Excellent update, a big thank you to the team, moderators and testers for their relentless pace and very hard work! 
Congratulations to @shu on his well-deserved addition to the Maid team! And congratulations to Nic for joining probably the best project in the world, welcome! 
The long awaited docs.autonomi.com is a great job, after reading the document you should be recognised.
No Safe, no wave.
He better not play the ābetween the code and the product, I am the interfaceā-cardā¦
Welcome Nic !
Weāll get there! (and so much faster with shu on board!)
First things node.
Fantastic progress on many fronts againā¦ the pace is impressive!
Regarding the DAG, some have asked previously if thereās a risk of the DAG becoming a significant burden on the network if transactions are free.
What is the teamās current thoughts on whether thatās likely to be an issue / attack vector (e.g. thousands of nodes making constant free transactions that all permanently fill the DAG)?
Hoping there are measures that counter this / reasons itās not going to be an issue 
Itās going to be an exciting few months following developments
With the network being meant for the world (upto 8 billion people) then those thousands per second will pale compared to the transactions done by people.
Now that the scale of the problem is even larger the question is more important just for daily operation.
Also, wasnāt it so, that making a transaction is a bit of work, so even though itās free to send, it consumes some resources before it can be sent? Resources from the device, not from the network.
This may be the case, but a small bit of work vs a transaction that needs to be stored forever on the network without nodes getting paid for that storage could still cause issues.
I would expect it would be best for transaction fees to at least pay for the permanent storage of the transaction data on the network at the going rate for putting data on the network. Anything less could lead to spamming, which wouldnāt be a big issue for network operation if it could just be rate limited to stop it bothering the network, but with the DAG, thereās permanent data being stored for every transaction, so it could become unwieldy / a big burden on nodes (DAG nodes or any node that stores DAG data).
Yeah. Though, if a āspamā client were able to create, for example, 10,000x the transactions of a normal user for free, you wouldnāt need that many spam clients to create a significant portion of DAG entries.
Spamming isnāt free just because transactions are free. Thereās still a cost.
The question is whether or not that cost is high enough, and if not what will the overall impact of having a fee be. Both aspects need to be studied, and any alternative mitigations considered alongside them.
The easiest spam is to āGetā just random addresses of non existent or existent data. The argument then would be pay for Get
I see it the same, we will put up with some spam and take some measures, but ask nobody to pay for transactions or reading data. They feel the same to me. Although transactions have a trail it need not be all transactions on every node, we have auto sharing here at the fundamental level and we will use that to great granularity.
Just wondering how the DAG will react once we are over a billion people using the network with the huge number of transactions happening. How is the DAG scalable? That is one area I havenāt explored yet
Thatās a significant difference. Spamming gets wonāt cause much disruption, or at least will have no enduring effects beyond any attack.
But spamming transactions could make the DAG swell to the point of being a burden on the network, and in my view, it needs to be clear that this canāt become a fatal threat if mitigations are not going to be considered (e.g. fees, or DAG pruning if thatās a thing).
If it is clear that the scale of the DAG canāt become a problem for the network, itād be great to hear why this is the case.
Charging a tiny amount for gets might also become something to consider if spam there becomes an issue, but given how spread out data is, and how many nodes there are likely to be, hopefully that wonāt be an issue. One node downloading at 100% canāt flood the connections of hundreds of nodes, but I guess a botnet all getting the same specific file could perhaps DDOS nodes hosting a specific file for a time, if caching etc doesnāt overcome that.
In the case of āget spamā, I guess nodes that host a targeted file could start charging fees for gets after a certain number of gets for specific hugely requested chunks within a given time?
True, but to the network, itās only cost and no revenue. And, to botnets, or any client, I expect it would likely be very cheap to spam large volumes of transactions.
I agree.
Given the tiny size of transaction data, I expect the cost to the network of storing each transaction will be so low that the fee would be pretty insignificant to the user while being big enough to offset the cost to the networkā¦ but yes, it all needs to be considered to ensure that either having no fee is not a threat, or that any fee is the required magnitude to offset any threat.