Here is a use case scenario to illustrate better what I mean:
I have used the name JohnSmith on SAFETube for over a year and uploaded lots of videos with that name and with the ID XYZ57432. Then I want to change my name to John and keep my ID since I want to keep using my same SAFETube account. Then the network will refuse my request for changing my name to John if there already exists another user with the name John and with an ID that starts with XYZ.
That sucks. Then itâs just like the name squatting problem again. And even if itâs unlikely to happen, for the users that it inevitably will happen to it will be really bad.
Exactly. You canât ALWAYS get what you want, but you can always make a slight compromise on your user name, in the very rare case that there is a collision. If you canât get exactly what suits your fancy, you might decide that thereâs not enough reason to change, or . . .
I donât see why not really. We just felt a limit is good to prevent abuse. Three possibly is a better maximum number though.
Its not really name squatting, if you try to be IBM with a close ID you cannot. You can be IBM though, but not pretend you are the original one or get an ID close to the original. If you consider the 10 digits (80 bits) there is another 512 - 80 bits so chances of this collision are very tiny. If you are trying to phish the chance is very high and stopped, but an actual collision is very unlikely.
Or, remove the requirement that the IDs have to look different. Use identicons instead for that.
Here is an identicon for: XYZ4132563
And here is the identicon for: XYZ4123563
I agree with your analysis @Anders, but not:
Then itâs just like the name squatting problem again.
Squatting would block only a small part of the ID space per account, so to cover a large part (necessary to create a market for squatted names) would require very large numbers of accounts to be created.
Someone could do some maths in this, but my feeling is that at $.50 per account, and two or three names per account name squatting just wonât be viable.
It does mean that common names are likely to be blocked by legitimate users though, and over time less common ones too. So that problem remains, but thatâs not half so bad IMO as people making a business model out if it, which only makes the problem worse.
Seems like a good compromise, but maybe we can improve it⌠Ideas? 
Ok, thatâs a good way of protecting brand names. People will recognize the real IBMâs unique ID.
I think that even this is far less than we might think.
I guess it depends a lot on what the network algorithm considers close, but even if itâs, say, he first three characters of a 10 digit alphanumeric string, the the colision chances will be really remote.
It will never be that you canât get the name you want. It MIGHT be that you have to change your account identifier to get it, but thatâs a call you can evaluate.
So, yes, the squatting model as itâs been done in the past is out for sure. But I have little doubt that those creative squatter types will find a game to play if there is one. 
Yes and the chance of a collisions will be along the lines of 2 ^ 80 (2^20 is the # of grains of sand in the planet, so collisions should be tiny). This will be for actual names, so there can be 2^80 - 1 Johns on the network. If itâs too small we can later increase the ID component, but I think it may be a far away problem. For sure itâs easily upgradable though, which is cool.
But arenât just the three first characters in the ID checked? Thatâs only 2^24. Even less if only A-Z and 0-9 are used. That would be around 5 bits per character times three = 2^15.
If for example the real IBM has the ID HAL7463933 then all combinations starting with HAL will be rejected by the network (for the name IBM).
Seems like I havenât fully grasped the ID system yet.
Yes you are correct itâs the first three (say) bytes (8 bits each)so 2^24. I will give links to the detailed design docs, itâs my anathema providing info people can understand and not too much. We can give the whole design in detail, but then folk donât want that much detail. Maybe best to say in overview collisions are tiny Even 2^15 is still tiny (hex char set so larger than this, I can get the exact figures but lets keep that to the detailed design parts which will get published (they are already).
If we say there are about the same amount of grains of sand on the planet as there can be unique IDâs then for most this will be enough. I Think you are right if we go into detail it will be around 2^18-20 depending on char et for collision avoidance and 2^80(ish) for full idâs.
This is precisely how to make unique items in project star Craft (the exchange project);
! Thanks !
From here I know all sorts of apps are securely enabled.
Looking forward to the upcoming Sprint
Can there not just be a general token name called anon that anyone can speak out of at any time? And when speaking from that, if MaidSAFE or ProjectSAFE got breached by some government or a DAO based on it, there would be no paper trail linking a unique anonymous name at the time to a unique identity name?
Its a variant of the two accounts notion I know. But anonymity has to be anonymity even under the very worst conditions. Sorry if Iâve missed something completely obvious here. If there is no absolute anonymity the gun pointed at the corporatocracyâs head isnât loaded and they will continue to escalate their dirty hand bs, it has to be made impractical and it will be an arms race to an extent, but the there is no choice.
Nope. And no need.
Anyone can create an account, or more than one. All anons are not equal, and most would insist on being their very own anonymous selves.
If one person uses âanonâ it wonât block the next person using anon? It shouldnât. There should be a cut out for a generic tag even if people also want exclusive anonymity tags or pseudo anonymity as well. Perception is an issue with uptake and even if people canât verbalize why or think it through completely I think any kind of exclusive anonymity will be intuitively perceived as lesser or somehow not fully genuine. Seems like an uphill battle to convince people that there would be no audit-able paper trail. Seems like the kind of stuff that could end up on a court transcript with arguments before some court that it had been proven the unique anonymous identifier was linked to a certain person. Its going to be taken as pseudo anonymity and that has its place but absolutely not good enough if thatâs all that will be on offer.
So despite saying that what Iâve suggested isnât needed and people donât prefer it, are we also saying that it canât be done under ProjectSafe or is prohibited or prohibited even under a fork? I believe it was Nick Lambert (not sure I donât want to use the @) in a response to another meta threat that mentioned a group is already working to produce a framework for completely anonymous forums. I guess I am suggesting a hybrid, but also questioning if pseudo anonymity will be passed off as something more.
Going back to your initial point in the OP, I donât see a balance I see a value hierarchy with the anonymity being more valuable than the verified identity even if it means bots are a problem and banking wonât work. If it would require two account types to get both thatâs a small price. But what again is the point of ProjectSAFE if it canât provide true anonymity? I sense there are some who just want a reboot of the present system based on sunk cost semi anonymous pseudonyms. Thatâs fine but it shouldnât get in the way of the potential of ProjectSAFE or even be allowed any kind of compromise. SAFE seems to be the opposite of compromising with commerce, its not about selling out.
Every ID can have safecoin sent/received (so you can even remove the last owner of a safecoin if you want to replace with a throw away ID, if you want).
Now that. Is true anonymity. Safecoin is thee best crypto when it comes to a medium of exchange.
I agree completely, except to say that âSafecoin is thy best crypto . . .â