Just another pre-installed program
Mathematica (like Cortana in Windows) was just another pre-installed program that I didn't use. If I want it, I can download it. Until then, it doesn't pollute my Pi.
Knickers have become ever so twisty over the last few days as fans of the diminutive Raspberry Pi computer and its Raspbian operating system noted that Mathematica had been "removed". The conspiracy theories kicked off when users noted two simple words in the release notes for the latest and greatest version of Raspbian: * …
Seconded. I'm glad Mathematica remains available, but like Mike, I don't use Mathematica on Pi, so pre-installing it does not help matters. But I guess some form of comms before this change would have been much more appreciated than the post-oops scramble.
What many commenters here don't seem to realize is that the grumbling was not about removing it from the default image (so it's no longer a "pre-installed program") but removing it entirely from the Raspbian repos. For a few days, it was gone for good. It was not possible to install it *at all*. I believe that most of us who need it would be quite happy to install it manually, but that possibility was removed too.
Given how quickly Wolfram reacted when people started to complain, I wonder if the Raspberry Pi foundation talked to them at all before removing it. I understand that the contract expired, but the reasonable thing for RPi foundation would have been to ask in advance, "Hey Wolfram, are you planning to extend the contract or is this the end?". Of course, we don't know what happened behind the scenes, but the impression they're giving is that they "quietly" removed Mathematica entirely "because no one uses it anyway" without consulting Wolfram at all, and left all RPi/Mathematica users high and dry.
I always found it odd that the RPi Foundation championed open source and yet chose to bundle proprietary maths software instead of SageMath.
Yes it's a nice gesture by Wolfram, but still it seems contrary to the whole RPi ethos, a bit like the proprietary VideoCore IV graphics driver (a problem which has since been resolved, unlike the proprietary codecs problem, a.k.a. the MPEG-LA tax).
Anyway, I'm not suggesting that nobody should be allowed to run proprietary software if they want to (I do), but on the other hand there's a difference between end users choosing to do something, and an organisation founded specifically to reintroduce real computer science back into the classroom (and education in general requires open access to information) actually promoting something that restricts access to information.
/end Stallmanistic rant
Go to http://www.sagemath.org/, look at the second sentence of the first paragraph:
"It builds on top of many existing open-source packages: NumPy, SciPy, matplotlib, Sympy, Maxima, GAP, FLINT, R and many more."
This _bundling of many packages is what the Mathematica design philosophy is trying to avoid_. The creator of Mathematica, Stephen Wolfram spent more time trying to get various math packages to work together than doing actual physics.
So he designed a 'kitchen sink' math program (Mathematica) where typically you just needed one download to do everything you needed.
How this ties into the Rasberry Pi philosophy of getting people started in computers? You just need a _single_ download to do everything (or a lot) and you don't have to worry about version conflicts within/between supported math libraries.
Mathematica (can be used as 'training wheels' for math (and also now for other scientific fields))
SageMath (if a person is further along in math and wants to continue in their own very specific direction.)
TBH, there has't been that much grumbling. Engineering preference is to have it in the repos rather than installed by default, as it's a huge chunk of the image that a large percentage of users don't need. As a user of Raspbian, but non-user of Mathematica, if you don't remove it, then when an update appears its takes a load of time to download and install it.
Having it in the repo means those that want it can download it. Those that don't save 700MB of bandwidth.
Those that don't save 700MB of bandwidth.
Unless your Internet access is provided by barbarians (the sort who impose usage caps on fixed-line access, duh), that's just a question of (a sizeable amount of) time, possibly as long as (calculates) seven or eight seconds.
Yes, I'm yanking your collective chains. At 1 Gbit/sec (my ISP just raised my fibre access to 1 Gbps down / 300 Mbps up for no extra money), it is, indeed, about that. At a megabyte per second (reasonable expectation for "up to 20" ADSL2+), it's more like 12 minutes or so.
It's not just the download time. When doing an update it can take ages to install, so much so that people have thought the device had locked up.
If you add up all the extra 700MB over all the Raspbian downloads from the last 5 years, (100 million? Dunno, might be worthwhile finding out) that adds up to, er, quite a lot of wasted bandwidth.
"If you add up all the extra 700MB over all the Raspbian downloads from the last 5 years, (100 million? Dunno, might be worthwhile finding out) that adds up to, er, quite a lot of wasted bandwidth."
So why was it ever included in the download? Why was it not made an external package sooner?
It was great to have it included in the download. It is great to not have it included in the download.
Having cake and eating it is the new normal.
Now imagine downloading it over a flaky dial up connection at a miserable 300Baud via an acoustic coupler, joystick port mounted, piece of crap modem. Through AOL. During a storm that causes so much lightning strike interference that you spend +50% of the time resending line noise mangled packets.
*Shakes a palsied fist*
Danged whippersnappers anyer newfangled fiber optics. Why, back in MY day the only fiber we got was from eating granola, and that's the way we LIKED it, dagnabbit.
*Brandishes a cane menacingly*
Now get off'n my laaaaawn!
=-D
"... There were a hundred and sixty of us living in a small shoebox in the middle of the road, cueing up once a week for a chance to use the village carrier pigeon."
at least you had a shoebox, we only had the lid..... and there was 500 of us taking it in turns for the carrier snail....
Replacing it with a menu line which would then go and install it might have been nicer than just vanishing it. Or is that what they did ? I don't know.
The problem for Mathematica is that while it is a very fine piece of work and in many ways a superior product especially with symbolic, everywhere I go seems to have standardised on Matlab. So GNU Octave tends to be more wanted on Pi. And now Python with mathematical and scientific libraries is right round the corner on the road map.
...I get the arguments about pre-installed apps and bandwidth wastage. The appropriate implementation would have been a .pdf on the Desktop instructing users downloading an 'educational' image how to install Mathematica. Or, have an image for download optimized for secondary- and post-secondary education that has Mathematica installed. Problem solved.
We need to remember that a primary goal of the RPi Foundation is the creation of a computer that is cost effective enough for even extremely poor school districts throughout the world. I think the theory is that we will build global wealth and address income inequality by helping others get the tools they need to build up themselves and their societies. Will it work? Who knows ... but I sure think this is a more decent and ethical approach than hand-outs and whatnot that just build dependency and help the richer people maintain a degree of smug superiority. AND we get and outstanding piece of kit; what's not to like?
To that end Wolfram Research (Mathematica) contracted to provide free access to Mathematica. This tool is absolutely fantastic for the physical sciences, theoretical and applied mathematics, and electrical engineering. I've used it for all three and one of my organizations spends literally tens of thousands of $ annually on contracts with them and does so gladly. My first exposure was on a NeXT Workstation (remember those?) at university in 1992, at which time we were told it would revolutionize pure mathematics. It did. And now Wolfram (the man) wants to share his creation, I say let's raise a pint but also eliminate the download pain.
Sure, offering this free on educational kit is kind of like offering hits of an addictive substance ... the marketing hope is obviously that students will get hooked on Mathematica and when they are in a position to purchase software later in life will tilt the scales in Wolfram's favor. Worked for Apple back in the day. But I'm just happy if we get disadvantaged students to the point in life where they are making software decisions.
Disclaimer: Yes I'm a major fan. But I also use MathCAD, Matlab, GNU Octave, Maxima ... basically whatever tool will make a specific job easier. I am a software polygamist.
If I buy a card with 700M of something like Mathematica on it, and it ships overnight, how slow does my ISP have to be that the former gets me the software faster? All these comments about old 300 baud modems make me wonder where the crossing point is. Let's see, 700M/24h is about 8K/sec? I don't remember get much over 8K sec even in the 28.8K modem era.
This is, sadly, another example of everyone, people and companies alike, pushing down responsibility for their formerly-featured work to those using the result. It's endemic, everywhere around us.
It goes like this: at product introduction, some added feature, at some acceptable internal cost, is added. In cases like this, where a partnership is involved, the motivation is around future business. Over time, market fluctuations and normal product evolution, that value pull-through is seen more and more as an expense and less as a valuable investment. At some point, the decision is made that the feature "isn't that widely used", and "it's better separated", and responsibility is pushed to the user. For the organization, it's a simple cost-cutting event, and it works wonders. What they don't realize is that they have just spawned a required set of actions that take time and effort on the part of their entire user base. Regardless of how few a percentage of all users have to do the work, N-users minus one (the now-fired worker) now have more work. The effort spent by the whole of humanity is exploded in huge amounts by that one decision.
Sure, for this one decision it's not "that much" extra labor across the earth. But, look around you these days, and you'll see companies and people doing this everywhere, pushing responsibility and effort to the user. A "safe, clean and secure" storage company that won't pay to control the rat population so you can't store food, bedding, clothing, cardboard boxes,....? A rental-car agency that makes you clean the interior in addition to filling the gas tank? Still taking your shoes off at airports? How many shoe bombings have been thwarted? Um...likely zero? Yet the human effort expended for zero benefit continues.
I'm not saying there's a solution, I'm just saying it's getting annoying. Put the thing back in and deal with your licensing issues without wasting our time giving a d@mn about it. Did we all have to work through the license terms way-back-when? No! If we had, we'd have not set it at just 5 years, duh. So why do we have to even care now?
Just sayin....
A year and a half after the debut of the $4 RP2040-powered Raspberry Pi Pico, the company is shipping a wireless-enabled version: the $6 Pico W.
Amazon Web Services has proudly revealed that the first completely private expedition to the International Space Station carried one of its Snowcone storage appliances, and that the device worked as advertised.
The Snowcone is a rugged shoebox-sized unit packed full of disk drives – specifically 14 terabytes of solid-state disk – a pair of VCPUs and 4GB of RAM. The latter two components mean the Snowcone can run either EC2 instances or apps written with AWS’s Greengrass IoT product. In either case, the idea is that you take a Snowcone into out-of-the-way places where connectivity is limited, collect data in situ and do some pre-processing on location. Once you return to a location where bandwidth is plentiful, it's assumed you'll upload the contents of a Snowcone into AWS and do real work on it there.
The Turing Pi 2 crowdfunding campaign has soared passed its $64,000 goal in a single day, currently standing at $1,027,428 with more than 3,400 backers.
Early-bird backers will be able to build their own Arm cluster in a box for $199 (compute modules come extra). Now the entry-level pledge stands at $219 for one Turing Pi 2 Cluster Board.
We covered the Turing Pi 2 at the end of last year and at last the waiting is over – in principle at least. It has been a year and a half since Turing Machines initially announced the product, and a good nine months since it published the specifications.
The astonishing PicoPuter emulation project can run a transputer emulator on multiple Raspberry Pi Picos, and clustering them using the transputer's native inter-processor link protocol.
The Raspberry Pi Pico is a surprisingly capable device at $4 apiece, and one of its less well-known features is its eight programmable IO state-machines on board. As programmer-archaeologist Andrew Menadue wrote in a blog post:
The Raspberry Pi gang will soon ship some new hardware in a form factor to delight builders still mourning the Compute Module 3. Behold the Compute Module 4S.
The Compute Module 4 emerged in 2020 and, as well as featuring some beefier hardware courtesy of the Pi 4, also changed the form factor from the SODIMM of old. While memory could reach the dizzying heights of 8GB and dual HDMI interfaces were available, the departure from that SODIMM form factor of old was a headache for some of the industrial customers at which the unit was targeted.
Jeff Geerling blogged about the new arrival last week, although details on availability outside of Revolution Pi are scant at present.
An update to the Debian Bullseye-based Raspberry Pi OS is being rolled out with both quality-of-life improvements and one very important tweak: an overdue departure of the default user.
Previously, all installs of the Raspberry Pi OS (formerly known as Raspbian) had a default user called "pi".
This was handy for quick setups and convenience but a bit of an open goal on the security front. "This isn't that much of a weakness," insisted the Pi team, pointing out that you'd need to know the password as well to get access (and you'd need to have enabled remote access for miscreants to do their dirty work anywhere but locally) but still.
Sometimes the best things are the most simple. A case in point: sending GitHub issues to an old thermal POS printer via a Raspberry Pi.
Developer Andrew Schmelyun was faced with the thorny problem of keeping track of issues posted on his various GitHub projects. A problem familiar to many: emails tended to get lost in the mailbox or issues were simply forgotten in the face of a growing to-do list.
What to do? Schmelyun spied a receipt printer spitting out orders while he was collecting a take-away meal and pondered if the same might be implemented for GitHub issues.
It is with a sense of inevitability that we can confirm somebody has managed to make Doom work on the diminutive RP2040-based Raspberry Pi Pico microcontroller board.
Running the '90s first-person shooter game on hardware ranging from ATMs to pregnancy testers is very much a badge of honor for hardcore tinkerers and the surprise is perhaps not so much that the RP2040 hardware is up to the job, but that it has taken so long for someone to do it. After all, it is just over a year since the board first arrived.
A quick glance at online stores shows that the Pico is currently not made of unobtainium and actually in stock at outlets.
Interview Today marks 10 years since the Raspberry Pi was made available to purchase. We spoke to Pi supremo Eben Upton about the last decade and what the future might hold.
The Raspberry Pi team has elected to mark the milestone on 28 February, it being the last day of the month this year – for many 29 February 2012 was when units were first available. Looking back at 2012, the numbers look almost farcical when one considers how things turned out.
The latest update to Armbian brings a mainline-kernel based Ubuntu- and Debian-compatible environment to dozens of small single-board computers.
This includes both Arm and x86-based hardware UEFI booting – and 64-bit builds for Raspberry Pi hardware.
Armbian supports over 60 different single-board computers, including various models of Banana Pi, nVidia Jetson, Pine64 and dozens more.
Biting the hand that feeds IT © 1998–2022