Trying to figure something out regarding "deletion" of data

Development
1

I’ve registered this forum account, got “trusted” enough to actually run the SAFE Browser in its alpha stage, surfed to a couple of safe:// sites and created my own safe:// test site. So far, it’s working as I expected, with the exception of the infuriating idea of not remembering the “secret” and “password” fields, as well as a very “unpolished” experience at this stage, including an extremely messy series of websites lacking downloads even for pages that appear to be download pages, and documentation split up into a million locations.

One question whose (chat room) reply really bothered me was this:

“So how do I delete my test “public id” as they call them (I call them “domains”)?”
“You cannot delete anything on the SAFE network.”
“…”

From that point, a long back-and-forth followed without the SAFE expert answering the basic question, asked over and over:

“Why can’t you delete YOUR OWN content, that YOU OWN?”

Note: I’m NOT talking about deleting random content. Only about the stuff that the SAFE network has established is YOUR content, because you have the “secret key”, so to speak.

Why shuffle around dead, inaccessible data perpetually, which even the owner doesn’t wanna keep? WHY? It doesn’t make sense. I can’t believe this inane design decision is actually part of SAFE, because the rest seems to be so sensible.

Google pissed me off beyond words when they refused to let me delete an AdWords campaign in 2004 or something. Not sure if they ever enabled you to do that, but their attitude was like: “We know better than you. You are to SAVE this and be HAPPY that we STORE it for you. You are NOT deleting ANYTHING, even if YOU made it, YOU are the only person who will ever see it in any way in practice, and YOU really are bothered by it. It doesn’t matter. You’re our bitch. You will keep it once you created it. You don’t get to delete or even hide it. We own you. Understood?”

As you can imagine, I got furious when the “Web Hosting Manager” SAFE APP didn’t have any “delete” button or anything. That’s what prompted me to attempt to look into this, but nobody has managed to give any sort of explanation that makes any sense. I really get the feeling that the secret purpose of a lot of software is not to provide something useful, but to fuck with and anger the users. Not saying that this is the case with SAFE, but it really gets me angry when I even suspect this kind of attitude.

If this truly is a conscious choice and design decision, believe me: it won’t fly. And there is ZERO technical reason why it “must be this way”, which I suspect will be a reply. No, it doesn’t. The whole point is that SAFE automatically and securely knows which “account” owns data… or so I thought.

2

The public ID deletion issue is associated with the test/alpha networks. I am of the understanding that this will be possible later on.

Although the key pair you create that is behind an ID will not be. But like Bitcoin keys you can just “forget” them since they will not for all practical considerations be created ever again.

For MD data you can delete, or will be able when full functionality is built.

For immutable data there is no owner of any particular chunk so the network cannot allow deletion of chunks since there is no way to know if you have the right to do so. This is intentional behaviour.

When you create a immutable file the network returns a datamap to the file and this datamap contains the information necessary for chunk order and locations. From this your can retrieve the file and the client can use the self-encryption module to decrypt the file for you.

If you modify the file the client will store the changed chunks and you will have an updated datamap.

You can effective delete your own file by “forgetting” the datamap. If the datamap is stored in a MD then you could delete the MD.

Noone can access your own files unless you release your datamap to them.

Public files have their datamaps publicly accessible.

Because you may not be the only one who uploaded the file. So the chunk maybe shared between multiple people. The network doesn’t know and the reason it doesn’t know is for security and anonymity.

Well if you store a file then its not Google or anyone elses, so you can just delete/forget your datamap and that data is gone.

If google was to do the same program on SAFE then it would own the data and could do the same thing to you. The advantage of SAFE is that people will be able to say to google get lost, we will own our own data.

Its alpha software, and only demo software not meant for the general market place. Delete will be a part of any sensible final product running on SAFE doing file storage such as a web hosting manager.

8 Likes
3

There are many reasons to do this in many designs. I will try and answer, but please understand it means I must try and ignore the genius / expert comments you are making and doing so in a poor way. If you ask politely it will reap much more rewards and then it wont be confrontational. We listen to the community and vice versa, but imagine you are standing in front of the person, if possible use that same language/courtesy.

SAFE has 2 data types,

  1. Immutable Data, this cannot be deleted (at the moment)
  2. Mutable Data, this is owned data.

MD at the moment cannot be deleted, but will be, but it’s not as simple as you imagine. Say for instance you have an MD, it points to your web site, youo decide that you don’t want all that any more and delete the MD, I recreate the same MD (different owner) and just republish your stuff, or worse. So there have been options like using versions to distinguish MD versions, use blank not delete RPC’s etc. It all boils down to a deeper decision regarding these types. So it’s not a simple simple data base entry we own, it’s much deeper and has to be. Delete will do what the user thinks, but may in fact use methods below the API that mutate the data instead of delete it. It all relates as well to the publish forever attitude we have, where data cannot be removed/edited after publish. For private data it will look like it deletes, whether we carry around “dead” data or not is not a simple question and has no simple answer, look at BTC for dead data, sometimes its required ,however later can be designed out (like blockchain efficiency decisions in the last few years)

Hope this helps

19 Likes