Less charitable or More cynical? How is Amazon supposed to track a 3rd party pulling their SDK and then reverse-engineering their own service side to work with the SDK? Assuming we're all okay with that premise to begin with, all sorts of other questions start popping up.
Do these 3rd parties get veto power over a feature they can't support?
Can they delay a launch if they need more time to make their reverse-engineered effort compatible again?
It seems a hard to defend position that this is at all Amazon's problem. The OP even links to the blog post announcing this change months ago. If users pay you for your service to remain S3-compatible that seems like its on you to make sure you live up to that promise, not Amazon.
Clicking through to the actual git issues, it definitely seems like the maintainers of Iceberg have the right mental model here too. This is their problem to fix. After re-reading this post this mostly feels like a click-baity way to advertise OpenDAL, which the author appears to be heavily involved in.
Requiring a header "just because you sniffed it to usually be there" is not Amazon being cynical, it's creatively-developing-overly-strict-checks. And it happens on the side of the S3-compatible service.
If your service no longer works with the AWS SDK because you crash at `headers["content-md5"]` just because "it seemed a good way to make things more correct" - it is on you to fix it, IMO.
Why does Minio mandate the presence of Content-MD5? Is it in the docs somewhere for the S3 "protocol"? No, it's not. It's someone wanting to "be extra correct with validating user input" and thus creating a subtle extra restriction on the interface they do not control.
I think you misread my response. I think assuming Amazon did this to hurt “s3 compatible” services is cynical. Amazon implemented a feature, well within their rights. Writing a blog post saying they “broke backwards compatibility” is cynical and disingenuous. Amazon never committed to supporting any random use of their SDK.
Hard agree. If AWS were offering “S3 compatibility certification” or similar I could see framing this as an AWS/S3 problem. This seems like the definition of “S3 compatible” changed, and now everyone claiming it needs to catch up again.
Traditionally, in poker variants with wild cards that enable 5 of a kind hands, it does IIRC beat a royal flush. )"Royal straight" isn't a thing; AKQJT straights are sometimes called "broadway", but they're never distinguished as a separate hand type. Whereas royal flushes do get distinguished from straight flushes, but not for any good reason.)
But all of this is moot because TFA doesn't define "suits" for the "cards" anyway. And of course the relative probabilities do change when you only have 5 ranks. (And we're also effectively "drawing" without replacement; there are an effectively unlimited number of each rank available.)
If an Uber driver caused you to miss a flight by driving around a parking lot in circles at a speed you can't exit the vehicle, you don't think it would be a reasonable request for the customer to ask Uber to make it right?
Fair enough, there is a difference. But now we are not looking at a missed flight so much as attempted kidnapping or imprisonment or some other much more serious crime. Which is interesting to think about with the Waymo example, but hard to take seriously in the context of the video since the rider declines to do what the customer service rep asks them to do (at least appears to for the sake of producing additional outrage for their video)
When you call from a cell phone, public safety answering points get your gps data from the carrier. Landline calls have their associated address data sent. This is called e911 and has been a thing for over 20 years.
The person answering the call will ask for a specific address, but does not need it to send help towards you.
Eventually, a 911 PSAP should (but does not always) eventually get decent geographical coordinates -- that's part of what the E911 system is intended to provide.
When it does work (and it usually does work): It's not always instant, it isn't always accurate (the first hit may just be the coordinates of a cell tower -- good luck!) and it doesn't always work inside of buildings. And for some callers some of the time, things like VOIP won't deliver the correct location to a PSAP because things are broken or databases are simply wrong.
And in the best case: It can only reveal the calling party's location, which is not necessarily the location where people actually need help. And it only necessarily reveals that calling party's location to the 911 operator.
It's still generally the job of a human dispatcher to relay that information to the boots-on-the-ground who will actually show up and help the people who need that help.
It's nice to think of it as some tightly-integrated system where somehow the information is, say, relayed automatically from the caller's phone, through CAD, and all the way to the dashboard satnav of an ambulance so they can just hop in and go.
But what usually happens (in my experience hanging around in 911 PSAPs) is that the location is relayed to first responders by human voice over radio.
And addresses are easy to relay by voice.
> The person answering the call will ask for a specific address, but does not need it to send help towards you.
It's important to provide an address because addresses are useful to the 911 operator. E911 is awesome, but it is not an all-knowing, all-seeing system that is somehow born from perfection.
Assuming you mean "undefined" as "unprovable", this still seems like an extreme stance to me. I also think, even if we had this the next goal-post would be "provenance of the video".
As a cost-to-entry for a casual player, this seems like a pretty hard sell. I'd rather just find something else to play personally, than deal with a webcam and some invasive recording situation to play.
for pvp competitive games where cheating is possible, you and your computer must be more public while the game is active. how else could cheating be observed?
many people already run a distinct gaming pc, either for kernel anti-cheat or other reasons.
i would prefer not to be surveilled on my iphone.
i would prefer to be surveilled while playing fortnite or tf2. along with all the other players.
Actually, I'm asking Google to serve me a file via a GET request. The response they send back includes the content along with an ad. Google would prefer I watch the ad but since the bits are on my computer and Google has no say in how I operate my computer, which bits I read, or how I allocate my time their wishes don't really matter after they've sent me the bits I care about.
Google could just say "no" and not send me any content at all. I'm not "taking" anything though.
If this is the analogy, why doesn't Google try something like protecting from sticks in the first place? For example, they could lock all of YouTube behind a login & paywall.
It seems like YouTube is attempting to FORCE people to adhere to its desired business model. If ad-supported video playback is not economically viable, then maybe it's time for a new model to be born? There is no obligation for me to pick up the stick and hit myself with it because Google says I should. I downloaded a bunch of bits that Google sent to my computer, and then I read the bits I cared about. That's how every webpage has worked in my experience.
They offer you a direct payment option which would quite well for not seeing ads while still supporting the people who make the videos which are so important to you.
And yet it's optional and they also keep sending me the same videos for free along with ads.
I could opt to close my eyes and ears for every ad they send instead and it would have the same effect at the cost of my personal time. Ad blockers are just time shifting Tivos in disguise. Google has no say in how I spend my time interacting with the bits they already sent to my computer.
Advertisers know you don’t always watch, but they price in the odds that the ad will catch your eye and pay.
Ad blockers drive those odds to zero, so the advertisers won’t pay.
Personally, I like paying people who make things I enjoy, but if you feel you’re entitled to their work without paying for it, just be honest about it and be you.
A similar realization had a major impact on how Blizzard approached end-game raiding in WoW. In the first 2 expansions, only the most dedicated players were seeing the final boss and culmination of the storylines. It took 40 players working together to get there, and those 40 players had to execute complex fights in order to reach the end. Though a very rewarding experience for those who could do it, Blizzard did the math and realized they were excluding the VAST majority of their playerbase from the coolest content. They were spending tons of money creating this content and no one was experiencing it!
In the 3rd expansion and ever since then, end game content has been tuned to be a lot more forgiving and to require less people. There are still complex and rewarding fights, but ramping up the difficulty is "opt-in" generally speaking.
This is likely one of the reasons the game remained so successful for so long. Prior to this mentality shift, it was very common for end game MMO content to only be seen by a small minority of the players.
A similar thing happened with Star Wars Galaxies. Part of the original pitch was that like <1% of players would have access to the Force. Only from there with super hard training could they be Jedi. This matches the lore of the universe, Jedi (especially in the OT era) are the most rare creatures.
This was accurate, and would make seeing a Jedi in the game world the coolest thing imaginable, but it was a disaster for the finances of running the game.
Obviously, EVERYONE wants to be a Jedi. So they re-tooled it so that anyone could. Eventually the whole game was replaced with another MMO that was set in the ancient ages when they were common.
Do these 3rd parties get veto power over a feature they can't support?
Can they delay a launch if they need more time to make their reverse-engineered effort compatible again?
It seems a hard to defend position that this is at all Amazon's problem. The OP even links to the blog post announcing this change months ago. If users pay you for your service to remain S3-compatible that seems like its on you to make sure you live up to that promise, not Amazon.
Clicking through to the actual git issues, it definitely seems like the maintainers of Iceberg have the right mental model here too. This is their problem to fix. After re-reading this post this mostly feels like a click-baity way to advertise OpenDAL, which the author appears to be heavily involved in.
reply