Feedback and Questions on the SystemDocs

4.1.1
Everything after NetWorkViewer is not visible in the gitbook the way it is written right now. Please let us know if these should be visible in the gitbook.

4.1.4

This means that data is stored at the nodes which are among the firstParameters::node_group_sizeclosest to a given group ID.

  • First or closest?

New node(s) will become closer to the group ID and will become amongParameters::node_group_sizeclosest to the group ID.

  • I don’t understand that sentence at all

Node(s) appearing in the logical group : Some of the group membernodes will move far from the group ID and will not remain underParameters::node_group_sizeclosest to the group ID.

  • Is that [Parameters::node_group_size] supposed to be a part of the sentence? I think I get the meaning, but these sentences are very confusing.

Nodes close to a segment of network will notice churn event by observing a change in their groupmatrix.

  • What segment of network?

I think the formatting is the issue here, the docs only mention the existence of the visualiser. I have amended the page to make this clearer.

It should be closest. Thanks for the correction, docs updated.

Yes it was confusing, I have amended.

I this would read better: Nodes close to a segment of network experiencing a churn event will observe a change in their groupmatrix.

It is meant to be there but could probably be phrased better. I’ll think about how to make this clearer.

Thanks for your help keeping us on our toes!

2 Likes

Node(s) appearing in the logical group : Some of the group member nodes will move far from the group ID and will not remain under Parameters::node_group_size closest to the group ID.

This is a good example, didn’t see that before. Will they not remain under a certain size of the group or will they not be close to the group ID anymore? Maybe you could try and rephrase that (or anybody else who gets the meaning of that sentence) and hopefully this will make the purpose of the Parameters::node_group_size clearer for the other examples as well.

Thanks for clearing that up Nick!

One other thing I stumbled upon. We should probably refrain from using the word “should” in the context of anything describing the inner workings of the network. Otherwise it sounds like we’re not sure if this is working as we think it should, or that it depends on the node being in the right mood, which I seriously doubt :stuck_out_tongue:

Case in point:

These Node(s) should delete data which they are no longer responsible for.

One last thing for today:

Here, the ClientMaidNfs is defined as an instance of NetworkFileSystem, with policies PutDataPolicy, GetDataPolicy, DeleteDataPolicy and ClientPostPolicy, all class templates, each taking template arguments for a signing identity and the persona type of the object, a Maid and Persona::kClientMaid, respectively, in this case.

a) That is one long, complicated sentence
b) what is a maid? (in this context)

I fixed the text around churn events so hopefully it makes more sense now.

1 Like
The requirements of the resilient SAFE Network are as follows:
...
* Dealing with very high levels of churn (computers going off and on, potentially staying off)
...

I can’t accurately get the word churn 's meaning and the whole sentence as well. Please explain :innocent:

Let’s say you have a network of 100 computers online at the end of week 1. At the end of week 2 you have 90 computers online, at the end of week 3 80 computers online. That would mean you have a churn rate of 10 computers per week.

I guess fluctuation could be used to describe it, although they’re not interchangeable. It’s also used to describe how many customers you loose (though I can only say that with certainty for the telecommunications industry)

1 Like

I’ve got it now, thank you! :smile:

1 Like

It’s about proof of resource mechanism.

 In this case the checking mechanism does not require to know the content of any data to be checked, but must know the data is in fact held and held in a manner that is accurate.

I don’t understand the second part of the sentence. Does it mean proof of resource must know that data is accurately stored?

This mechanism is triggered on Get requests and during account transfers. 

What are account transfers exactly?

Vaults are rewarded for supplying resources by being awarded safecoins at a rate proportional to the Vault's rank. As a result users are not locked to a particular Vault.

I also can’t get the second sentence here.

Users do not require to have their account managed, even in a secured manner.

Em… managed by whom/what? Of course, users don’t want it, or what is meant?

Rank is an internal measure of the Vault's stored versus lost data.

Stored versus lost data? This means stored data as compared with lost one, right?

Yes, the network will know that the correct data, verified by the hash, has been stored, although it won’t know what the data is. This process is explained in the 5 steps listed.

During churn events, as nodes are turning off and on, data chunks, or information about nodes (for example) are passed to continually changing close groups. This is account transfer. I have updated the docs to make this clearer.

One user can create multiple accounts and therefore have multiple vaults

This point is just emphasising that unlike a centralised service, the SAFE Network does not require data access / integrity /security to be managed by a third party.

Yes, the rank is comprised by the network randomly checking that it can retrieve chunks stored on vaults and sets that against chunks that it could not retrieve. Lost chunks in this sense means a data chunk could not be retrieved during the POR process.

4 Likes

Just as a reminder, we should think about replacing the word maidsafe in the SystemDocs. Should we replace it with SAFE Network or decide that on a individual basis?

I have to say that I’m still a bit sceptical about the name. First of all, the project became known as maidsafe. Second, the name SAFE network is not really a unique name. I’m not a big fan of these fancy sounding names or big marketing strategies (as Doug Stanhope famously said: A good product doesn’t need advertising. Or have you ever seen an advert for cocaine?) but I do think that establishing a good brand is not something that should be dismissed.

Just my random thoughts on this. :wink:

2 Likes

I’ve removed all reference to MaidSafe in the context of the technology (i.e. MaidSafe client becomes the SAFE Network client) in the System Docs. Hopefully I have caught everything.

2 Likes

4.1.9

And we do not appear in the first K

K = 1000?

MAID (MaidSafe Anonymous ID)
PMID (Proxy Maidsafe ID)
MPID (Maidsafe Public ID)
MSID (Maidsafe Share ID)

What’s the name for that? Just SAFE? Replace the M with an S?

or would-be hackers

can we please remove that? That doesn’t sound convincing nor professional and it doesn’t go into specifics, so I think it shouldn’t be there :smile:

If no record exists then each of the Data Managers of this group closes to the NAE being stored selects a connected node as a PmidNode to store on (if the message has a Data Holder hint then the closest of the data managers of this NAE attempts to store there).

I guess that should be closest, right?

  • What is a ManagerDb/AccountDb? It’s only mentioned twice in 4.1.9, nowhere else.

There is a mention of MaidNode and MaidManager, so we should probably look for maid instead of maidsafe.

PS: I just addded a sentence to the first page that states:

Please be aware that this book was translated by community members. It may contain translation errors. If you should see an error, please provide feedback to the community which you can reach here https://safenetforum.org/.

I hope that’s ok, and I think we should add that to all the translations.

That’s it for today. I should be finished with the translation sometime tomorrow, 4.1.9 is really a lot of text :stuck_out_tongue:

2 Likes

Fantastic job and perfect timing @hillbicks !!!

4.1.9.

  • I’m not sure what is meant by cost? Is it the “amount of safecoin” that has to be paid in order to ‘PUT’ data onto the network?

Important data is faster when it’s important and slows downs otherwise, making good use of network resources.

  • I think we’ll have to see how that pans out in reality, but it looks to me now, we’re breaking net neutrality with that design. This is just as a reminder for myself, because that would be really bad in my opinion. Link to previous thread that talks about this Net neutrality

Close nodes can become temporary stores of data very easily, allowing many advantages to system designers.

  • such as?

One thing I’d like to mention again is the use of words like may and should. I’m not sure why you choose to use these in the SystemDocs because it seems kind of strange to me that there are scenarios where it is not clear what will happen. I’d think I’m not the only one who raises an eyebrow at phrases like that, because there is no further for those scenarios. The only explanation I have right now is that it’s assumed to behave that way, but further tests are need to ensure that behaviour, which is totally fine by me, but I think we should clarify that. Or am I missing something here?

PS: @happybeing: Why perfect timing? Do you know something we don’t? :open_mouth:

I have no privileged information, but my sense it’s that things are about to happen. Let it be a test of my intuition :wink:

2 Likes

:smiley:…aahhmmm mmmfff…the new socially improved nicer me has just bitten my tongue to shreds…lol. Only joking…hope you’re intuition is right… :smiley:

Hehe, glad to see you not asleep on the job. :slight_smile:

1 Like

OK @dirvine, I’m back from hols now so it’s OK to start testnet 3 :wink:

4 Likes