Shrinkflation, SanDisk Style

https://ounapuu.ee/posts/2023/02/15/shrinkflation/

by hddherman • 3 years ago

343 227 3 years ago

ounapuu.ee

acadapter• 3 years ago

Looks like it's just a way to utilize chips with defects, instead of throwing them away.

Some of the missing gigabyte is probably used for a memory mapping algorithm.

Anyone who has dug deeper into this issue?

leoedin2• 3 years ago

The author looks at one of the drives, but they bought 2. What size was the other one?

It may just be that there's some distribution of "usable flash blocks" in every flash chip, and they happened to get one which fell at the bottom end of the acceptable range. You'd have to sample quite a few before you could be certain this isn't just an outlier.

fluoridation• 3 years ago

It's clearly stated. One is 16 billion bytes, the other is 15.4 billion bytes.

hddherman2• 3 years ago

I bought two of the same model from the same retailer and the other one reported the exact same capacity.

Femtiono• 3 years ago

Good Take.

I already though it would be really weird if the flash ship is really smaller.

wongarsu• 3 years ago

Which is great. And if they would have followed their own logic (1GB == 1 billion bytes) and labeled it as a 15GB or 15.5GB stick everything would be fine.

I don't think memory mapping a defective stick really need additional memory, regular memory sticks already have to internally remap flash memory because of how flash memory can only delete multi-MB blocks at once.

miga• 3 years ago

So they need to overprovision, and compensate for effects - that is fine.

However, only advertising "usable size" is truthful, everything else is misleading.

It is sad that the legal system tolerates such cheating because it makes it harder to find the actual properties of the product. Or even makes it impossible to actually know how the product will actually work.

Y_Y5• 3 years ago

"Up to 16GiB of storage!"

It's shockingly common nowadays to advertise things with an upper limit, rather than something like expected value. They're just saying, "we won't tell you what it is, but there's no way it's more than this". Bizarre that it seems to be effective.

throwaway_3850• 3 years ago

Bizzare that democracies can't/won't regulate this.

schroeding• 3 years ago

The local telco store had an advertisement for a certain mobile phone contract in the store window with "up to 50 MB/s or more!".

Sure sounds better than "some amount of MB/s"! :D

hackernewds• 3 years ago

Uber did this early on with "Earn up to $35/hour!" when they really meant "Earn no more than $35/hour!"

helmholtz2• 3 years ago

I saw a billboard for a lottery, it said "estimated lotto jackpot 55 million." See, I did not know that shit was estimated. That would suck if you won and they go, "oh, sorry, we were off by two zeros. We estimate that you are angry!"

LanceH• 3 years ago

Was just looking at this with internet service last night. Though I can understand where they're almost forced into that, since explaining that you can't get 1Gig speed from a random server that isn't serving that fast would be impossible to a lot of people.

But, the cable company doesn't disappoint -- since I can't find the price of an upgrade without calling, nor mention of upload speeds.

hulitu3• 3 years ago

Just like streaming companies. Selling 320x240 MPG1 scaled to 1920x1080

laomai• 3 years ago

Yet keeps legal eagles in business :-/

jopsen• 3 years ago

However, only advertising "usable size" is truthful, everything else is misleading.

Sadly, this is extremely common for smartphones. 64 GB iphone, who knows what usable storage it has?

It is sad that the legal system tolerates...

Just because it's not prosecuted doesn't mean it's not illegal.

crote• 3 years ago

That's different. An iPhone with 64GB of storage will have 64GB of storage - some of that is storing system files, but it's still storing something!

quitit• 3 years ago

I dare say that "actual storage less" (and other equivalent phrases) are now past the point of covering for the discrepancy.

5%+ of missing capacity is definitely large enough for the actual expected capacity to be used for marketing.

This is also something that should find its way into smartphones. Samsung's are notorious for using significant amounts of the advertised storage for bundled marketing apps that cannot be removed.

bee_rider• 3 years ago

I think they would say that technically the capacity is there, they are just using it.

It is annoying though. They should just sell it as a 15GB drive.

gruez2• 3 years ago

I dare say that "actual storage less" (and other equivalent phrases) are now past the point of covering for the discrepancy.

I think the original reason for that disclaimer was due to filesystem formatting, not for the GB -> GiB discrepancy or providing less usable bytes on the block device than advertised.

dehrmann• 3 years ago

If they had a run where half the blocks are bad, would they sell it as 16 GB, but with only 8 GB usable?

stavros• 3 years ago

What the hell? Why is this legal? I assume it's very illegal to say "12 eggs" and then the box only contains 9, how is this allowed?

NKosmatos• 3 years ago

The same way that unlimited data plans are capped at a few GBs/month, unlimited SMSes and minutes have limits as well, unlimited bank transaction have a limit of a few ATM transactions per month and so on…

Hamuko3• 3 years ago

My unlimited data plans have always been unlimited.

tenebrisalietum2• 3 years ago

it's not illegal to say "12 eggs, some eggs used for internal storage maintenance. User accessible eggs will be less."

Aachen6• 3 years ago

If that's on the box. If this can't reasonably be expected and is hidden in an FAQ somewhere, not obvious when purchasing, then a box that says 12 eggs should contain on average 12 eggs. Dunno to what extent variability is allowed (when not purchasing huge amounts); for discrete items like eggs probably not but for gram weights usually it's +/- some amount.

Edit: found a comment relevant to allowed variability https://news.ycombinator.com/item?id=34802080

Tepix4• 3 years ago

The way to stop this practice is to return them rightaway.

Unfortunately most consumers aren't smart enough to even notice it.

amelius2• 3 years ago

I sometimes wish a power of two didn't end up so close to a power of 1000.

Ekaros• 3 years ago

I think it was chosen as it was so close to power of 1000...

If you think 2^10 as unit is weird, 2^8 would make lot more sense and be quarter of 1000...

Y_Y5• 3 years ago

This is a good point and feels like three kind of thing Conway would have written about.

One good way to find rational approximants is to look for large values in continued fractions. Looking at log_2(10) we get the continued fraction representation

    [3;3,9,2,2,4,6,2,1,1,3,1,18...]

The 9 isn't particularly large, but if you truncate just before it you get

   log_2(10) ~= 10/3

and hence

   2^10 ~= 10^3.

If you keep going and cut just before the 18 you find that

    2^325147 ~= 10^97879

This is correct to six places, as you can easily verify

    $ python3 -c print(str(2**325147)[:8])
    10000003

No doubt manufacturers will use this to further swindle us as soon as we have storage that requires 40 KiB pointers.

suddenclarity• 3 years ago

Any other reasonable explanations to explain a measurement error? Partitioned differently? I might be a bit too critical but it seems farfetched for me that they've withheld 579 mb only for no one to notice before. I'm predicting this will become major news on tech blogs. People love "company X misleads customers". It just seems too good to be true.

nousermane• 3 years ago

Unlike RAM, where you always get exact power-of-2 number of bits per chip, modern flash storage normally ships with defects, plus error-correction codes to deal with those.

Number of defects vary. Chips coming from the same factory, even same batch, are likely to have different number of defects, and will be binned accordingly.

Devices with larger (but manageable) number of defects will simply have larger ECC region reserved - leaving smaller space to show to host computer. OP's new USB stick is like that, that's it, I reckon.

exmadscientist• 3 years ago

There's more going on here: the read and write times have also changed. This looks like a completely new hardware spin. And I'd bet it's got TLC flash now instead of MLC, or something of that sort.

I hate TLC. But at least it's not QLC.

fomine3• 3 years ago

General trick is to have NAND chips in GiB and sell product as GB. So difference 16GiB - 16GB = 1.1GiB can be used to management area / wear leveling.

Tepix4• 3 years ago

No, this really happens. Even if you repartition them they will not have the usable bytes they are advertised with.

satysin4• 3 years ago

The author mentions they don’t “trust” VenToy yet (they don’t state which kind of trust) my question is why?

I’ve used VenToy just fine for a while now. Am I missing something?

theblazehen• 3 years ago

In my experience, it doesn't handle "unusual" boot disks well. For example, images not using GRUB / ISOLINUX, or the usual modern windows bootloader. I've had issues with Haiku OS and Redox OS for some examples. I still keep a VenToy disk for most uses, and another one for when VenToy doesn't work

satysin4• 3 years ago

Yes I would sort of expect it to struggle/fail with more unique disc images. I only ever use it with official ISOs for this reason. If I don't see the OS listed on the compatibility list I tend to not waste my time seeing if it works or not.

dspillett3• 3 years ago

I read that as VenToy being relatively new to them, compared to years of using a stick per OS, and they'not found enough time to play to prove it would work in all circumstances they might need/want it to.

satysin4• 3 years ago

Fair point. I would highly recommend to the author they take an hour and try it out. The amount of time VenToy as saved me over the past few years has to be in the 100+ hour range by now. Fantastic little tool.

Retr0id• 3 years ago

I'm not OP, but I've had all sorts of bad experiences getting "non-official" bootable USB media to work properly, in various edge-cases and weird hardware.

And by "official", I mean an .iso or .img directly dd'ed onto a drive - this is the setup that I've found to be most reliable.

I haven't tried Ventoy, and I suspect things are slightly nicer these days now that EFI has matured, but I'm personally not prepared to spend any longer than necessary debugging why my bootable USB isn't working properly.

satysin4• 3 years ago

This makes sense, I guess when I read they didn't "trust" VenToy my brain went to "I worry it might modify my install ISO with malware" or similar rather than the reliability/compatibility aspect.

I've not used VenToy with anything but official ISOs so cannot comment to more unique uses but I've never had an issue with the official images of Fedora, Debian, Windows, Ubuntu, etc.

For me VenToy has been great at making life a little easier and saves me time. I have a 128GB USB drive with two dozen ISOs on it. Much easier to work with the one drive vs half a dozen I have to keep rewriting ISOs onto.

GTP2• 3 years ago

For what is worth, I use Ventoy and it works very well for me. Only problem is it fails to boot if you have secure boot enabled, but I didn't even try to look for a solution since I don't use it anyway. So maybe there's an easy way to have it signed so it works with secure boot too.

duffyjp3• 3 years ago

Depends on what he's using it for. Ventoy won't boot on lots of intel Macs and certain questionable EFI systems. If they're refurbishing machines and time is more important having separate USB sticks for each OS is worth not having to fiddle with it.

paateetee• 3 years ago

what is wrong with using ventoy?

ulnarkressty• 3 years ago

Apparently multiple antiviruses flag it [0]. A tool that can inject whatever during the OS installation process should be closely scrutinized. Some also have trust issues with software that comes from China.

https://axebase.net/blog/2021/11/12/ventoy-verdacht-auf-malw...

GTP2• 3 years ago

Well, antiviruses usually flag also pentesting tools like nmap (this is just an example, I didn't actually test it but I wouldn't be surprised if it is flagged) because those are "hacking tools" even if they don't do anything by themselves. So it would be more interesting to know why it gets flagged, just saying that it is flagged doesn't mean much as like in the previous example, they could decide to flag something just because it is unusual.

GuB-42• 3 years ago

That should be illegal, they should use the same rules as for food packaging.

For a 1kg bag of flour bought in the EU, that would be

- On average, you should get at least 1kg of flour

- No bag should be less than 985g

- Packaging does not count

For a USB stick, that would be

- On average, the actual capacity must be no less than what's advertised

- For a given product, the actual capacity must be no less than what's advertised, within a small margin

- Only usable capacity counts

josalhor• 3 years ago

Totally agree. Something similar should also apply to laptops and smartphones. If your phone comes up with 30GB of bloatware on a 64GB device you should definitely not be allowed to advertise 64GB of space.

Vogtinator• 3 years ago

Please no "on average" or margins. If it says 16GB, it has to be able to store at least 16GB.

phirus• 3 years ago

Yes, the margin for potatoes is there because potatoes are discrete. A single potatoe weighs something around 100-300 grams, so hitting exactly 1kg is not that trivial. But bytes are bytes. If it says 16GB, I want at least 16GB.

sowbug• 3 years ago

And the distribution should remain normal: https://houstonstatisticians.wordpress.com/2013/01/11/poinca...

hinata08• 3 years ago

I would expect the same for smartphones and computers.

it was really frustrating back in the times, when your brand new 8GB android smartphone only had 1 GB of free space out of the box (The 8GB edition of the Sony Xperia M4 Aqua only had 1.25GB of space for the user !)

Nowadays, devices have larger memories. But an android with mandatory apps and 'necessary data', or a Windows 11 can take around 20GB away. It can't be neglected on low end systems. There are still devices with 32, 64 or 80GB out there ! (your redmis or galaxy A XX)

Since most smartphone manufacturers won't let you use another OS or a custom ROM, all of that space should be reserved for them, and not advertised to the customer ! It's the same for computers. Most users buy the computer with the Windows, and will never replace it.

oarsinsync• 3 years ago

Using the rules of food packaging… 958 / 1000 * 16GB = 15.76GB.

JadeNB• 3 years ago

Using the rules of food packaging… 958 / 1000 * 16GB = 15.76GB.

The figure was 985g, not 958g, but that's what you seem to have used in your calculation anyway. (Amusingly, when I pulled out bc to test which way you computed it, I immediately made exactly the same typo.)

unpopular424• 3 years ago

Would that mean that the flour that sticks to the packaging walls does not count as unusable?

pbhjpbhj3• 3 years ago

I think it would be fair to make the regulation either way, so long as they put 1kg flour in the bag.

I'd imagine the trading standards would only check to about +/-0.5% (5g); whilst the flour on the bag's surface is probably around 1g for a kilo bag (1kg)?

jiri• 3 years ago

Situation is really messed up.

I understand that 16GiB (16*2e30 bytes) chip is present (delivered to user), but some of it is broken and not available (industry standard). Therefore capacity is less than 16GiB. Also I believe that there is a threshold, that manufacturer is using to throw away chip that are too broken to be sold. Ok so far.

So proper declared capacity should be threshold value (minimum) or threshold to maximum value declared range. I understand that this proper capacity is not very nice round number.

Problem has past: Around 2000, manufacturers started using GB (10e9) instead of GiB (2e30) that gives them some margin (7%) for chip defects. Later margin was not enough, so they started using more misleading specs like this. It seems that margin is at least 10% nowadays instead of <7% in the past.

Personally, I would be satisfied with stated margin size (like 7%, 10%, 12% etc) to maximum capacity, f.e. "USB flash drive 16GB max capacity, at least 90% usable."

pwg2• 3 years ago

Problem has past: Around 2000, manufacturers started using GB (10e9) instead of GiB (2e30) that gives them some margin (7%) for chip defects

It actually started somewhere around 1995 -- with mechanical disk makers all choosing to advertise GB drives using the power of ten instead of the power of two, while maintaining the same labeling on the outer box.

And the reason was not to account for "chip defects" -- it allowed them to ship slightly less capacity, but print on the box a number that appeared larger to those who did not read the 4 point type fine print on the back.

Aachen6• 3 years ago

(Note that 1e9 is 1 000 000 000 and 2e30 is 2 000 000 000 000 000 000 000 000 000 000, different from 2^30.)

enlyth• 3 years ago

The 32GB and 64GB SanDisk variants are actually cheaper on Amazon than the 16GB one. Yes, the advertising is misleading, but OP could've saved money and quadrupled his space if he wanted to.

water-your-self2• 3 years ago

Your comment is not relevant to this post.

hddherman2• 3 years ago

I did not buy mine from Amazon, I was ordering other things from a local IT retailer and had room in the budget for two of these USB sticks. The price made sense in this context for me and I had little use for that extra space.

mkj• 3 years ago

Are the amazon ones genuine?

iakov• 3 years ago

Ordering from Amazon is like gambling. Will you receive a genuine product as pictured or described, or a broken out-of-the box piece of crap? Only Jeff knows!

I'd rather buy from a local electronics retailer. Sure, might be a few bucks more expensive, but at least they have a brick-and-mortar location where I can come and complain if the products suck.

timbre12342• 3 years ago

Sadly OP is behind the times here. The storage industry has pretty much co-opted GB to mean 1000^3 bytes, which is why you see folks in the know refering to GiB for the power-of-two-numbers. This is super-frustrating, but it's been like that for decades (literally: this was finally "officially" resolved in the late 90's by the IEC).

It's frustrating that SanDisk used to give you extra bytes and they stopped -- and everyone including me HATES it when products get worse with no external indication that they changed.......but let's be honest: it's kind of SanDisk's primary MO to buy the cheapest NAND they can find and sell it on the consumer market.

Dylan168076• 3 years ago

Please look at the numbers in the article again.

The old drive had 16 billion usable bytes. The new drive has 15.4 billion.

Base 2 has nothing to do with this problem.

Y-bar• 3 years ago

You are entirely correct, and the person you are responding to is wrong. But the author does themselves no favour by including this incorrect tidbit in the article:

[...] Operating Systems define 1 GB as 1,073,741,824 BYTES.

Mac OS, iOS, Ubuntu, and Debian operating systems at the very least all use base 10 for representing disk and storage space.

timbre12342• 3 years ago

Yup, looks like I mis-skimmed the article. Thrown off by the (somewhat out of place, in retrospect) quote about it. My bad.

dspillett3• 3 years ago

The storage industry has pretty much co-opted GB to mean 1000^3 bytes

G has always been 1000^3, M 1000^2, and K 1000, in the storage and communication industries. OS designers, and programmers more generally, started using 1024 instead for convenience, but that came later. The storage industry is doing it right, using the correct meaning of SI units, and programmers co-opted GB to mean 1024^3, it isn't the other way around.

There were times when the storage industry and programmers worked together to really stuff things up and cause further confusion by mixing & matching: the 1.44MB of HD 3.5" floppy disks was actually 1474560 bytes so 1.44*1024*1000.

pbhjpbhj3• 3 years ago

You say M "has always been" 1000000 in your first para, then tell us in the second how it didn't used to be that way.

hyperman1• 3 years ago

I've tried to find relevant EU law about this. IANAL, of course.

According to 'misleading and comparative advertising' - https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CEL... - it seems this could fall under article 2b and 2c, defining misleading resp comparative advertising. The definition of misleading here is:

  [..] any advertising which in any way, including its presentation, deceives or is likely to deceive the persons to whom it is addressed or whom it reaches and which, by reason of its deceptive nature, is likely to affect their economic behaviour or which, for those reasons, injures or is likely to injure a competitor [..]

This is a directive, unfortunately, so the actual law resides with (and is different in) each member state. OTOH, this means inhabitants might complain to their regulator, and don't need to sue sandisk themselves.

I was hoping this one would help. It says which errors are allowed when measuring things. Unfortunately it only speaks of weights and (physical) volumes, not other measurements like number of bits. https://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CEL...

In general, I don't have much idea about how to navigate the EU consumer protection landscape. So while I do have a lot of protections, it is sometimes hard to know which ones. Does anyone know a good site or something?

Y_Y5• 3 years ago

https://commission.europa.eu/live-work-travel-eu/consumer-ri...

I think that it would be normal to go to your national regulator first if that's an option.

dark-star3• 3 years ago

There are tools out there that let you reconfigure the flash controller to have less overprovisioning. You're losing reliability this way of course, but if you want to get exactly 16'000'000'000 bytes out of the stick there are ways to accomplish this.

screamingninja• 3 years ago

Which one would you recommend?

dark-star3• 3 years ago

it really depends on the type of controller chip that is used in the USB stick. I have used MPALL in the past, for Phison chips.

Here is a list: http://thehowto.wikidot.com/flash-drive-recovery

Downloading the actual tools requires searching around russian and/or chinese forums though, so you should take appropriate care...

csdvrx• 3 years ago

Do you know tools to change the amount of SLC in a modern NVMe? A part of the MLC is often configured as SLC, and used for caching. I'd like to set up drives to be 100% SLC, even if it divides the usable storage by say 4 for QLC

dark-star3• 3 years ago

sorry, I haven't had the chance to play around a lot with NVMe disks yet. Let us know if you find out something!

thrdbndndn3• 3 years ago

I'm confused by the report from F3 probe even for the "good" one.

It says:

    Device geometry:
          *Usable* size: 14.91 GB (31266816 blocks)
         Announced size: 14.91 GB (31266816 blocks)
                 Module: 16.00 GB (2^34 Bytes)

So why it's only 14.91? I assume it uses GiB (1024) -- but then, why "Module" is 2^34 B? Shouldn't it also be 16000000000 B or 14.91 GB?

rhn_mk1• 3 years ago

You still need overprovisioning on NAND flash devices.

vitus2• 3 years ago

This is just GiB vs GB. The usable / announced size is more or less 16 GB (roughly 14.91 GiB), which you can also verify by multiplying the number of blocks by the block size (512B).

thrdbndndn3• 3 years ago

I get GiB vs GB for usable size.

I'm asking why Module actually has 16 GiB (as it says, 2^34).

jbverschoor2• 3 years ago

The new one also has lower performance

Aachen6• 3 years ago

Higher, I think? The nanoseconds for a write operation are 14 instead of 25. Or did I read it the wrong way around / is another metric worse?

jbverschoor2• 3 years ago

That's weird. You're right. Not sure how I mixed that up

AlexandrB• 3 years ago

To me this reeks of Western Digital style dishonesty[1][2]. I was wondering how long it would take for the new parent company to rub off on SanDisk and I guess I have my answer. My suggestion is to stay the hell away from anything Western Digital (and now SanDisk) if accurate specs are important to your use case.

[1] https://arstechnica.com/gadgets/2020/05/western-digital-gets...

[2] https://arstechnica.com/gadgets/2020/09/western-digital-is-t...

deadbunny• 3 years ago

I missed that WD had purchased SanDisk. Well that's SanDIsk off the list of suppliers...

coastermug• 3 years ago

What supplier would you recommend? WD and Sandisk have been my go-to for all things storage, but I prefer to pay a premium up to 20% for better quality.

goosedragons• 3 years ago

Hmm. I also have a 16GB SanDisk Ultra made in China like the old one. I don't remember exactly when I bought it but it was several years ago. Mine shows up with a usable size of 14.77 GB (30965760 blocks) from the same f3probe command the article runs. Testing my terrible no name 16GB stick shows 14.65GB usable.

blagie2• 3 years ago

Here's my problem: I'm glad to pay a premium for Sandisk if it guarantees reliable, honest, and accurate devices. I can buy generic flash drives for a fraction of the cost.

If Sandisk is starting to cheat, why would I pay the premium? And if I'm willing to pay a premium, whom should I buy from?

Sandisk, prior to WD, was expensive but honest. I was concerned this might happen, since WD has a bad track record here.

glintik2• 3 years ago

SanDisk isn’t premium at all. And known for various issues.

brookst2• 3 years ago

The used to be premium? Or at least I considered them to be. TIL. Who is a good premium brand today?

acdha2• 3 years ago

They used to be. Then Western Digital bought them, and this is how they advertised hard drives for years.

comboy• 3 years ago

Then which brand is? And I'm mostly looking for reliable rather than premium.

kvetching• 3 years ago

Same story with many brands. Rather than raise their prices, they cut quality. Do companies not understand why people were buying them to begin with? Bounty paper towels are the latest brand that I noticed. They always were superior, thick large sheets. Now they are on par with generics.

beezlebroxxxxxx• 3 years ago

At a certain point in the lifespan of the brand the reason the brand is well known becomes less the quality of the product than the recognition and "stickiness" of the brand itself.

A certain mindset and approach has taken over a lot of companies where they want to pull a fast one on their customers and improve margins at the same time. They rely on the brand name to keep customers coming back (because research has shown that people are less likely to try a new brand in a product space over their chosen one, "'ol reliable", even if they complain about its falling quality) while lowering the quality of that product until the brand identity is all that differentiates the product from a competitor. This makes their margins better, competing with competitors, at the expense of the quality of the product.

At least until customer perception sours. Then the brand engages in "artificial" rebranding by "returning their product to quality" and advertising it as such. A recent and famous example of this cycle was Domino's pizza, which was a stellar marketing campaign built on...their 'old' pizza sucking. Many hyper margin conscious companies rinse and repeat this cycle.

cruano• 3 years ago

Remember "brands" are not a thing, all companies are people and people care about themselves. If you join a company and cut the costs to "save money", you'll probably get a big bonus and then jump ship somewhere else so you don't suffer the consequences.

amelius2• 3 years ago

Kingston and Samsung make decent drives.

inasio• 3 years ago

Samsung used to have the best regarded SSDs, especially their Pro line, but that's not the case anymore [0]

[0] https://www.pugetsystems.com/blog/2023/02/02/update-on-samsu...

ciupicri2• 3 years ago

"SSD shadiness: Kingston and PNY caught bait-and-switching cheaper components after good reviews", https://www.extremetech.com/extreme/184253-ssd-shadiness-kin...

botverse• 3 years ago

It’s like the probably fake story of the MBA that removed an olive from the salads in the airline. They way they should finish by the story is with another MBA removing the rest of the olives, then another the space between seats, then another started to charge per bag…

They just can’t see it.

permo-w2• 3 years ago

one thing that isn’t clear from this article is where he bought it from and what it said in the listing. all he says is that he bought one of the same size and brand. if the listing says 15gb, then I don’t see a problem. it’s up to you to see that and perhaps choose a different option

zwirbl• 3 years ago

Look at the fotos, 16GB is even printed on the stick itself. Without the 'usable memory less' clause

permo-w2• 3 years ago

ah yeah missed that. comment retracted

rvba• 3 years ago

I wonder why there are so few usb flash drives that have a physical "read only" lock.

That was so useful to defend against viruses.

A cheap flash drive will cost say 15 dollars and one with a physical button/lock to make it read only will be 150?

The cost of materials should be maybe 5 more dollars...

On a side note, if you go into one of those "print shops" that can print things for you and use an USB stick, then how can you sanitize your stick from the multiple viruses / worms that you will get? Only thing that comes to my mind is to use a computer where the hard drives are physically disconnected and some live CD linux. But even then it is unclear if you can really format the usb flash memory to make it safe.

prmoustache2• 3 years ago

I also realized a few years ago that the physical switch on the sdcard is just a marker of intention. Hardware as well as software can just ignore it.

Aachen6• 3 years ago

I'm having trouble making my software ignore it. What command would write to the card when it has the write lock set?

tenebrisalietum2• 3 years ago

If you zero the USB drive out using `dd` or other equivalent tool that literally writes zeroes to each LBA, its safe. To avoid excessive wear and tear you can really just get away with zeroing the first couple MBs - without a valid partition table or recognized file system, no OS will do anything when it's plugged in other than ask you to format it.

yrro3• 3 years ago

See also the wipefs command.

codazoda• 3 years ago

This has been going on my entire career, starting in at least 1996 or so. Maybe SanDisk was an exception? People have been complaining about this for a very long time so I don’t think it’s quite right to call it Shrinkflation.

But, maybe I’ve been conditioned into complacency. I do agree, manufacturers should have to post usable space. I don’t see this changing though.

Edit: Removed note about 1,000,000,000 bytes. The article isn’t specifically about that, but I’m not convinced that’s ever been the only factor.

scajanus• 3 years ago

The article was not about that. The old memory stick was around 16,000,000,000 bytes; the new one 15,400,000,000.

So even with the 1,000,000,000 bytes to a gigabyte it doesn't round up to 16 GB.

gladiatr722• 3 years ago

Snap. Yeah, but it's a smoke and mirrors strategy that the storage industry has been using for decades.

We gave you what you paid for because we redefined what it is that you payed for.

gladiatr722• 3 years ago

Yes. The tech business keeps redefining technical terms for marketing purposes, and we end up with GIBibs.

Conjures images of those rusty cages they used to put people in to die..

drakenot• 3 years ago

This article was about SanDisk thumb drive having less than 1,000,000,000 usable bytes.

VectorLock• 3 years ago

What the article describes is not that. They're saying its an extra shrinkage ON TOP of the usual 1,000,000,000 bytes = 1Gb thing drive makers have already done for a long time.

Tepix4• 3 years ago

I have had the same happen to me with Toshiba MicroSD cards. I ended up returning them.

I also bought a Transcend 128GB USB stick, which i unfortunately did not return (yet). I have communicated with Transcend support and they said it was an issue with partitioning and filesystems - which is wrong, because the tools i use show the capacity at a lower level. They kept stubbornly repeating these statements.

I may end up returning it which shouldn't be a problem even after almost 2 years because the defect has been there since the beginning.

Here's what diskutil has to say about it:

    % diskutil info /dev/disk4
    Device Identifier:         disk4
    Device Node:               /dev/disk4
    Whole:                     Yes
    Part of Whole:             disk4
    Device / Media Name:       Transcend 128GB

    Disk Size:                 123.7 GB (123718336512 Bytes) (exactly 241637376 512-Byte-Units)
    Device Block Size:         512 Bytes

For comparison, i get this for a Lexar 32GB USB stick:

    Disk Size:                 32.0 GB (32008830976 Bytes) (exactly 62517248 512-Byte-Units)

You can clearly see that more than 4GB are missing on the Transcend 128GB stick. It should be around 119.2GiB (128GB) but it is 115.2GiB (123.7GB)

duffyjp3• 3 years ago

I'd be curious what the output is of the f3probe tool used in the article. I assume you have a 128GB module with the missing bits used as reserve for when blocks fail.

I'm fine with that if it's sold as a 124GB drive, not 128GB. I like to buy Intel drives because they'll sell you a 480GB drive and under the hood it's a 512GB module with tons of reserve.

Tepix4• 3 years ago

Sure, if they had sold it as a 123GB drive i'd be perfectly OK with it.

On macOS, f3probe isn't available, i'll have to hook it up to a Linux PC.

yrro3• 3 years ago

This is simple to resolve. If I buy a 16 GB flash drive with fewer than 16 billion bytes of usable capacity then I will demand a refund and return the device at the retailer's expense. I will also complain to trading standards.

rfoo• 3 years ago

I don't know if it is possible where you live, but can you demand a refund without returning the device?

That's how people in China countered fake 2TB thumb drives on taobao. They knowingly buy the device and ask for refund (because it's fake) without returning. They have an online forum full of "which merchant is selling fake drives today".

The trick is there are "Mass Production Tools" (I think it actually means firmware configurator) for the flash controller used in these scam drives, which can turn them back to their original capacity (usually 4GB or less). So by doing so they effectively win some free tiny-ish thumb drives.

yrro3• 3 years ago

In England & Wales, goods must correspond to their description [Sales of Goods Act]. If not then the seller must offer a full refund including the cost of returning the mis-sold item [Consumer Rights Act and other acts relating to the implementation of the EU Distance Selling directives].

If all or part of the payment for an item is made with a credit card then the credit card issuer is jointly and severally responsible ["Section 75"]. So if a retailer doesn't play along will find themselves answerable to their payment processor; the customer doesn't have to get involved.

It's not unheard of for sellers to issue a refund and request the customer dispose of a faulty or mis-sold item instead of bothering to return it. But that's only if the customer is OK with it; if I wanted to return, say, a washing machine and the seller refused to co-operate then I'd be in my rights to dispose of it at my expense and refund the balance minus my reasonable costs to the seller.

lisplist• 3 years ago

Tangent, but I think the practice of shrinkflation should be made illegal altogether (although not sure how you would enforce it). If you need to increase prices, fine, but it's a complete waste of packaging materials to just give less of something for the same price in the same package, not to mention the environmental harm.

account422• 3 years ago

Reducing package size can be in the interest of the user sometimes, e.g. if most people need less or won't use the whole package in time for perishables.

However it should be required to advertise the size reduction on the new packaging, including before and after price per amount.

kevincox4• 3 years ago

I find this is what supermarkets do. They sell you basil and other fresh herbs in MASSIVE packs that you will likely only use 1/4 of for any meal. The rest goes bad and gets tossed out.

It's basically a scam to make you feel like you are getting an ok deal because you are getting so much. But really it is just a way to charge you more without seeming like a price raise.

mxfh• 3 years ago

Burying some way more interesting unnoticed metric in the logs.

In practice, I would trade a ~75% faster drive for <10% of storage in a thumb drive use case with zero hesitation:

Old: Probe time: 1'49" Read: 463.8ms / 4814 = 96us Write: 1'48" / 4192321 = 25us

New: Probe time: 1'02" Read: 338.2ms / 4814 = 70us Write: 1'02" / 4192321 = 14us

ryukafalz• 3 years ago

I might too, but only if I knew that's what I was doing.

Aachen6• 3 years ago

Was thinking the same, but with it not being called out in the article at all, I was wondering whether to trust these numbers. The tool says "probe", not "benchmark". Not sure what to make of it

marcosdumay• 3 years ago

So, guess what, advertising that it's a bit smaller and much faster is probably the best way to get people like you to buy.

daneel_w2• 3 years ago

I stopped buying WD many years ago because of their dishonest marketing tactics and, frankly, shit products.

roeles• 3 years ago

Any recommendation for reliable disks, for NAS? Or external USB?

WD used to be the brand for these around 2010, according to my uni friends.

daneel_w2• 3 years ago

I've used only Toshiba drives the past 10 or so years. Not a single issue so far, so I'll happily swear by that brand. Currently I use a stack of their L200 2TB 2.5" drives because of how little power they use (<2 watts) and because of not having intermediary flash storage like e.g. Seagate's 2.5" drives do - because I'm convinced that the flash portion is just a liability which will fail sooner than the actual mechanical drive. The L200s are SMR drives so low write speeds (about 25 MB/sec seq. at worst) will occur intermittently if that is a concern.

fwlr• 3 years ago

It would be nice to to have something like caniuse.com or dpi.lv, but for “usable storage space on device”. I doubt it would shame most manufacturers into actually providing what they advertise, but it might just create enough visibility for one or two manufacturers to commit the extra resources to delivering on the promise.

Aeolun• 3 years ago

How can SanDisk redefine a SI unit to suit them? Don’t we have GiB for that?

kevincox4• 3 years ago

In theory yes. In practice GiB is rarely used and GB is widely used for 10^9 and 2^30.

It's a shame, but due to the rare ruse I doubt it would prevent much confusion in practice.

Yannik_Sc• 3 years ago

I'd bet it's because your new one uses an MBR partition as your tools show you clearly, that indeed both sticks are 16 GB sticks

m4632• 3 years ago

one thing I've learned - IODD is cool

Shrinkflation, SanDisk Style

Related Stories