the equipment is being deployed in remote locations. i didn't explain _how_ remote but often it requires cherry pickers or climbing up ladders to install the kit. it's already 16kg: a UPS would a) not be enough b) would be too heavy (regardless of capacity) c) would be additional cost.

also, the answer is in the question itself, already. any UPS is going to be what... £50 minimum? at the lower capacity end (16gbyte to 80gbyte) the cost differential between a power-loss-protected SSD and a consumer-grade unprotected one is around a £15 variation.

so you'd be making it dangerous for on-site engineers to install the kit (too heavy to lift one-handed whilst up a ladder) and actually *increasing* the cost of the equipment. not sensible! :)

i think it reasonable to assume that the write cache is DDR RAM. on the Innodisk 3MP SATA Slim the DDR RAM IC was clearly visible, as was Innodisk's CPU and the Toshiba NAND ICs.

probably by accident. it was hard enough to find out the details of drives as it was. imagine having to do several hours worth of google searching per drive, and having a list of suppliers whom you're asking "can you find out if this drive has power-loss protection" and they ask "what's that" - it gets really really boring after a while.

  the 250gb one you reference is for example far too much money: the requirements were, after all, a minimum of *16 gbytes*. not a minimum of 160 gbytes: not a *maximum* of 160 gbytes, a *minimum* of 16 gbytes. the 120gbyte Samsung 840 however is actually within budget. over 100gbytes of its capacity would be completely wasted for the application it would be considered for, but that's ok.

  so thank you: i'll put it on the list and we'll consider having a look at it if i am permitted to.

the number of drives that _weren't_ investigated because they were a) too expensive b) had no datasheets that could be found even with exhaustive testing c) were far too large in capacity was much *much* larger.

*sigh* yeah. the very early Vertex's were fine [i say fine: they were "no worse than any other consumer-grade SSD that had no power-loss protection"]. we surmised that it was only later that OCZ decided to ignore Sandforce's advice. sadly, *all* the OCZ's got tarred with the same brush. the investigation that found that the crap drives could all be firmware-downgraded came *after* they'd all been pulled at enormous expense.

I note that there are quite a few risks associated with CASH including:

* theft of physical wallets that are used to store CASH
* absence of realistic application of frameworks to tackle customer problems, disputes and charge backs with CASH.
* exposure to potential losses because of high volatility in value of the CASH currencies
* legal and financial risks assocated with trusting strangers with your CASH,
* and breach of anti-money laundering laws where people can carry CASH because of lack of complete information on tracking of CASH between anonymous random strangers.

so. wonderful. india's announcing something random about bitcoin. great! let's watch the value go up again just like it did in china as it comes to more peoples' attention. yay!

brian: define "value added". more to the point, whose "values" are you considering: yours? or the kinds of people for whom access to FSF-Endorseable hardware is absolutely critical? you cannot dictate what other people find find to be "value".

WHERE?? WHICH ONE, DAMNIT!!

... but with a modular tablet you'd be able to transfer - in seconds - the entire applications and data over to a replacement unscratched tablet chassis with a new battery which would cost you *less* money than the equivalent monolithic product.

you need to remember to view this from both sides. it's possible to replace *either* the CPU Card *or* the chassis, and in each case you have significant advantages and lower costs.

when did you ever buy a hermetically-sealed product that you could upgrade? the clue is in the word "hermetically-sealed".... :)

true. i have a design which uses PCCARD 3.3mm. if you have around $250,000 for the tooling costs for all the parts (assemblies, housings, sockets, casework) i can get it done... maybe in about 6 months time. or... we could use off-the-shelf parts and get immediately into production.

which would you prefer? perfect waffle-ware - more expensive due to the investment and NREs - or actual product that's reasonably-priced because there's no investor overheads?

simply not true.

you know what? whoever you are, foobar bazbot, i'm amazed and delighted to see that you clearly Get this concept. there are a couple of things that you left out:

1) from a CPU Card manufacturer's perspective, they love the fact that a short-lived SoC in a ready-to-go pre-packaged product can be sold in much bigger volume because it's shared - for the relatively short duration that the SoC has its day - across potentially dozens of mass-volume products.

2) from your perspective (1) translates into cost savings due to the CPU Card manufacturer being able to take advantage of stable huge volume pricing, as well as the Foundries, having larger orders, being able to dedicate and optimise a fab to get the yields up. both the volumes and the better yields automatically (one might hope!) translate into lower pricing

3) from a cost perspective, the fact that there is about an extra $6 on the BOM when compared to a monolithic product... this is *completely* dwarfed by the immense cost saving when you buy one or more EOMA68-compliant "chassis" and share a single CPU Card between them. laptop and tablet are the two obvious examples, with the clear additional benefit that applications and data transfer conveniently *between* the products.

4) from an environmental waste perspective, EOMA68 significantly reduces e-waste by making it possible to re-purpose older CPU Cards down a chain. today's latest-and-greatest laptop/tablet CPU Card becomes tomorrow's router/NAS/SoHo server CPU Card.

so there is an enormous amount going on here in what appears to be an otherwise unobtrusive "wtf??" moment. i haven't begun to describe the benefits to the linux kernel developers yet (but have posted a number of times on LKML explaining the N CPU Cards plus M products instead of N*M monolithic designs.)

there's nothing to stop anyone from creating OSHW EOMA68-compliant CPU Cards. a good starting point for anyone wishing to do so would be Dr Ajith Kumar's work on a GPL-compliant KiCAD board, or any one of the boards from TI or Freescale which have full schematics and even CAD/CAM PCB files - complete - available.

they are indeed. tracking down a cost-effective desirable SoC from - and this is also a really important bit - a fabless semiconductor company that respects the GPL - is very very hard. let's go through the list so far of CPU Cards that i've 30-98% made the PCB CAD/CAM drawings for (the A20 one is the only one that's reached 100% completion so far)

* AM3389 CPU Card. GPL-compliant: yes. cost-effective: most definitely not. desirable: well, it turned out that there was a proprietary blob for HDMI, and it was to be an FSF-Endorseable CPU Card, so no.

* iMX6 CPU Card. GPL-compliant: yes. cost-effective: at $35 for a quad-core SoC in 1k volumes when the competition is $USD 12: mmmm.... no. desirable: yes.

* Ingenic jz4760 CPU Card. GPL-compliant: yes. cost-effective: yes (around $7). desirable: as it's only a 1ghz single-core MIPS with no HDMI output... mmm... no not really.

* Rockchip RK3188 Quad-core CPU Card. GPL-compliant: no. only "leaked" source code is available. cost-effective: yes (around $12. for quad-core! amazing). desirable: yes (good features). but, the GPL-compliance nixes it. that and the huge NREs demanded by rockchip for their development board details.

the list keeps going on and on like this. much of these issues go away once we have some sales. so if you'd like to see this project succeed, help out by buying one of these engineering boards. in the future you'll be able to re-purpose the old CPU Card by getting an alternative chassis (just the chassis), or you'd be able to sell the old CPU Card on ebay.

windmills - chaaarge!

reactos was the real reason why i ported samba-tng to w32, using mingw32 to compile it up. worked absolutely great. unfortunately you cannot effectively run samba-tng/w32 under windows (without changing the port numbers) because the ports 137, 138, 139 and 445 as well as the critical NamedPipe services are already occupied... by microsoft's implementation of SMB as well as microsoft's implementation of the critical MSRPC logon services (LSASS, NETLOGON and so on) without which it would be flat-out impossible to even log in to the box in order to see if the services were running!

likewise unfortunately because wine has had to implement MSRPC (completely independently), although it would run successfully you likewise would have to change the MSRPC pipe service names as well as the TCP and UDP port numbers of the endpoint mapper (port 135) because wine has had to implement \PIPE\winreg, \PIPE\srvsvc and many others which are *also* implemented in samba-tng.

the amount of cross-over between samba, wine and reactos at the core fundamental networking level (much of NT's design was based around networking and RPC services, even when run as a stand-alone system), is just crazy. especially when you consider that it takes about 250,000 lines of hard-core intensive c code just to get even the _fundamentals_ of MSRPC correct. it's been over twelve years so i've had to stop letting people know about the duplication of effort and just let them get on with spending their time learning the hard way that they're working on exactly the same thing... without sharing any effort between them.

there's some absolute golden nuggets in amongst the wine/reactos code. periodically - every few years - i have a go at extracting the DCOM implementation from wine - to build a stand-alone GNU/Linux + w32 DCOM library. the last person who tried that called it "TangramCOM". he forgot to commit some critical bits to the repository (such as the IDL compiler). if anyone's ever worked with DCOM at a high level (using e.g. python) you'll know that it's just stunningly easy. DCOM was - still is - why microsoft has been so insanely successful after all this time. the equivalent in the MacOS world is ObjectiveC, which achieves similar results (without the networking) at the compiler-level which is pretty ambitious and nuts but highly effective all the same.

ahh, what can you do, eh?

i'm just going over the batch of OCZs that we had to pull from locations all over the world. the cost of the recall was far in excess of the cost of the drives. over 200 of them. if you have an OCZ Vertex drive with firmware revision 1.11, it *will* fail spectacularly. all you need to do is set up 64 sets of parallel writes, run them for 10 minutes, and you *will* get data corruption. you can do this in a shell script (i used python) by spawning "cp -aux" of a directory hierarchy with 1500 subdirectories and 3,000 small files. 64 parallel sets of copying (and then deleting) i.e. if you do around 1.5 million file-directory creates and deletes you are *guaranteed* to have data corruption.

the strange thing: the very first Vertex OCZs released were absolutely fine. what i learned just yesterday was that *even* with a drive that has been consistently failing, if you downgrade its firmware to revision 1.7 *it becomes absolutely fine*.

the problem that we have is that upgrading units in-the-field when the firmware upgrade system provided by OCZ is an ISOLINUX cd image with FreeDOS and a firmware-flash program is going to be rather tricky when none of the systems have a screen let alone a keyboard.

by contrast, we have somewhere around 500 Intel 320s installed world-wide. there have only ever been 3 failures.

for the selection of the new drive (Intel 320s are end-of-life) i'm endeavouring to replicate that test system which was reported on slashdot to have destroyed 12 different SSDs within under an hour per drive. i have managed to destroy one already: Crucial M4. it took 2500 power-cycle interruptions (the program's still in development) so the M4 failed in under 24 hours. so don't get that one. still on the list: Innodisk 3-MP Sata Slim, Toshiba's new SSD, and Intel's new S3500.

the toshiba i can already tell you, if you interrupt its power you will find that, on power-up, some of the outstanding write requests will *not* have been actioned. this is partly good news: it means that the drive is detecting that it doesn't have power, so doesn't risk corrupting the drive. i'm looking forward to properly testing the 3-MP because they're cheap, small, and the datasheet has, unlike any other manufacturer, a heck of a lot of details about how they actually do power-loss protection. most other manufacturers don't even bother to mention power-loss protection, that's if you can find a proper datasheet at all.

i did an analysis of the Quark X1000 based on the Galileo schematics, and the assessment isn't good:
http://lists.phcomp.co.uk/pipermail/arm-netbook/2013-October/008979.html

the key failure is that there's absolutely no I/O multiplexing. given that intel actually designed the PXA series of ARM processors before selling them to marvell you have to wonder what was going through the minds of the engineers behind the Quark X1000.

the main points of the above link which automatically and very unfortunately make the Quark X1000 a complete failure are:

1) there's no video outputs, and the only options are USB2 (DisplayLink with no 3D capabilities and too slow to do video), SPI (for character-based LCDs) or PCIe. to match a 0.4 watt processor with a 20 watt 3D PCIe Graphics card is completely insane. there are therefore no good options for video display of *any* kind.

2) there's no "industrial" or "embedded" style GPIO. no CAN bus, no PWM, no ADC, no DAC. there's also no audio. there's not even I2S and there's certainly no SPDIF. so to make up for that lack you'd have to add something like a Cortex M0, M3 or M4 embedded controller... and given that those usually come with built-in Power Management, NAND Flash and SDRAM, for the majority of purposes where you'd need to use an embedded controller with a Quark as a GPIO expander you'd be better off, cost-wise, with... just the embedded controller.

overall then there really aren't *any* markets that this chip could be useful for. if i'm wrong about that, and anyone can actually think of good uses for it, please do speak up.