La Raspberry Pi es Una Porquería
El otro día me llegó una Mele a1000 que en unos días va a tener un review completo, creo. Pero tengo que decir esto: es tan pero tan superior a una Raspberry Pi que no tiene gracia.
Pra empezar, cuesta el doble. Eso es malo, pero aún al doble de precio, es más barata. Porque si sos una persona para quien U$S35 hace diferencia, entonces tal vez tampoco tengas una tele con HDMI. Y la Mele funciona con HDMI pero tambien con RCA y VGA, o sea que cualquier monitor o televisor fabricado en los últimos 20 años te sirve. Podés recuperar monitores antiguos (¿uno que da 1024x768 fabricado en 1996? ANDA 10 PUNTOS) y esos los encontrás en cualquier lado.
Además viene con Wifi, que la Pi no. Y por supuesto, Ethernet, que la Pi modelo A no tiene.
Trae 4GB de almacenamiento interno, y 512MB de RAM. ¿La Pi? Nada y 256MB. ¿Se vé el patrón? Pagás el doble, pero viene con más del doble de cosas adentro.
Tiene dos conectores USB, (Pi Modelo A: 1, Pi Modelo B: 2).
Una cosa importante: viene en una caja. Una caja linda, robusta y chiquita, que se puede poner en un lugar para que no se te parta el circuito en 4.
La CPU es mucho más rápida, y varias generaciones más nueva. La GPU es un poco menos potente, pero no te vas a dar cuenta, porque lo único que la Pi, con esa CPU y esa memoria, puede hacer con la GPU es decodear video por hardware. Para cualquier otra cosa remotamente interesante, vas a tener un cuello de botella.
Las dos GPUs son propietaris, pero por lo menos hay alguien trabajando en drivers abiertos para la Mali400.
La Mele es útil de entrada. Le ponés Ubuntu, o usas Android. Viene rooteada, no hay que hacer nada raro, es imposible de brickear, instalar Ubuntu es poner una SD (igual que la Pi!) pero como la CPU es mucho mejor, las cosas funcionan mejor.
¿Ya mencioné que mientras la Pi tiene un slot SD para almacenamiento, la Mele tiene 4GB de flash y un slot SD, y un slot eSata? ¿Y que si le ponés un disco está buena como fileserver?
Y lo mejor: no es una caridad. Sale U$S70 porque le estás pagando a una empresa para que cree estas cosas, y no depende de conseguir chips con descuento de Broadcom, es tan sólo un buen producto a un buen precio. Y eso está buenísimo porque significa que en dos años vamos a tener una cosa como la a1000 al precio de la Pi, y algo mucho mejor al precio de la a1000, y que eso es sostenible, porque la gente que lo diseña, lo construye, lo envía y lo vende se está ganando el puchero con esto, y le patea el traste a la caridad. Y eso es bueno. Y si querés comprar 100000 de estas, las pedís y te las mandan, en vez de pedire pre-compras, y tenerte meses esperando, y producirlas en cuentagotas.
Así que, compren esto, o algo parecido.
Además: U$S92.60 with envío gratis
Hey man, i love this article : D
I am Tom Cubie and i sell Mele A1000/A2000 on aliexpress.
I am also a developer contributing code on A10 kernel and u-boot.
I also donate the devices to developers.
This is my shop.
http://www.aliexpress.com/f...
Too bad I didn't know about it before or I would have ordered mine from you!
Thanks for the cool toys :-)
It actually has 3 USB ports (2 on the back, 1 next to the SD card slot).
The only Pi advantage (beside the price) are the availability of GPIO connectors and some others interfaces, which is great for hardware hackers, but those are not in the majority of Pi users.
Missed that one, thanks!
To make things really cheap/small, you have to make compromises. I agree Mele seems to be better in speed/memory etc. The most important thing Raspberry Pi has is a very enthusiastic community.
Anything that enthusiastic community implements, I can have running on the mele in one day, except for RISC OS stuff.
Estoy de acuerdo con Egon por el transfondo de su comentario: RaspberryPi ha sabido publicitarse y al final va a acabar siendo un proyecto (a largo plazo y con soporte por parte de la comunidad) en lugar de un mero producto. Esto me recuerda a la Sharp Zaurus que tengo olvidada por algún cajón de casa...
Evidentemente esta es sólo una opinión, y el tiempo lo pondrá todo en su lugar.
PD. Me he pasado por la web del producto y además de estar prácticamente todo en chino (o parecido) ni siquiera te puedes bajar la datasheet.
Obviamente está en chino. Es un producto chino para el mercado chino. Cuando lo recibís, está en chino, también.
yeah sure, all the ARM HW community can potentially live in harmony unlike the old x86 eat or be eaten mindset if you want to grow and prosper rather than be chained to x86 SOC for another life time, but i still think the pi guys should have made their base hardware from a generic Cortex + NEON SIMD base rather than this old ARM11 as now the pi code will not be able to be NEON super charged :)
and so the pi students cant learn and actually contribute new NEON assembly and C code to the wider ARM cortex code bases that need and want this extra speed boost (like apple did/do to all their core PPC and now ARM code) for the real world markets place.
perhaps the pi guys should now make another play and produce a Cortex A9 dual/quad with internal 1gig Ethernet and SATA as their next progression hardware as the minimum if not a quad like i always thought they should/would for long term educational use for instance Freescale do the imx6 quad and do 15year support contracts so long term supply is not a problem but their GFX cores are 3rd party http://www.vivantecorp.com/... and not open source OC so that needs sorting or reverse engineering....
Sorta missing the point of the pi there. The pi is a small CHEAP computer designed to be used for teaching, it doesn't need a modern powerful processor or loads of RAM. It does need to be cheap enough to mean that if you blow it up, or attach it to your homemade drone, or whatever pet project you may and you break it it's only £25 to replace, not £60. I want one to attach to the back of my TV and run XBMC, I want another to be low powered file server in the other room. pi will do these perfectly, the Mele is more than I need.
There is also the advantage of support, the Mele may be a better device but it's less accessible and less well known, more people developing for pi means more interesting uses and more diverse software written for it.
A school can also buy twice as many pi's for the same money and teach twice as many pupils, the extra value of having more ram and connections on the Mele does not outweigh the doubling of the cost of the pi when the pi's main purpose is teaching, not having super-duper hardware.
No, the school can't buy twice as many pi's, unless they are buying them to teach hardware hacking to the students.
If the idea is to teach programming, then there are going to be extra expenses with the Pi:
1) Boxes. You can't have bare circuit boards on a school, they won't last.
2) Monitors. The Pi requires a more expensive monitor.
3) Wiring. The Pi has no Wifi, so you have to run wires to all stations, and buy switches.
4) Storage. The Pi comes with no storage, so add SD cards (cheap, yes)
5) USB Hub. With only 2 USB ports, they are going to be taken by keyboard+mouse. If you want to plug a webcam, a BT dongle, a mass storage device, or *anything*, you need a hub. The Mele has 3 ports.
How much closer to U$S 70 is the Pi with all that? I am guessing pretty close. And then, for the same cost, you have a much lamer device.
The issue with the Mele in an educational context is that lots of useful features are included in the machine. The benefit of the Pi lacking a lot of these features is that students can learn computer setup and configuration (in terms of both software and hardware), contributing to their general computer literacy.
I would love three examples.
Also it has power source. And after recent ICS upgrade, Mele A1000 is usable even with Android.
Yes, did the ICS upgrade last night, and it's *much* better.
¿Que pensas hacer con la raspberry?
Does the A1000 has 4 GB of internal storage or just 2 GB? (the specs in DealExtreme reads: "2GB Nand Flash" for the A1000, and "4GB Nand Flash" for the A2000).
Mine has 4GB I think. I would have to look.
We shouldn't forget one factor that is very important: Time.
I think that the time between the moment the final design of the Raspberry Pi was set in stone and sent to manufacturing and the moment Mele launched these Allwinner can be measured in months, and there lies the difference.
Any manufacturer of this kind of devices need to deliver and must draw a line and end the design phase, maybe they could wait a few more weeks to be able to use a more powerful CPU or GPU or with better connectivity and/or cheaper. but that's a never ending story.
Case in point: Allwinner and other chinese ARM licensees whose products power Android-compatible tablets/STB (Mele Ax000 are among them) that started to flood the market in Nov/Dec '11 have launched multi- (mostly dual-) core designs a couple of months ago and products based on them are already starting to appear.
Products that surely will make the Mele A1000 itself look underpowered in a way similar to your comparison with the Raspberry Pi.
All of this In a period of six months or less.
+1: very valid point. The Mele is very nice, but there's no need to slag off the Raspberry Pi. It echoes an earlier initiative by the BBC with the BBC Micro, which brought a lot of young people to programming. And not through schools, either; it made it affordable to get your own, easily programmable computer and play with it at home. So what if the Raspberry Pi people don't have their own fab and deep pockets? They still tried to make their corner of the world a little better, and there's no shame in that.
I am all for anyone starting a cool project for whatever reasons. I know I start a lot of things for no reason at all. OTOH, the final product should not be sheltered from criticism.
I know that the A1000 will look underpowered in 3 months. That will not make it any less powerful thhan it is. OTOH, will the RPi foundation have a much more powerful product in 3 months? Will they have made the Pi cheaper? My guess is, probably no.
If they can't, that has to be considered when evaluating the merits and results of the product and the project.
That's with 20:20 hindsight, is it? The RPi people went with what options they had at the time. There will still be people who can afford the lower-priced product, but not the higher-priced product. If a few of them opt for the RPi because of price, then there are no losers, are there?
Agreed 100%. This is June 2012, and we do have hindsight. I am not saying they did anything wrong, just that the result is not all that great.
It's not hindsight. Moore's law is very predictable. People have been saying for several months that by the time the Raspberry Pi Foundation can meet demand, there won't be any left because it will be obsolete on arrival. This is just a repeat of the OLPC which was another lame project where they promised vapor and took way too long to deliver. If we consider this type of deceptive advertising acceptable in this industry then I have a kickstarter project you may be interested in.
I'm going to make a Core i7 computer the size of a credit card with 8x USB 3.0 ports and sell it for a $1. I don't want to commit to a release date because I'm a con artist hiding behind a good cause, and am also full of shit. Since when did Tech Sector startup product releases turn in to options trading? I'll sell you 9 gazillion mega flops today, to be delivered somewhere between Q1 of 2013 and Q3 of 3011.
The joke when first announced was the name Raspberry Pi was a ballsy foreshadow to the fact they never intended to deliver on a reasonable schedule and would be blowing raspberrys at their suckers^H^H^H^H^H^H^H customers all the way to the bank.
Thanks for the pointer to this interesting device. Could anyone who has one tell me whether or not it is CE certified?
Sorry! There are no stickers or certification logos I recognize, and have thrown away the manuals.
The Rasberry Pi has been made in order to resemble the 1980ies Homecomputers, aka ZX 81, ZX Spectrum, VIC20, C64, etc. People on the university of Cambridge say: "We got a lot of talented IT students, it's just, they don't know anymore what a computer is", meaning, that today's way of development is highly abstracted, away from the machine. They designed the Rasberry Pi so, that you could boot directly into Python, afair, which would resemble the booting into BASIC on the old machines. The Pi NEVER was intended for the market, it has found now. Oh, and don't forget all the pins on the Rasberry Pi board. You don't find that on the mele.
Interesting article and I agree with you. But I'd like to raise two points:
Firstly, the RPi may not give as much "bang per buck" as others, but perhaps it offers enough "bang" for many people and their purposes?
Secondly, the RPi has many I/O pins for hardware experimentation. I don't know if these A10-based devices have any general-purpose I/O. If they do, it's nowhere near as easy to get to.
Ningún problema excepto pagar el impuesto :-)
not bothering to wade through all these comments to see if i'm repeating someone else's sentiment. so you thinking kicking charity in the 'ass' is a good thing? let's hope you never need charity.
Looks like we have a confusion here, or maybe I misspoke.They are a charity. They don't *do* any charity.
In other words, they are a non-profit that provides a mediocre product. I would rather products be provided by companies trying to make a profit, since that's much better for the economy in general.