Raspberry Pi 500 and monitor arrive in time for Christmas The Raspberry Pi 500 and monitor have been released just in time for Christmas. The machine itself is an undeniably impressive bit of hardware, but the monitor is slightly puzzling. The Reg was provided a Pi 500 before its release and used it as a daily driver for a few days. It works well and runs swiftly – subjectively …
I love the "confusion" it causes in the US when using a UK keyboard.
Some guy talking about some C code: "So, we pound-include the header file".
Me: "You mean hash-include?"
Guy: "No, you use the pound symbol".
Me: "You mean this one (shows £ symbol) ?"
Guy: "No, you use #".
Me: "Like I said, the hash key".
I don't even bother to move on to the the € that I also have...
More confusion arrises with Twitter/X, where people are happy to use # to "hash-tag"...
Having worked as a field engineer and moved around a lot of sites in a lot of countries with people from various different nationalities...you get used to different logical layouts pretty quickly because most international layouts are based on the physical ISO layout...not many nationalities use the ANSI layout (which is US)...technically speaking, ANSI is the weird one...it's objectively shit as well because the pipe/backslash button is in an annoying place...it's like an afterthought.
Also, in most national variants, the more-than and less-than signs are on the key between left-shift and the letter key to its right (often Z, but maybe Q or Y) that is missing from the USA variant.
I am curious about one thing: Why did they mess with symbols on shift+6, 7, 8, 9 and 0 compared to how they were originally paired in ASCII?
omg another member of the number line gang...those numpad users are fucking weird, am I right?
You know the best way to fuck with a numpad user? They think they're used to it because of phones etc...point out that the numpad layout is not the same as a phone numpad layout...make them actually look at it.
It's like trying to consciously remember how to piss without letting your body do it automatically. Some people really struggle to piss if you tell them to consciously make themselves piss. It's the same with numpad users when you make them conscious of the fact that a keyboard numpad is not the same as a phone numpad their brain just fails.
I know that by simply posting this, I've just ruined the day for some people.
Numpad users, you're lying to yourselves...break free. Use the number line.
I don't think most people that can touch type formally learned how to touch type. I didn't...I could just touch type one day...and yes, all 5 fingers. 10 if I use both hands. Crossed over if I wish and on two keyboards at the same time if I'm being fancy.
You can spot the ones that had lessons though, they all do the wrist flick before placing their fingers on the home row.
It's very rare for me to struggle with a different keyboard layout unless it is dramatically different like ANSI (on ANSI, for a short while, I'll miss the return key by hitting the edge for example).
International layouts aren't a problem as long as I've had a chance to see the keys first. For example, if you sat me in front of an AZERTY keyboard in a dark room where I have no chance to see the keys before I type...I probably won't be able to identify it quickly or click my brain into place to be able to touch type on it...however, if I sit in front of it, look at it and realise "oh it's a French keyboard", I'll be fine.
The one international layout that made me pause for a second was the South African layout...it's an ISO keyboard but with "kind of" US layout...i.e. @,#, pipe etc are in the ANSI locations but the return key etc are ISO sized and spaced.
It's like they smashed ISO and ANSI together.
Years ago we had the local Apple dealership. We had a customer with a French keyboard. IIRC we also had customers with German and a couple of other languages.
The keyboard was just the start of it, the menus were all localised as well - lets see, though it's been decades, I still vaguely recall that save was fichier, and print was imprimateur (or something like that). German was fun as the menus expanded so much in width that they could go off the end of a small screen. We got CD's from Apple with all the localised versions of Mac OS on - so sometimes we'd try them out ... just because they were there. Japanese was cute, they replaced the multi-colour (if you had a colour monitor) Apple logo with a Mount Fuji.
The fun bit was seeing where localisation didn't work. It seemed that most of them had most of the menu items in the localised language, but with odd US-English ones. This was most pronounced in kana based languages like Japanese, where you have a menu full of what looked to us like squiggles with a few English words tossed in and looking very out of place.
And one place I later worked at, we did actually buy a language kit for one user who sometimes worked in (IIRC) Japanese - it came with a sheet of sticky labels to add to the standard keyboard.
I've been using the US International layout for years because it's actually better than AZERTY to type in French. Even when my keyboard is physically French.
Fun fact: many French people wrongly believe that capital letters shouldn't be accented because the AZERTY layout prevents typing them.
(in case some French readers doubt they should be: https://www.dictionnaire-academie.fr/article/QDL005 )
I dunno what I'd do without the .XCompose file. Though I still have to keep adding to it because I can never remember so many of the existing sequences; and even when I can, some of them seem to be particularly awkward and/or prone to typos: I'm glad I rarely have to enter anything in Portuguese, ã is particularly irritating and if I'll always mistype part of it unless I look at the keyboard in the process (yeah, I'm still a terrible typist at least 45 years after first experimenting with a typewriter; partly the same reason I play bass instead of guitar too, it's harder to get the wrong string, though I still manage).
ã is particularly irritating
On Windows, of course, Numlock + Alt + 0227... 'ã'. I am enough of a saddo that I have several useful ASCII codes memorized for this method. Colleagues at work used to think I possessed powers of witchcraft, entering the occasional ® (0174) or ™ (0153) with barely a pause in my typing rate!
The icon is a cerveja, just for you.
I've never really used PCs except for gaming. It took me a long time to even realise what the VT220-style compose key did: for many years I'd press it by accident when I missed the space bar and then it'd seemingly lock up the keyboard for the next couple of keypresses. I eventually discovered it was configurable in the terminal settings so redefined it as escape, though that had much the same problem.
But I was also "that person" who took an age to wean off ed and use vi. I was originally taught to use emacs but the reams of documentation didn't highlight the actually useful bits like global search and replace, and ed was basic enough that it was easy to find stuff like that.
Now the only problem is finding all those 8-bit characters and their "what code page is it?" malarky typically display as something weird on UTF-friendly applications so I'm constantly having to update them. And then I get to play the "so what's it supposed to be?" game... D:
You can get used to AZERTY just like you did when you started with QWERTY, however it's almost everything else about the keyboard that's utterly insane. The French government acknowledged this and has come supported the new AZERTY which tries to be logical, but I've not seen it anywhere - and plenty of people who buy blingy keyboards by QWERTY imports...
https://norme-azerty.fr/en/
All the computers I encounter except mine are AZERTY and it is infuriating to say the least. All my personal kit is QWERTY/ISO/UK so I'm perfectly happy with it using family computers is painful - even for them and they are French....
The reason why USians call it "Pound" is back in the day of 7 bit ASCII, on many serial terminals (and printers) sold in the UK, there was a switch or a menu setting that changed the "#" for "£" on both the keyboard and the screen. This was so that currency could be displayed correctly anywhere that used the pound as their currency.
So when reading something like UNIX shell, the comments would look like
£ This is a comment
So Americans used to think "£" and "#" were the same thing.
Other countries tended to do very different things, as they often had more differences in character sets than just the currency symbols.
This all started to go away when extended 8 bit ASCII started being used, with the upper half of the character set containing national characters for a country, and completely left us when multibyte character sets like UTF-8 came along.
Strictly speaking there's no such thing as 8-bit ASCII, by definition ASCII is a 7-bit code.
IIRC it's officially not allowed to use $£ and similar symbols to represent financial information because they do change/move between variants of code tables, and so might be misleading when displayed or printed on different devices. It's why banks always use standard abbreviations like USD and GBP.
I sort-of agree, but what I called it was "extended 8-bit ASCII", and "extended ASCII" is a well used term for 8-bit character sets where the lower 128 characters are the normal 7-bit ASCII control/print characters, and the upper 128 are used for other characters, including those for many different countries. The upper page is nearly always different for different countries, and I know that IBM defined multiple 8-bit code pages for different countries. They used to ship a booklet with many of their systems that listed them all!
I must admit that I did not know about banks not using currency symbols. I guess that must be for international transactions, because I'm pretty certain that my bank uses £ on my bank statements, and I don't recall anybody telling me that when I worked at one of the big UK banks (although I was in IT, so it did not really come under my remit).
Yeah, it's the likes of Dow Jones that use the three-letter symbols rather than Lloyds high-street banking. I hadn't really thought about it back in the day, but I'd no idea what half the relevant symbols actually are (well, probably much more than half; I see one every now and then and I'm all "oh, that's what it looks like."
Isn't the problem more that £ could be British pound, Egyptian pound, or historically Cyprus Pound, Italian Lira[1], or Irish Punt? And $ could be any number of different currencies, not all dollars eg Brazilian Real or Mexican Peso.
[1] Lira was actually ₤, but I suspect a lot of people used £ for it.
Not sure this is strictly true, my understanding was that "#" was used for pounds (weight) in the US and OED seems to support a use in 1923 https://www.oed.com/dictionary/pound-sign_n?tl=true
"2. U.S. The symbol # [...] 1923 Special Signs and Characters..#..Number or pound sign"
Which is pre-ASCII (although doesn't attest a relation to weight). ASCII did develop from Teletype terminals, but you wouldn't think there was enough interchange of equipment in that era for the #/£ keyboard key to be the origin (and of course screens weren't a thing at that point either).
What does always cause mild irritation is people (including people in computing) calling "#" in code a "hashtag", c.f. twitter. What do you call "#something" if "#" is a hashtag? A hashtagtag?
Quite a few terminals had the same firmware for US and UK, as they were so similar. As most ended up in the US, they were pre-configured for the US market. It's only if you were in the UK that these switches ever needed changing, so I'm pretty certain that most Americans would not have even known that this could be done.
I had been told about some countries using # for pound weight, but I've never heard of it being used that way in the UK.In my experience, we always postfix a weight with "lb" to indicate pound (something that someone, probably jake) explained as being due to something the Romans did).
What does always cause mild irritation is people (including people in computing) calling "#" in code a "hashtag", c.f. twitter. What do you call "#something" if "#" is a hashtag? A hashtagtag?
If you're feeling irritated, try to do something to relax. I'd suggest listening to some nice soothing music, perhaps in the key of G #minor
"The reason why USians call it "Pound" is back in the day of 7 bit ASCII, on many serial terminals (and printers) sold in the UK, there was a switch or a menu setting that changed the "#" for "£" on both the keyboard and the screen."
Not quite. It goes back farther, at least as far as when someone needed non-digit inputs to a phone and decided to put a # on one of them. The United States called and still calls that the "pound key", in the same way that a lot of other countries call it the "hash key", and Canadians call it the "number key". Can I explain why those names got chosen? No, I can't. It does predate a lot of terminals and computer usage, though, suggesting that the names for the symbol might have come before both the phone and the computer usage.
The # sign was originally the abbreviation of libre pondo, lb, (which for some reason had the p upside down) and was often written as the ligature ℔, often with an extra flourish across the top. It was later simplified into the # we now know and pound repeatedly while trying to get through automated phone helpline menus.
My discrete structures (comp sci) prof actually spent a class period on the many (65+) names of "#" and something like 18-20 different theories on where it came from.
He actually counted off on exams if you called it something other than "octothorpe"
Weird bloke. But then he did teach discrete structures.
"Hashbang" was the mostly-spoken way of referring to the sequence #! which used to be used as the first two characters in Unix/Linux shell scripts, as in:
#! /bin/sh
The sequence told the shell which program to feed the second and succeeding lines to. This works with any program which uses # as a comment delimiter.
IIRC it was BT/GPO usage. When DTMF keypads were introduced and phones acquired additional "star services" like short codes and voicemail they didn't want to follow US usage of "pound" since it clearly wasn't a £, and the official "octothorpe" was just too much of a mouthful.
Computers weren't common enough for people to know "hash", and someone decided that it looked like a gate.
I blame AT&T. The British Pound was first. The symbol had been established long before telephony came along. There's probably an interesting history on who's to blame.
Nearly since birth, I've used a DTMF touch-tone phone. Every voice prompt in every PBX or switch I've interacted with has brainwashed me press the STAR key and the POUND key. Was this not the case with DTMF in other English-speaking countries? Did anyone hear "followed by the hash key" or "followed by asterisk?"
For me, the move to computers changed star to asterisk on occasion, but pound remained pound until the 2010's and now we have both. These days, I'm still going with pound unless referring to social media. It rolls off the tongue easier than "hash tag." When I hear "hash" out of context, I'm thinking of a breakfast that includes potatoes.
I've been looking for a new keyboard recently, and thought I had found a decent one. German made, so the engineering quality should be good, and described as a UK layout. No pound (currency) visible anywhere on it. Shift-3 was the hash symbol.
I also have a laptop that has the pound symbol as shift-3, but no backslash (useful for Windows directories) or pipe symbol (which is really quite a useful key under Linux) visible. It turns out you can type them, but you need to know which meta key to use with which unmarked key.
Back in the 80s we received a delivery of DEC VT220 terminals, fitted with a UK keyboard by default because we hadn't specified anything in particular.
They not only had a £ symbols, but { and } were replaced by ½ and ¼. Not so great for C programmers...
It must have been a common issue, DEC supplied a conversion kit consisting of bags of replacement keycaps.
It's a computer that's built into a keyboard. So if you want to talk about perverse layouts:
The Commodore 64 has only two cursor keys, one for horizontal motion and one for vertical. Use the shift key to pick left from right and up from down. Which you might or might not think is worse than machines like the original ZX Spectrum where cursor keys are a shift combination with other keys. Either way be careful when typing if you have an Acorn Electron: the break key, which resets the machine, is right next to right cursor and not in any other way differentiated from the rest of the keys. It feels the same, presses the same, etc.
The original ZX Spectrum also distinguishes itself via a space key. Not bar. Key. The Oric 1 manages a bar, but all of its keys end up being bars because they're very short but a normal width. Get the Atmos if you want an Oric.
Also probably to avoid: the original Commodore PET was ortholinear. Endless machines are completely flat membranes, but at least we're all used to that now with phones.
In the early days, cursor keys were optional. That's why the semi-standard "Wordstar cursor diamond" existed using Ctrl S/D/E/X (and associated keys for word right/left and page (or screen) up/down. So, to be fair, those early 8-bit computers actually having cursor keys at all was a bonus. Choosing how to lay them out was an exercise for the keyboard designers, who were not really thinking about keyboard control for gamers. In fact, not all even did full screen editing, so even less need for efficient and ergonomic cursor key arrangements. Personally, I grew up with a TRS-80 clone so up/down under the left hand and left/right under the right hand and either thumb for Space Bar (fire!!) was perfect for me and the soon to be "standard" arrow cluster (of which there were variations too in the early days) seems rather clunky and silly for gaming (Joysticks were not only optional, some games just didn't support them at all unless they were wired across the relevant keys of the keyboard matrix)
And not to mention that teletypes had no need for a cursor at all, never mind keys to control it :-)
I hate flat keyboards. I'm not even that keen on stepped keyboards and prefer the C64's curved profile, though I wasn't so fond of its mushy (or dampened) feel. Probably why I liked the Model M so much when I first encountered one. Which is, ironically, a membrane keyboard, just with a fancy mechanism on top.
One thing I remember from the early '80s computer boom that I haven't seen often enough (in fact I don't recall seeing it anywhere else) is the Atari 800 which protected the reset key with an immovable bar across the top and bottom to lessen the risk of it being pressed unintentionally: only the middle part moved. Dunno why that sort of thing wasn't more commonplace; perhaps fingernails (I always kept mine short) but more likely the manufacturers being tight-arses. DEC's approach was (eventually) to move it out of the way at the end of the function-keys; IIRC the key next to it didn't do anything particularly useful so it wasn't likely to be pressed by mistake.
"The Commodore 64 has only two cursor keys, one for horizontal motion and one for vertical. Use the shift key to pick left from right and up from down."
The Acorn Atom did that too, but put them at the left hand end of the keyboard instead of the right.
Also like the Electron, Break (reset) is just another key on the Atom keyboard, next to up arrow (its version of caret) and above delete. Quite easy to delete rather more than you were expecting, although fortunately the OLD command gets your program back if you do press it.
Yes, typing on the Oric-1 wasn't a particularly pleasant experience with that keyboard.
We all get used to a layout. I don't think there's a better or a worse one, just a normal one and the one that you keep making mistakes with because the keys are in the wrong places. For example, if you're used to an ANSI layout*, you might get used to a longer shift key and dislike the little key they shoved in between the shift and Z key which you keep hitting when you want to capitalize something, whereas if you used the ISO layout and were now given an ANSI one, every time you try to run a command and hit something other than enter because they made it thinner would get on your nerves.
* For once, this isn't a US versus everyone else thing. ANSI keyboards, what has been called the US keyboard layout for this thread, are commonly used in a lot of countries, not just the US. ISO is common in many European countries, but most other continents have more ANSI keyboards.
There was a time when the US got it right - Classic Mac, Amiga, and Atari ST keyboards, but ANSI decided to standardise on IBM.
All non-ANSI layouts have a vertical rectangle or a backwards L for Return which is fine. The ANSI layout has a horizontal rectangle for Return which is unusable because you can't just press the top of Return with your little finger. It is known.
They're on The Pi Hut, where happily you can also see the price in £.
Seriously, El Reg can't even include the price in both currencies?
I had a quick look, they are listed on the site but appear to not be available for purchase/pre-purchase, they are "coming soon"...
I seriously hope that other keyboard layouts will become available since I use neither UK or US.... Here in mainland Europe there is a plethora of layouts...
I imagine they will cater for other layouts just as they have done for the Pi 400 and the Pi standalone keyboard.
That would go against the nostalgia involved in having a keyboard attached to the computer. Maybe there's a benefit other than that, but I don't really know what it is. One fewer cable doesn't worry me very much, and it lets me have a keyboard of my choice. Still, they are selling enough of these that they made another version, so people must like having the keyboard and computer integrated.
In a case that you put the main board in? That's at least what I've done. I have too many USB keyboards to justify buying another one that's probably of similar quality to the ones I got for free. Clearly, some people find value in having them integrated that isn't making sense for me.
This post has been deleted by its author
Entitlement?
That's hardly fair.
He is thinking different priorities from the Raspberry folks, that's all.
Raspberry Pis have been cheap because they're looking to served the widest base market, and that's the SD card crowd.
However, many hardcore Pi users are now are now the sort of people who add NVME. An NVME adapter is always an added aftermarket cost, and fitting it in the base model would be cheaper as it would leverage economies of scale. But it would make it more expensive for the core market, even if only marginally. Raspberry don't want to do this.
They have three audience groups now:
The kids' toy crowd
The maker crowd
The not-really-maker crowd who want to have a device that they can do interesting stuff on without fighting with either hardware or software.
I suspect that last one isn't particularly profitable....
You missed an audience group.
Industry.
Pi sell over 70% of Pis to industry or commercial customers. Which is good, because those sales means the prices can be kept down for the other groups you mention (everyone pays the same price, which is often forgotten).
The Pi 400/500 are great thin clients. Do they need NVME? They are great replacement desktops for people who just use a browser to access the cloud? Does that need NVME?
They are great replacement desktops for people who just use a browser to access the cloud? Does that need an NVME?
Are you saying that is all they can be?
You can see on the official site the amount of clamour for a M.2 supporting device and the number of people that having got wind that the PCB is laid for one are choosing to wait.
This may change. The large system model of software is proving to be an expensive failure when combined with complex systems and EOL issues with hardware and OSs. There is the opportunity for a new, simpler OS that simply lasts, offline, without updates, on cheap kit, with modular software, to replace PCs and the sort of the stuff Oracle are trying to implement for Birmingham council. These machines would only do what their users needed them to do, 80s/90s style. It would reduce public services' exposure to the shockingly expensive private sector contractors and outsourcers.
That said, I would like to know why there are so many 'Opened but never used' Pi 400s on ebay. And why these machines are not being sold in high street stores to give them more penetration amongst domestic users. Things won't change if you hide them.
I would also like to see a simpler OS that would allow more homebrew software.
"That said, I would like to know why there are so many 'Opened but never used' Pi 400s on ebay."
Probably a combination of two groups: A) people assuming that they can buy anything with the Raspberry Pi label on it and resell it at a massive markup. It worked for Zeros and during the pandemic, but some of them are going to be disappointed and I would expect those trying it with the 400 to be some of those people. B) people who thought the 400 was cool, but eventually decided that it wasn't all that different from a normal Pi with a keyboard plugged in.
"And why these machines are not being sold in high street stores to give them more penetration amongst domestic users."
The stores don't want to try to explain to their employees, let alone their customers, what Linux is and how they're going to have to image an SD card to make this computer turn on. Raspberry Pi doesn't want to try to convince stores to sell these with the typical retail markup when they can already sell them online.
"I would also like to see a simpler OS that would allow more homebrew software."
What about Linux is not allowing for more homebrew software? I can at least understand what a simpler OS would be, though not why you want it, but I'm not sure what is reducing the allowance for user-written software.
The Debian "base stack" is pretty complicated these days, what with systemd & other things.
Have you considered a more streamlined base OS, something like Alpine Linux, with Busybox instead of the GNU utils (they can still be installed!) and MUSL libc instead of glibc? As a Debian user for three decades I am very impressed with Alpine, and I keep it to hand on an SD card for experimenting.
This approach might be more appropriate to the "maker" use case, and more maintainable for you guys in the longer term.
Just a suggestion . . .
"Have you considered a more streamlined base OS, something like Alpine Linux"
Alpine has already been available for Raspberry Pis for some time, I use it (as a server OS) on all my RPIs.
I haven't yet personally tried using Alpine as a desktop OS but intend to try it on a RPI 400 in the not-to-distant future. Obviously Alpine doesn't (currently) have the RaspPi OS specific desktop changes but those patches/changes could be applied on Alpine.
Why does this matter for most programs? Most of the time, you're not plugging your program into the init stack, which is why most programs run identically if there's systemd or not. In those exceptions, it's usually because your program is running as a service, so you need two service definition files, one for each system, and that takes about ten minutes. Only a few programs are going to interact with the init system in a complex enough way that the choice of one would directly impact the program. If that's your reason why homebrew software is being disallowed, I think your idea of homebrew software is probably not the most accurate.
"So you expect RPi to procure the extra components, integrate them into the design, and drop the price?"
Is that what they said? I don't see that in their comment. What I see is a suggestion that they consider adding M.2 support, potentially increasing the price, in order to have a better product. It would cost slightly more to have that on the board, but at the scale they're producing them, not a lot more. The chips they're using already support NVME, so the parts that would need to be added aren't very disruptive or expensive. They clearly believe that the enhancement would be worth the cost.
And that is not entitlement. That is a preference. They're expressing their opinion that the 500 is not as good as it easily could have been. You are free to express your preference that the additional cost would not justify the NVME support, and thus that the 500 as it exists is better than their suggestion. Then you could debate how much adding the hardware would have costed and how many people would have benefited from it, and probably at the end go away with the same opinions. None of that is a sense of entitlement.
The additional cost for the NVME components is actually fairly high. Its not just the board components, but a new casing with hatch, because splitting the case to fit the drive is a no-no for a consumer product.
Of course, there is no need for the NVME drive for many many use cases, the faster SD card capabilities of the 2712 SoC when combined with command queueing means that even with an SD card the device is very useable, and for most people good enough.
That may be true. As someone without a strong preference as I'm not likely to buy a 500, I would encourage the two who do have strong and opposing preferences to debate that. I can see why some people want the NVME capability, because a 500 is of most use as a desktop, and a desktop is also where you'll notice the slower speeds of SD cards most often. However, a lot of the people buying this are specifically going for a cheap desktop, so if NVME is that expensive to add, it would put some of them off. Without knowing the price, I could not even begin to estimate how large that tradeoff would be.
I do have to ask for more details about some of the basis for the higher costs you expect. Yes, the case would need to have two parts so you didn't have to break it to install the drive. You wouldn't need a hinge to manage that, assuming that's what you meant by "hatch". The 2712 already has support for NVME. Other than the case changes and the board components, are there other parts that I'm missing? And if those are the only two components, is the case really the more expensive of the two as your comment implies? At the scale of the production runs the 500 will get, how much more expensive do you think a two-part case would be? From the sound of it, the case already got redesigned, so it's not like they got to reuse 400 cases anyway, so we're mostly considering manufacturing cost differences. I'd have been less surprised if it turned out that the M.2 required some expensive component I didn't know about, but I didn't expect a case hole to be the biggest issue here. I have no manufacturing experience, so maybe I'm clueless about the difficulty of plastic cases. It is an interesting thing to learn about.
Yes, I do remember netbooks. I remember how they died, even though I liked them, because some users refused to buy them because they were underpowered. That's exactly why you can't only go with cheapness; at some point something becomes too underpowered to be in demand. Netbooks could have stayed a niche product and tried to appeal only to enthusiasts like me, but they tried for a mass market approach without the hardware to match and a lot of people who bought one tried to do things the hardware was not capable of.
I don't know that M.2 is that for the 500. I'm not planning to buy one anyway, since although I have a bunch of Pis, I don't use many as desktops and when I do, I am fine connecting a keyboard of my own to it. I could see why someone might think it was, since that allows for significant speed increases over an SD card, and when I have seen people use Pis as desktops, speed is often something they comment on if they use it for long enough. However, most people using a Pi as a desktop are in the enthusiast category which should be more aware of the limitations and more willing and able to work around them, so that might not be as big an issue. It would probably limit people trying to sell them to the general uninterested user, but I don't think that's likely to work in the first place so it may not be a good reason to change the hardware. As I said in my original comment, I did not come to argue one of these sides. I came to explain how your disagreement about what was best was not entitlement on the part of someone preferring NVME. Those who have a stronger preference could discuss this with the reasons behind their preferences.
I got downvoted to fuck for saying touch screens were a good invention.
Apparently an adaptable, variable user interface that accepts touch input isn't a useful thing.
Like mobile phones would have been popular with a keyboard and mouse....
Form factor is relevant here.
I have a touch screen on my laptop, never, ever used it, because there's no point, the mouse & keyboard do everything a touch screen interface can, without getting grease marks all over the screen...
I'm retired and have been doing some pro bono work on setting up Pi 5s for use by retirees. Task analysis shows that they never need/use LibreOffice - All they want/use is "in the Cloud" - Google/Docs, MS/365, banking, medical appointments, etc., with a browser and printing. We set up various combinations of RAM, disk, and Pi 5 cases. There was little/no difference in our perceived performance between 4GB and 8GB of memory.
It's summer here, and we noticed that the RP fan was often on at logged-out base-load with the RP official case/heat-sink/fan. This did not improve with the Active Cooler. In the end we found the FLIRC passive case worked well, but could feel hot under heavy load (Chromium and Firefox both open, each with 10+ tabs open with a lot of forced updates, while watching/listening to a 2K YouTube video). As old people's skin can be easily damaged, I added "temp_limit=60" to the boot config file. There was little difference in performance under stress and the case felt cooler (the CPU dropped from 2.4 to 1.8GHz).
We standardised on Chromium and basics like Mousepad, calculator, the viewers, and File Manager for the "user" account; with management and checking software for the "admin" account. When the user logs out the system is reset. This causes about 200MB of disk writes, which initially concerned me - I considered using RAM storage to ameliorate this, but in the end went with a slightly more expensive but slower 64GB High Endurance MicroSD card (expected lifetime 5+ years?) without a measurable performance hit.
In summary: The performance difference was slight but measurable between the M2 SSD and MicroSD setups - opening Chromium took 4 seconds instead of 3, with a slight improvement in responsiveness with more than 10 "busy" tabs open. If this carries over to the 500, there is probably a cost saving as the keyboard is included - At the moment the Pi monitor looks a bit expensive as a local price for a normal 24" screen is ~AU$20 less
The price is pretty good considering there is a monitor.
But then I remember that a second hand ThinkPad is ~£40.
And then I look behind me at my ridiculous stack of ThinkPads and realize I don't need any more.
So now I can spend all my money on bills and nothing but bills. Being an adult is lame.
drinking a G&T at 10AM
I say old chap! The sun isn't even over the yardarm. Whatever this sun thing is? I think it disappeared behind clouds about 3 weeks ago...
But, as you say, adulting can be fun too. I discovered a new whisky on Saturday night. The Balvenie 12 year old American Oak - and it and I became good friends. A state I was helped to recover from by the hotel's excellent fried breakfast - and large supply of tea and danish pastries. I'm donating blood in an hour - so hopefully the alcohol is now out of my system - because we did become very good friends indeed...
There's definitely some cool potential for power bank running. A beefy dual-output power bank could run the Pi 500 and the monitor for a quite a while (probably 25W total?), and you could connect it to your phone's mobile hotspot to have a fully off-grid desktop PC solution (assuming the mobile network keeps going during a power outage, which most of it should, with base station backup).
The monitor would be very useful for the school I've worked with, which used a stack of Pi 400s. They had these horrible, ancient, donated monitors with VGA/DVI ports, and the Pi 400s never liked them, even with a variety of HDMI adapters trialled. So a cheap, compact, dedicated HDMI monitor would actually be ideal for them, even before they potentially upgrade the Pi 400s to 500s.
I'm biased, but I have been using the monitor on and off for a year, and really like it. The three mounting mechanisms work really well, and being able to power from the Pi or from a standard USB PS has made exhibition demonstrations much easier to do. I hope schools really like them.
The monitor sounds nice. I wanted a photo frame from the kitchen, and the displays on Amazon were around the same price for 7". The first one had speakers, but died after a week. The second is fine, but no speakers. Both required faffing around to get native resolution from the HDMI (the displays are basically laptop/tablet displays, and they don't have EDID on the HDMI as cost saving. Hopefully this one has
"power from the Pi"
I wonder if there'd be any market for a combined/double-headed USB-C/HDMI lead just for that use case? For some people, anything that cuts down on "cable mess" is a good thing :-)
(I suppose getting two of the relevant cables in the same length and some cable ties/spiral wind would do in a pinch)
(assuming the mobile network keeps going during a power outage, which most of it should, with base station backup).
I don't think most cellular towers have more than a few hours battery backup, I know my nearest one died the last time the power was out for half a day...
"I don't think most cellular towers have more than a few hours battery backup, I know my nearest one died the last time the power was out for half a day..."
Many mobile phone antennas these days, certainly in UK towns/cities, are provided by monopoles ("https://pedroc.co.uk/content/vodafone-o2-monopoles" and "https://pedroc.co.uk/content/ee-three-monopoles") rather than the old style towers/pylons.
These monopoles tend to be placed next to public pavements with 2 or 3 on-pavement cabinets housing the telecoms kit. Typically in the UK there seems to be only a small (1 hour max?) battery backup in the cabinets. As these "mast sites" are housed on/next to public pavements then they obviously don't have any generators to provide longer run-times.
"A beefy dual-output power bank could run the Pi 500 and the monitor for a quite a while (probably 25W total?)"
I just watched a Youtube review. 10W draw at "full"[*] brightens/volume powering the screen from the Pi500 playing a 1080p video. Powering the screen separately and ramping brightness/volume to actual full levels, playing the same video took it up to 11W. He didn't go any further in that testing, so probably nowhere near 100% CPU/GPU, but interesting nonetheless.
* It "knows" you are powering it from the Pi500 and restricts the brightness and volume settings to ~50%
You're going to need a pretty huge power bank. It's the perpetual problem of the Raspberry Pi, at least for me. I wanted to do so many things with them that involved powering them from a battery, and it always worked... for a couple hours, but my laptop worked for a lot more hours and my phone even more. There is a reason that every new Pi has come with the release of a higher current 5V power adapter. We just made the leap from the 4's 3A one to the 5's 5A one. In both cases, that's not actual power usage because those values are intended to provide lots of expansion room for peripherals. However, if you try to run a Pi 5 off a 2A adapter, chances are that it will work just fine for quite a long time until it suddenly doesn't and you have to get more power, something that is frustrating for the knowledgeable user and even more so for the user who thinks something must be broken because they don't know why the computer just rebooted itself.
A laptop with a more powerful processor might consume more power when at maximum CPU, but its CPU, unlike the Pi's, was designed to shut parts down when idle, meaning that average usage is probably lower. It also comes with a battery that's designed for that power profile, whereas the power banks you're likely to use are mostly intended for charging phones and do things like reassigning loads when multiple devices are plugged in, sometimes dropping power to all of them for a second until they've done it. The Raspberry Pi doesn't appreciate that. If you really need to run without power, you need to plan out your equipment more thoroughly.
...and therefore be chained from the Pi itself, all we need is to add PoE and what we have here is a single-cable machine. I like to imagine a room full of them netbooting a nice thin-client OS and attaching to a chonky heavy-lift compute rack in the basement. Everything old is new again. It'd be like being back at school.
I can actually think of a few applications for the monitor. Having a standalone monitor with a 5V power requirement is quite handy in labs where the weight of power requirements of a standard monitor is over the top
Not using the compute 5 module (plus lack of NVME support) seems a major mistake though, and makes this a minor upgrade rather than the major change it should be, so i will probably skip this variant
[Num Lock key is no longer, and a power key has been added – infuriatingly close to the Delete button.]
That will get annoying really fast.
Also who really uses touchscreens on a desktop? A keyboard and mouse not only work better and faster but you avoid cleaning the monitor so much so that's a plus. Any extra use you may get of a touchscreen on a desktop computer gets faced with the fact most Linux software is not made with touchscreens in mind
Yeah - loads more effort required to hold your hand up in an awkward position to smear your greasy fingerprint onto the screen. If only someone could invent some sort of pointing device that just sits conveniently on the desk, on which your hand could rest...
"Touch is great for touch-designed interfaces. Desktop UI's are not touch-designed."
Windows 8? :-D
And that Ubuntu thing they tried to use cross devices, some sort of phone/tablet interface also for the desktop?
(Although I also remember the derision heaped on Win8 for similar reasons mentioned today. Arm strain by constantly reaching out to a large desktop screen :-))
I'm so disappointed. I was sure this version would have M.2. I didn't buy the 400 because I was waiting for the 500 with M.2, and they still haven't put it in, despite the hardware supporting NVME. The cost of a connector is about $0.05, the additional passives probably about an $0.01 in the quantities they are manufacturing. Maybe less.
There is simply no excuse for not including M.2 on the board. I see the board has the layout for it, but the connector is missing. Jeff Gerling wasn't able to get it working, despite soldering a connector to the board.
Sorry, but without M.2 it's just a toy. Frankly, junk. Not buying. So disappointed. :-(
Without M.2 it's just a toy and a piece of junk? Uh-huh.
You do realize we're talking about a minimum cost SBC here, right? And that M.2 requires more than just the presence of the connector? And that the obvious excuse reason for not fitting a small, modestly spec'd SBC with top performance M.2 support is that your estimate of single dollarcent costs are (to put it kindly) naively optimistic?
> You do realize we're talking about a minimum cost SBC here, right? And that M.2 requires more than just the presence of the connector? And that the obvious excuse reason for not fitting a small, modestly spec'd SBC with top performance M.2 support is that your estimate of single dollarcent costs are (to put it kindly) naively optimistic?
Nope. *All* the support for NVME is *already* built in to the chipset. The only additional hardware required is the connector, *which is already laid out on the circuit board* (take a look at some of the reviews on YT where the reviewers open the machine up), and some passive components for noise reduction on the data lines - i.e. resistors, and capacitors. These are fractions of a cent. There's simply no excuse for not including it. The connector solder pads are already on the board, and the legend is already printed on the board for M.2, M.3 etc. So, they laid out all the tracks, laid out the connector, then didn't fit the connector. Why?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
