Adding 1 byte or 1 second to the start (1st chunk) has a flow on effect assuming the file is larger due to the addition.
4 Likes
I thought there was a work around to this in self-encryption to better encapsulate chunks and not have a cascading effect?
like this:
One of the interesting use cases of the rolling hash function is that it can create dynamic, content-based chunks of a stream or file. This is especially useful when it is required to send only the changed chunks of a large file over a network and a simple byte addition at the front of the file would cause all the fixed size windows to become updated, while in reality, only the first âchunkâ has been modified.
The âchunkingâ is executed on the client doing the uploading. He could modify the chunking, maybe even insofar that by using a different chunking algo he would create completely new chunks without even the need to modify the actual content.
3 Likes
If an uploader did this other clients wouldnât be able to decode the chunks to read the files.
hmm? a filemap is just a list containing links to chunks, that can be put together by simple concatenation. The chunking algo sets the size of the chunks, why should that make decoding fail. Theres no other use for enforcing the chunking schema by the downloader, beside this issue, and hinders future progress.
1 Like
I appreciate that there may be strong resistance to a long-contemplated network design element, but I donât think PtD, PtP, or any other Pt{ } should be implemented at all. I think that doing so is risky on many levels that could provide a similar network without such trickable bloat a real opening to supplant the original. Briefly:
-
Security
These additional Pt{ } could increase the potential attack surface in ways currently unseen compared to if the only way to earn tokens is by providing nodes and nothing else. -
Perception of value (and fairness)
If the token can be obtained by tricking the network, perception of its value will suffer. If the Safe network is to become a force of nature, it has to be kept simple and un-trickable, with no baked-in advantage for anyone, not even for the creators of it. See bitcoin and satoshi.
(Specifically re:PtD, I understand that MaidSafe might want that additional 5% of the token to be able to support continued network development, maintenance, etc. So maybe take it upfront at genesis and not continually into perpetuity given potential appearance risks, not to mention potential questions as to the fairness/need/use of that tax should the tokenâs value increase dramatically). -
Better alternatives to PtD, PtP, etc.
Given the way the space has evolved, itâs clear that developers and content creators have plenty of ways to monetize their work including token issuance, NFTs, tipping, etc. In fact, many would prefer to have the freedom to set the price for their work rather than have the network decide for them. I would therefore say that itâd be preferable to give devs and content creators the ability to issue their own tokens as a replacement for Pt{ }.
9 Likes
Had been contemplating a much more poorly worded reply attempting to express the same thing as @Bogard here, very nicely put.
A more fair and honest solution seems to me to be paying people directly for their work. I think this would be surprisingly powerful if it were really an easy option, anonymous, and you knew your payment were going to the person you wished to give money to.
This would be very possible and elegant if there were a solution to another tricky and oft-debated issue which Iâm not sure where we are at with technically, namely, proof of unique human. It would be beneficial for the typical use case that people imagine, like the singer or the author, but also the whistleblower, the person looking for crowdfunding, the community looking to get some project off the ground, the journalist, the documentarian, the legal defendant, the political group, the researcher.
Cutting out the middlemen effectively in all these scenarios seems more far-reaching, and would effectively take care of the struggling artists and writers and software developers (hihi) too I feel.
One place people might disagree: I am assuming that it is a falsehood that people only pay for things under threat of legal repercussions or fear of social ostracisation or something, because I think it is. Enough people understand and appreciate the work of othersâ to pay, we can manage a few freeloaders and poor people getting things for free.
I think a system like this could genuinely work if it were very easy and very anonymous and very sure your payment goes where you think itâs going.
3 Likes
Of course some will pay, the issue is how many, how much, and also whether this achieves better or worse results than PtP (such as a fairer distribution, easier operation, efficiency etc).
This is why we should try both, or rather a range of different approaches. I donât think any of @Bogardâs points are strong, and not enough to say we shouldnât even experiment with PtP. Weâve had endless debates about this in the past and no clear resolution because views are too much based on opinion and speculation, and not enough on experimental evidence.
2 Likes
Please keep the Safe Network simple and let the different dapps on top of SN evolve the network. The focus should be to make SN flexible, so different dapps use cases can be implemented.
3 Likes
Not everything can be done in an app. If you keep it to simple it will be less useful. Decentralised computing for example will probably need enhancement in the network itself.
Itâs fair enough to say thatâs enough, but how can we decide that without first doing experiments to understand the parameters?
1 Like
I am all for experimenting if there are available resources to do that. Also updating the network to adopt different use cases for apps. But making a general decision on PtP on a network lever is in my opinion a mistake. But making changes to SN so that an app can use a PtP function that can be tweaked is just making the SN more flexible. I am all for that.
2 Likes
PtP is a network level feature so experimenting means doing so at the network level, which is why I made the comparison with decentralised computation / smart contracts.
1 Like
Then PtP is a mistake in my opinion and interferes too much with the fundamental incentives of the network. Let the network grow based on the fundamental incentives of privacy and freedom. Pay ones and store forever. I agree with @Bogard
4 Likes
In which case smart contracts are also a mistake?
Either way you are making a value judgement about something (PtP) without the data which experimentation would provide.
This has been the position of a surprising number of people here. The counter position would be to insist that PtP be implemented without first gaining this data. Neither position makes sense to me.
I donât understand why people are willing to take strong positions without data.
Smart Contract adds a new function that you can choose to use and doesnât come in conflict with the other use of the SN. So I donât see the compersion.
No, I donât want to make a value judgment and I donât want SN to make it either. That is the main point for not making it a part of the fundamental network.
2 Likes
Are you saying PtP conflicts with something, if so please explain?
You are IMO making a judgement on the size of risk you see, versus any potential advantages. And doing so without experimental data on what PtP might involve in implementation and how useful it would be compared to alternatives.
1 Like
I agree with you here, @happybeing. I think experimentation for Pt{ } can be done, but on a separate test network, not on the main network. I actually originally liked the Pt{ } aspects of the Safe network years back, but my views have evolved from a better understanding of what makes the average person give value to something (counter to my own biases) and how the decentralized space has evolved, which offers a picture into what people/devs want. They want their tokens and the freedom to trade and set their own prices, but do it all fast and without fees, which the Safe network can give them (on top of all of the other advantages which we all know and love here).
Edit to add: an alternative could be to set up two networks (one with Pt{ }, the other with token issuance) and see which one is adopted.
3 Likes
I donât think itâs achievable for an app trying to replicate PtP to get the same anti-spam measures (caching) as at the network level. Thatâs why itâs important and still something that is constantly being resurfaced. Also would more seamlessly integrate with your SafeIDâs linked wallet.
1 Like
sod the libertarians
I understand there are a lot of issues that need to be solved and new solutions need to be made to overcome these challenges. We all agree that it must be possible for the âproviderâ to get paid for their content/service etc. I believe that this must be solved on the app level. I donât think SN should be an experiment in complex economics and incentive structure.
2 Likes