CVS and E*Trade both refused to accept my fairly standard [email protected] address during initial registration, but had no issue changing to that address once the account was created. It would be nice if their internal teams communicated a bit better.
The full email address syntax described in the RFC cannot be precisely matched with a mere regular expression due to the support for nested comments. The need to track arbitrarily deep nesting state makes it a non-regular language.
If you remove the comments first the remainder can be parsed with a very complex regex, but it will be about a kilobyte long.
So you’re not remapping the source ports to be unique? There’s no mechanism to avoid collisions when multiple clients use the same source port? Full Cone NAT implies that you have to remember the mapping (potentially indefinitely—if you ever reassign a given external IP:port combination to a different internal IP or port after it’s been used you’re not implementing Full Cone NAT), but not that the internal and external ports need to be identical. It would generally only be used when you have a large enough pool of external IP addresses available to assign a unique external IP:port for every internal IP:port. Which usually implies a unique external IP for each internal IP, as you can’t restrict the number of unique ports used by each client. This is why most routers only implement Symmetric NAT.
(If you do have sufficient external IPs the Linux kernel can do Full Cone NAT by translating only the IP addresses and not the ports, via SNAT/DNAT prefix mapping. The part it lacks, for very practical reasons, is support for attempting to create permanent unique mappings from a larger number of unconstrained internal IP:port combinations to a smaller number of external ones.)
The most valuable thing is an experienced team who thoroughly understand both the specifications and the implementation as well as the reasoning behind both. Written specifications are great as onboarding and reference material but there will always be gaps between the specifications and the code. (“The map is not the territory.”) Even with solid specifications you can’t just turn over maintenance of a codebase to a new team and expect them to immediately be productive with it.
Who is enforcing this and how?
Liability would be decided by the courts or another form of binding arbitration. Obviously. Harming someone through action or negligence is a tort, and torts are addressed by the judicial branch. Both sides would present their arguments, including any scientific evidence in their favor—the FDA or similar organizations could weigh in here as expert witnesses, if they have something to offer—and the court will decide whether the vendor acted reasonably or has liability toward the defendant.
If you knowingly sell me a car with an engine about to fail, you are in no way accountable.
If you knew that the engine was about to fail and didn’t disclose that fact, or specifically indicate that the vehicle was being sold “as-is” with no guarantees, then you certainly should be accountable for that. Your contract with the buyer was based on the premise that they were getting a vehicle in a certain condition. An unknown fault would be one thing, but if you knew about the issue and the buyer did not then there was no “meeting of the minds”, which means that the contract is void and you are a thief for taking their payment under false pretenses.
Anyway, you continue to miss the point. I’m not saying that everyone should become an expert in every domain. I’m saying that people should be able to choose their own experts (reputation sources) rather than have one particular organization like the FDA (instance/community moderators) pre-filtering the options for everyone. I wasn’t even the one who brought up the FDA—this thread was originally about online content moderation. If you insist on continuing the thread please try to limit yourself to relevant points.
No, I am not okay with bans like that. You should be able to knowingly buy products with mercury in them. Obviously if someone is selling products containing mercury and not disclosing that fact, passing them off as safe to handle, that would be a problem and they would be liable for any harm that resulted from that. But it doesn’t justify a preemptive ban.
You misunderstood. It’s not a middle ground between “can regulate” and “cannot regulate”. That would indeed be idiotic. It’s a middle ground between “must judge everything for yourself” and “someone else determines what you have access to”. Someone else does the evaluation and tells you whether they think it’s worthwhile, but you choose whose recommendations to listen to (or ignore, if you please).
What if they never even hear the FDA recommendation?
Then the FDA isn’t doing a very good job, are they? Ensuring that people hear their recommendations (and trust them) would be among their core goals.
The rare fringe cases where someone is affected indirectly without personally having choosen to purchase the product can be dealt with through the courts. There is no need for preemptive bans.
To put it another way: do you think we should have the FDA? Or do you think everybody should have to test everything they eat and put on their skin?
There is a middle ground. The FDA shouldn’t have the power to ban a product from the market. They should be able to publish their recommendations, however, and people who trust them can choose to follow those recommendations. Others should be free to publish their own recommendations, and some people will choose to follow those instead.
Applied to online content: Rather than having no filter at all, or relying on a controversial, centralized content policy, users would subscribe to “reputation servers” which would score content based on where it comes from. Anyone could participate in moderation and their moderation actions (positive or negative) would be shared publicly; servers would weight each action according to their own policies to determine an overall score to present to their followers. Users could choose a third-party reputation server to suit their own preferences or run their own, either from scratch or blending recommendations from one or more other servers.
In what sense do you think this isn’t following the email standard? The plus sign is a valid character in the local part, and the standard doesn’t say how it should be interpreted (it could be a significant part of the name; it’s not proper to strip it out) or preclude multiple addresses from delivering to the same mailbox.
Unfortunately the feature is too well-known, and the mapping from the tagged address to the plain address is too transparent. Spammers will just remove the label. You need either a custom domain so you can use a different separator (‘+’ is the default but you can generally choose something else for your own server) or a way to generate random, opaque temporary addresses.
If you want to talk about non-compliant address handing, aside from not accepting valid addresses, the one that always bothers me is sites that capitalize or lowercase the local part of the address. Domain names are not case-sensitive, but the local part is. Changing the case could result in non-delivery or delivery to the wrong mailbox. Most servers are case-insensitive but senders shouldn’t assume that is always true.