Talk:Acorn Archimedes/Archive 1

What, no photo???

-- User:70.149.166.7 17:45, 6 October 2005

You could also try requesting one at Wikipedia:Requested pictures#Computers -- Solipsist 17:04, 6 October 2005 (UTC)

The current photo could be replaced with one showing a native CRT monitor. One such image is this A3020. But personally, I think an older machine would be preferable, as a reference to the origin of the series. -- Trevj (talk) 15:53, 23 December 2011 (UTC)

There are some others on flickr but they show an Atari monitor. -- Trevj (talk) 16:22, 23 December 2011 (UTC)

Silly me! Soon we'll be able to use the A310, A540 or similar from Chris's Acorns -- Trevj (talk) 16:31, 23 December 2011 (UTC)

backward compatibility

Was the Archimedes backward compatible with the RISC?

Huh? The Arm processor the Archimedes used had a RISC architecture, if that's what you mean. Cpc464 01:31, 16 February 2006 (UTC)

IIRC, the Archimedes was the first RISC computer designed for home use --Rbanffy (talk) 15:06, 2 March 2010 (UTC)

Future work

I'd like to get this moving again. Aside from a picture (we definitely need a couple) it's lacking simple things like an infobox and a refs section. It also fails to mention that 4th Dimension were the best games company that ever existed, which is objectively true. Chris Cunningham 15:37, 30 August 2006 (UTC)

I must say I don't like the current picture; the flat screen monitor, while cool, is not really typical. I'm sure loads of people have one kicking around, but if not I could take a picture of my A310 with the 'original' and IMO much more typical rebadged-as-Acorn Philips monitor (AKF11 IIRC). (It's probably better if someone else has a photo, the flap covering the controls on mine has broken off and hence has a visible chunk of blu-tac holding it on. ;-) ) - S --195.137.91.247 23:27, 29 October 2007 (UTC)

I really agree a photo with a matching Acorn monitor should be added. And one can always digitally restore the device (it's not like "cheating" as long as it reconstructs the original appearance) --Rbanffy (talk) 15:06, 2 March 2010 (UTC)

A420

I seem to recall the A420 is an urban myth: only the A410 and A440 were released, though all three were released as A4xx/1 versions. And if I'm going to be pedantic, none of the 300 or 400 models had an 'A' prefix: they were called Archimedes 410/1 (for eg) 81.179.140.82 16:23, 15 December 2006 (UTC)

Not true, I've got one sitting here. It was the A410 (not A410/1) that was never actually produced. 85.211.7.182 (talk) —Preceding undated comment added 14:14, 1 August 2010 (UTC).

Impact

If I remember correctly:

• Tesco only switched to RM after Acorn ceased producing desktop computers.
• The major cause for blame for Acorn losing market share to the IBM PC was not multimedia capabilities - even latterly for the most part the same software was being produced for all 3 platforms including 3 in 1 CD ROM. The biggest impact was the growth of parent school governors who wanted to see machines like they used at work.

80.177.52.130 (talk) 12:13, 13 May 2008 (UTC)

Why no mention of the ARM processors in the "Impact" section? Whilst the Archimedes might be consigned to history, and RISCOS very much a niche OS, the processor technology developed for these machines is now ubiquitous. Should it not be mentioned? Captain Sumo (talk) 10:20, 12 April 2016 (UTC)

Agreed, very much so. Andy Dingley (talk) 10:43, 12 April 2016 (UTC)

List of models

The recently added 'ARM core' column is incomplete as I don't know these and don't have time to check them at the moment. --Trevj (talk) 12:30, 22 June 2011 (UTC)

I'm willing to be corrected, but I was always pretty sure that there were essentially only the ARM2/ARM3/250mez/ARM250 cores? The A3xx & A4xx were all ARM2 machines. The A4xx at least could also be upgraded (user upgraded in fact, as it was a socketed chip) to ARM3. a_man_alone (talk) 14:08, 22 June 2011 (UTC)

A quick look through the enviable and comprehensive Chriswhy pages - which are heartbreakingly nostalgic - seems to confirm that they were all Arm2. I've updated the list. a_man_alone (talk) 14:14, 22 June 2011 (UTC)

Thank you I suspected Mr Whytehead would have the info! Me, I jumped straight from 6502 to ARM610! --Trevj (talk) 14:20, 22 June 2011 (UTC)

Discussion at Template talk:Acorn computers#Proposed move/new title

You are invited to join the discussion at Template talk:Acorn computers#Proposed move/new title. Trevj (talk) 18:03, 18 October 2011 (UTC)

Fred Harris introducing the Archimedes, with Roger Wilson

These video links will be useful here. -- Trevj (talk) 18:43, 25 October 2011 (UTC)

Those links are dead, but this appears to have the same content: http://www.youtube.com/watch?v=mou73QdF-NU --Guy Macon (talk) 12:14, 1 February 2013 (UTC)

Comparison articles

Seeing as this article is about the series of computers, I thought it'd be worth comparing content with similar articles. Some examples are Apple II series and Atari 8-bit family. What others are there? -- Trevj (talk) 15:16, 27 February 2012 (UTC)

And some high quality articles (although not series) are Macintosh Classic, iPad, ZX Spectrum, MacBook and MacBook Pro. -- Trevj (talk) 16:19, 27 February 2012 (UTC)

Old discussion

"it could run a piece of software faster and with better visual quality than the more-famous Commodore Amiga" - I was an Archimedes owner, so I have no axe to grind, but is that true? Faster maybe, but I'm not sure the standard graphics capabilities were unambiguously better than the Amiga. For a start, didn't the Amiga have some sort of hold-and-modify display which allowed all 4096 colours on screen at once, as opposed to a maximum 256 colours on the Archimedes? And the palette was not very flexibly redefinable in 256 colours on the Archimedes either, but I have no idea what the Amiga's capabilities were. -- S — Preceding unsigned comment added by 213.219.24.100 (talk) 12:38, 4 November 2003 (UTC)

The Arc graphics chipset was actually inferior in speed to the Amiga one, but the seperior CPU meant it could out-perform it anyway. Compare Zarch (Arc) with Virus (the mildly stripped-down version of Zarch so the Amiga/ST could handle it). The Arc was better at handling higher screenmodes than the Amiga. I was an Amiga500 owner, & I think the Arc was the better of the two. The CPU was so weak that it greatly limited what could be done with the chipset's capabilities. Crusadeonilliteracy 13:44, 4 Nov 2003 (UTC)

I looked it up: The CPU in the Amiga500 was 0.7MIPS, whereas the 8MHz ARM CPU in the early Archimedies was 4.5MIPS. Ouch was the Amiga was underpowered with its 1970s CPU. I knew the 68k was slow, but I didn't realise it was in the same league as the CPU in the Apple II. — Preceding unsigned comment added by Crusadeonilliteracy (talk • contribs) 13:06, 28 January 2004 (UTC)

(While this discussion is old, that last statement still deserves a reply for the sake of others who might read it.) "I knew the 68k was slow, but I didn't realise it was in the same league as the CPU in the Apple II." It wasn't. You have to very be careful when comparing MIPS values from different architectures. The 68000 was a very different architecture from the Apple II's 6502, and its instructions were much more powerful. Even if you consider memory bandwidth alone, an 8 MHz 68000 had about 4 times as much bandwidth as the Apple II's 1 MHz 6502. The same warning about MIPS comparisons applies when comparing a CISC architecture like the 68000 to a RISC one like the ARM. Colin Douglas Howell 06:09, 31 October 2006 (UTC)

It's quite an invalid comparison indeed. The Amigas CPU may have been inferior to the Archimedes, but it's laughable to compare it to a 6502. — Preceding unsigned comment added by 193.69.146.66 (talk) 13:29, 1 March 2007 (UTC)

Indeed, 1 mips for an 8 bitter means 1 mio 8 bit adds per second while it means 1 mio 32 bit adds for a 32 bit cpu. That means at the same mips rating a 68000 is roughly 4-5 times as fast as the 6502 as you have to do the carry stuff on your own there too. Then again the 68000 has only a 16 bit ALU and 32 bit adds take 1.5 times as long as 16 bit. In reality we compare 0.33 mips 8 bit vs. 0.7 mips 32 bit so that the Amiga's 68000 is roughly 8 times as fast as the C64s 6510. — Preceding unsigned comment added by 193.102.155.35 (talk) 16:11, 11 January 2010 (UTC)

It is a hoax that the 68k was much inferior to the ARM2, the same CISC instructions cannot be compared with RISC, since CISC can complete operations in fewer clock cycles than RISC

Synthetic comparison: ARMv2 is 0.5MIPS / MHz; 68k is 0.2MIPS / MHz. Closest comparison to the real world: ARMv2 is approximately 300 Dhrystones / MHz while 68k is approximately 250 Dhrystones / MHz. About a 20% more, better but not a lot.

In the Amiga 500, the cpu power was secondary, the important thing was its custom chips that made the system look smooth and even with several simultaneous programs, I do not remember another that from 1985-1990 was comparable in graphics and sound capabilities in same price range, the Atari ST being somewhat more economical did not have blitter or sampled sound, etc. and much less Acorn. — Preceding unsigned comment added by Hooankee (talk • contribs) 04:07, 30 April 2021 (UTC)

---

We had access to a fun beast which was an Archimedes on a ISA card which we installed inside a 30386 PC; they shared the keyboard and display IIRC. The main idea was that we could run our Smalltalk port on the Archimedes. It soon became obvious that the 25MHz PC was being absolutely crapped on (in a left-behind-eating-its-dust kind of way) by the 8MHz Archimedes, with better graphics to boot. Phil 16:02, Nov 27, 2003 (UTC)

---

I believe initially that Acorn set up a lab out on the West Coast of America to develop an operating system for the Archimedes. However, I this effort foundered and the abominable Arthur was hacked up until something decent could be developed. RiscOS (2 ?) was then developed to replace Arthur, and provided many of the features seen in modern GUI systems today. In some cases, such as saving files using drag and drop, RiscOS was better than even current systems. Jonathan

Yes, see ARX. I wonder what happened to it. It's a pity Acorn didn't dump Arthur for it, for as good as RISC OS was, an OS on par with MacOSX in the 1980s would've been quite a selling point. Crusadeonilliteracy

---

A4 - 2 MB RAM (A5000 hardware in a laptop case)

This page states that the A5000 was an A4 in desktop form, not the other around around. Which is correct? Crusadeonilliteracy 03:54, 31 Jan 2004 (UTC)

---

The story about archimedes being ditched in favor of win '95 is kinda pov, and sort of wrong too I think. Archimedes was obsoleted by the RISC PC (which is basically a much faster dual-processor Archimedes)) from the same company. I'm pretty sure Win '95 wasn't the major factor. (My poor wrists still remember Win '95 :-/ ). [For completeness: PCs were cheaper than RISC PCs though, so some people may have (unwisely :-P ) followed that 'upgrade' path ;-) ] Kim Bruning 10:55, 23 Sep 2004 (UTC)

---

Without resurrecting the infinite number of "my computer's better than yours" battles from the 1980s, comparing the speeds of main CPUs when contending for "most powerful" system is simply nonsense. On many 80s machines, the speed of the main CPU was irrelevant, as it did no work other than to orchestrate all the other chips, which greatly outpowered the main CPU. If you compare Amiga and Arc games, you'll see that scrolling/sprite games are generally slow or very slow on the Arc, as the ARM is not designed for blitting and scrolling fullscreen games and its VIDC is no help. However, CPU intensive games (such as 3D games) perform far better on the Arc than on the Amiga. Kyz 02:20, 26 Sep 2004 (UTC)

Indeed, the basic Lander demo blew people away at the time.

Actually, MEMC and VIDC do provide support for hardware scrolling the screen. The start, end, and start-after-VBL addresses for the video buffer can be programmed in the MEMC at addresses that are multiples of 16. This easily enables vertical hardware scrolling. The 16-byte minimum increment is too much for direct horizontal scrolling, but the VIDC allows programming the start and end of the graphics and the colored border part of a scan line independently at pixel resolution. If border and graphics overlap, the border overrides the graphics display. This can be used to create a graphics screen that is slightly wider than displayed and use the border to mask off columns of pixels on the left and right sides (at the price of having an always black border). MEMC 16-byte increments can be used every couple of frames to jump the picture horizontally and VIDC scan line shifting/masking can be used to smoothly scroll the frames in between.

What the Archimedes lacks, compared to the Amiga and Atari ST, is separate bit planes. All bits for a pixel (1, 2, 4, or 8) are contiguous in memory. This makes impossible the popular trick on the other machines of using some bit planes for background, some for foreground, and being able to redraw these independently. MEMC and VIDC also support a single hardware sprite, but that is intended for the mouse cursor and too limited for much else. IIRC, the hardware does not provide a horizontal raster interrupt (and nothing like the Amiga's Copper co-processor, of course), but an IOC timer-triggered interrupt can be efficiently used for the same purpose.

—Naddy 01:06, 14 October 2005 (UTC)

The Amiga had a dual playfield mode, where you could independently scroll the two layers. But the bit planes sucked. For N planes that you want to affect, you have to do N blits. That means N times the setup overhead. It's less efficient than packed pixel. Mirror Vax 08:07, 14 October 2005 (UTC)

That is not fully true. You could blit all planes at once on an Amiga if bitplanes are interleaved linewise. Which can be realized easily with an additional line offset. Bitplanes just became less efficient when it came to per-pixel texture calculations like in DOOM. (Philipp, unregistered) — Preceding unsigned comment added by 213.218.5.34 (talk) 11:34, 3 September 2015 (UTC)

Back in the days I switched from Amiga 1200 stock to A3010 and boy did I miss my Amiga. I mean what's the point in having such a powerful computer when you just do not notice the speed? Games? Amiga had better ones. OS? Preemptive multitasking made a huge difference for me. Then again the Archimedes had tons of fantastic software and it was a dream to program. — Preceding unsigned comment added by 193.102.155.35 (talk) 16:11, 11 January 2010 (UTC)

Much too long

This article is too long.

In August 2022 a bot put a notice to that effect at the top of the article. Then there were 22,622 words. Now (25 January 2024) the article has 24,488 words.

Wikipedia:Article size § Size guideline recommends that when the number of words is >8,000 thought should be given to the size. An article with >15,000 words "Almost certainly should be divided or trimmed".

Alongside this is giving consideration to what is actually making the article so long. Wikipedia:Too much detail informs us that "Wikipedia is not supposed to be a collection of every single fact about a subject".

One contender is taking the A3000 section together with the A3010, A3020, A4000 section into a new article as because they were similar and for the same market.

The section RISC OS only needs to be 2 or 3 paragraphs to briefly summarise as there are already articles RISC OS and List of RISC OS bundled applications. In its present form it has much duplication.

It is hard to make an article shorter. I have various models in this product range so I am interested. But "Articles over a certain size may not cover their topic in a way that is easy to find or read" in Wikipedia:Summary style § Article size is relevant here. BlueWren0123 (talk) 23:50, 25 January 2024 (UTC)

It was actually @Millstream3 who added the maintenance tag in this edit. As I noted in the exchange with you on my talk page, one of the challenges is to unbundle sections of this page into other articles, but I do not want to get into a notability dispute with people about whether they think the A3000 series is worth another article, only to see it get deleted. Moreover, separate articles seem to suffer from duplication and maintenance problems, such as the Risc PC article which has non-subjective quality issues.

One thing I particularly want to emphasise is that a lot of the information I have added relies on its context. The RISC OS section is meant to communicate the software situation from the point in time that RISC OS, as opposed to Arthur, was first delivered for the Archimedes. Moving that content to the History of RISC OS article, whilst avoiding the wrath of the clique who continue to insist that RISC OS is some kind of modern contender, risks losing that context and thus putting the content at risk of degradation by being deemed off-topic in its new context.

I wouldn't mind, but the History of RISC OS article exhibits some of the same copy-pasting from the inaccurate history previously found on the Acorn Computers article, along with copious trivia: Element 14 being named after silicon - no, really?! I don't want to have to follow my contributions around Wikipedia and be saddled with fixing up yet more articles that people scrawled their recollections into fifteen or so years ago.

By the way, I am not listing "every fact" about this topic. But I am trying to give the treatment of the topic sufficient depth so that people do not keep trotting out myths and misunderstandings about what they thought happened thirty or so years ago. To take an example that I noted in my talk page discussion, people have had wild ideas about the performance of the Archimedes and had seemingly endless arguments about what "MIPS" means, all of which was put to bed by just reading a couple of sources and digesting what Acorn had claimed, which had nothing to do with what a lot of those arguing believed it did. I don't doubt that the discussion on performance can be condensed, but, sure, we could certainly trim all of that away and let people start arguing again in all their ignorance.

I don't know, really. People seem to want short and simple narratives these days, typified by the one continually doing the rounds about how Steve and Sophie designed the ARM (which random YouTube commenters take to mean that they "invented RISC") and then the eventual glory and market success was simply inevitable (it's in everybody's phone, those commenters have to point out), when a combination of luck and outside expertise actually intervened to stop the whole thing from slowly fading away or at least being substantially less successful. When the actual story is more interesting, why wouldn't anyone want to hear that instead? Sometimes, I really do wonder whether people want to learn about these topics or whether they just want their existing perspectives reinforced. I don't contribute to Wikipedia in order to facilitate the latter.

Having written of all this, it might be constructive to consider other articles that supposedly demonstrate a better division of content. I noted that there are articles describing other computing systems that could indicate a formula to adopt here. However, finding a good one is difficult: Atari ST does not have enough content, Macintosh-related articles have been sliced and diced to absurd levels, the Apple II article links out to all the different models but fails to cover the software platform in any detail; maybe the Commodore Amiga and Commodore 64 articles show a reasonable approach. PaulBoddie (talk) 17:29, 2 February 2024 (UTC)