatever people may think of MicroDigital, the Omega *HAS* to ship soon, and be a stunning product. If it not, the RISC OS desktop market will die completely.

Mico: In addition to the usual USB/sound issues, here are some more posers:

Does the audio on the Mico actually work? Does it use a software driver to do pseudo DMA where the CPU copies the audio data to the audio hardware itself?

Does the audio hardware actually support audio sampling and digitisation?

Does it have joystick drivers?

Does it have MIDI drivers?

Does the Mico wake up on anything other than the front button?

Does the Mico have DDC monitor support?

What sort of ethernet card does it ship with? Does anyone have any performance figures?

P.S. Their web site, along with those of most other RISC OS companies, is a disgrace. Go and look at a good one, like http://www.apple.com

  ^[ Log in to reply ]
 
Anonymous Message #89316, posted at 13:54, 26/10/2001, in reply to message #89315
Unregistered user So where are MD's legal team?! Are they on the phone to tib.com now?! :-)
  ^[ Log in to reply ]
 
Jim Message #89317, posted at 14:19, 26/10/2001, in reply to message #89316
Unregistered user So what if it's slower than Kinetic? At least the DMA might work!

Who told you that anyway? Someone close to Castle perhaps...

With hardware MPEG playback, JPEG decoding and graphics acceleration it'll blow Kinetic and anything else out of the water.
I won't buy one immediately - I'm going to wait and see if Evolution appears. Although the spec isn't as snazzy, it's probably the better machine.
  ^[ Log in to reply ]
 
Whinge-a-lot Message #89318, posted at 14:40, 26/10/2001, in reply to message #89317
Unregistered user Having just seen it at the RISC OS 2001 show, I want a Phoebe.
  ^[ Log in to reply ]
 
Guy Inchbald Message #89319, posted at 16:28, 26/10/2001, in reply to message #89318
Unregistered user nothing like a good maon on a Friday afternoon...

... and that lot was nothing like a good moan.
Why don't you all go home for the weekend?
  ^[ Log in to reply ]
 
Andrea Lloyd Message #89320, posted at 19:45, 26/10/2001, in reply to message #89319
Unregistered user So, you think it has hardware MPEG playback? Fantastic - MicroDigital's wonderful marketing towards idiots seems to be working - at least they do something well.
  ^[ Log in to reply ]
 
Michael Stubbs Message #89321, posted at 21:35, 26/10/2001, in reply to message #89320
Unregistered user I wish people would stop saying "if this doesn't happen by [insert time here] then the market will die". What bollocks. It is said by people who can't possibly know the real state of the market, who have no idea what Castle and RiscStation are up to and who moan at everything.

So what if RISC OS depends on custom hardware? We just need faster custom hardware with RISC OS updated appropriately. That doesn't make it poor.

Buy a Kinetic, enjoy the speed then buy an XScale machine when it's out. If it's from Castle or RiscStation, at least you know you'll get decent support and that it works.
  ^[ Log in to reply ]
 
Rob Kendrick Message #89322, posted at 14:01, 27/10/2001, in reply to message #89321
Unregistered user I think what most people are actually thinking is "If *somebody* doesn't release something soon, the market will die" rather than just MicroDigital. RISCOS Ltd. seem to be the main bottleneck here, not the hardware producers.

Depending on custom hardware is expensive and tends to lead to lower performance. If RISC OS boxes could be built using off the shelf componants, we'd have much faster machines, because it doesn't mean you have just one or two companies competing to produce better products, you have hundreds. Using 'faster custom hardware' will just mean we'll face the same problem at some time in the future.

Also, your suggestion of "Buy a Kinetic" and then "buy an XScale machine when it's out" doesn't suit everybody - I'd love a Kinetic, but I can't afford it - I recently nearly broke the bank buying a couple of 710 RiscPCs to aid me developing the new Slayer. People who would buy a Kinetic, bin it after a (well, lets be hopeful) few months and buy another machine have more money than sense.
  ^[ Log in to reply ]
 
Jon Hall Message #89323, posted at 14:18, 27/10/2001, in reply to message #89322
Unregistered user MD's site is broken.
  ^[ Log in to reply ]
 
Annraoi Message #89324, posted at 15:26, 27/10/2001, in reply to message #89323
Unregistered user Not having seen an Omega and had a chance to run benchmark code on it it's difficult to say its faster/slower or the same as the Kinetic.

Another thing often said (and does not make it any truer) is that RISC OS is at fault for being designed to use "specific" hardware - that is a perfectly reasonable objective if the OS is intended to only ever run on two specific devices IOMD/VIDC.

Simply getting it to run with "off the shelf" devices is not necessarily benefitial in that most existing software won't work and secondly there are well known issues with some of the PC hardware out there (we'd simply be substituting one set of problems for another and needing a substantial processor speed increase just to offset in loss of performance due to the introduction of a HAL).

Get real guys the bulk of code runs with EXISTING RISC OS software and hardware if you change the hardware to a system that is utterly different software will break all over the shop and (let's face it) most software vendors are not there still to update their stuff to continue working. So let's get real, the best way to improve the platform is to UPGRADE the VIDC/IOMD and do so in such a way that RISC OS continues to work (with minor tweaks).

Improve the IOMD bandwidth, give VIDC some (simple) video processing capabilities (at least bit blitting) and a wider memory bus for VRAM and that would do the trick.
  ^[ Log in to reply ]
 
Rob Kendrick Message #89325, posted at 15:31, 27/10/2001, in reply to message #89324
Unregistered user Annraoi: What I'm saying is that it should have been designed in the first place to have a HAL - Although a HAL may decrease performance, RISC OS has many areas that need to be improved, one of them being general performance - Just look at Linux and Windows - both have HALs, and for everything except the usability of the GUI (which has nothing to do with the hardware) are much much faster and a lot more stable.
  ^[ Log in to reply ]
 
Lee Johnston Message #89326, posted at 16:29, 27/10/2001, in reply to message #89325
Unregistered user Changing the hardware will only break software which bypasses the RISC OS APIs and there aren't any useful applications that do that are there..

..oh hold on there are some that do it in an attempt to squeeze performance out of the machine. Oh dear.

Custom hardware is great if it can match the performance of off the shelf equivalents otherwise it becomes expensive in comparison. As good as the lightning chipset may eventually be I can't see it coping with the DirectX 8 performance benchmarks I've seen recently - now they are incredible (did need a 1.3Ghz athlon and GeForce 2 card mind).

Also the custom hardware needs to be available to anyone who wants to use it. I bet the reason we haven't seen StrongARM machines from RiscStation is because Castle have the only reliable source of VIDC20 and IOMD chips. This, and improving performance, is why Evolution wanted to use off the shelf components.

Mind you, right now I'd settle for anything new and a definite roadmap as to where the OS is going.
  ^[ Log in to reply ]
 
Rob Kendrick Message #89327, posted at 18:27, 27/10/2001, in reply to message #89326
Unregistered user Anybody fancy doing the open source RISC OS clone? ;-)
  ^[ Log in to reply ]
 
Guy Inchbald Message #89328, posted at 23:08, 27/10/2001, in reply to message #89327
Unregistered user well, if Red Squirrel can do an emulator for PC, maybe an emulator for ARM Linux is easier?
  ^[ Log in to reply ]
 
Annraoi Message #89329, posted at 16:51, 28/10/2001, in reply to message #89328
Unregistered user You'll find that PC performance (whether it be under Linux or Windows) is helped in no small way by having a fast memory system (133 or 266MHz (if using DDR)) while traditional Acorns use Ram clocked at around 16MHz. Additionally the fastest ARM is only clocked at 300MHz but squanders its processor bandwidth (66MHz) by hooking up to the I/O system at 16MHz while PC's have processor clocks of 1.4GHz+ and I/O at 33/66MHz. I'd love to see Windows XP running on a 233MHz P-III with EDO RAM (60nS) and an I/O bus that creaks at 16MHz. The point is that BECAUSE of its non-use of HAL and its close relationship to the hardware RISC OS manages to be much more responsive with less bandwidth and processor cycles than the alternatives.

I'd go so far as to say that the best course of action is for an OPEN IOMD/VIDC system to be designed (this could be prototyped using VHDL/Verilog design systems) and the hardware vendors could then pay to have it transferred to ASIC or higher end FPGA (either would confortably beat the existing VIDC/IOMD). This could be done under something like the GPL (so that vendors would have access to but not exclusive rights to the design - so all vendors would have equal rights to the design - if they want to make their product better than the competition they'd produce it on a fast ASIC rather than FPGA).

The alternative is an "open" RISC OS that runs on an ARM with PC hardware for I/O and memory. This is a steep mountain to climb, there may be licenising issues, problems with some of the PC chipsets and non-availability of some information needed to complete the task.

The idea of simply running RISC OS on Windows or Linux would seem to me to miss the point (you have your long Windows (or (slightly faster Linux)) boot and then a boot into an "emulated" RISC OS. What would the performance be ? (not good) and to cap it if the underlying OS (say Windows) crashes people might blame RISC OS or a flakey emulation rather than Windows (that's the sort of thing that WOULD finish off the platform).

  ^[ Log in to reply ]
 
Anonymouse Message #89330, posted at 19:36, 28/10/2001, in reply to message #89329
Unregistered user To all those who keep repeating 'we need a HAL'...If were moving towards standard PC hardware, PCIcards, graphics cards etc. you don't need a HAL. Look at any desktop PC today and you'll find theyall use the same HAL. Why? because they're all thesame hardware. Each pci or graphics card has it's own driver.That's the problem we face, getting drivers forthe hardware we want to use. In my oppionion,changing RISC OS to use a HAL would be a totalwaste of effort.
  ^[ Log in to reply ]
 
Lee Johnston Message #89331, posted at 10:59, 29/10/2001, in reply to message #89330
Unregistered user Annraoi - now the idea of an open VIDC / IOMD design is not bad. I hadn't really considered that before and it would certainly solve the problems of only one hardware manufacturer having access to the necessary parts.

Anonymouse - I'm tempted to think that post is a troll. The only reason you can use all those drivers is because of the HAL. The drivers implement the specific parts of the HAL for the particular piece of hardware in question. Of course the HAL is the same otherwise the OS wouldn't be able to talk to it. Getting the information to write the drivers is difficult but not impossible as not all the hardware developers operate a closed door policy.

Even if it's decided that off the shelf hardware isn't the way to go a HAL would improve things in the long run - for a start you wouldn't need to build custom hardware that was backwards compatible all the time.
  ^[ Log in to reply ]
 
Guy Inchbald Message #89332, posted at 15:11, 29/10/2001, in reply to message #89331
Unregistered user I was wondering if a (relatively) quick-and-dirty ARM Linux port could kick things off, with key bits progressively transferred to WINE-style system calls, until the whole thing eventually ceases to need Linux.
  ^[ Log in to reply ]
 
Annraoi Message #89333, posted at 15:52, 29/10/2001, in reply to message #89332
Unregistered user Thanks Lee: I think the Open IOMD/VIDC concept is a valid concept.

And just the way a whole crowd of people got behind Linux and made it happen I cannot say the same could not happen with this concept. It addresses a number of issues, not least that of improving the performance of RISC OS based hardware.

Just by way of comment this sort of thing has been done with "Freedom CPU" design (a site discussing this is www.f-cpu.org), they even have a manual at http://www.f-cpu.de/man/i7/part1.html

Regards


Annraoi
  ^[ Log in to reply ]
 

The Icon Bar: News and features: Post show quickies