問題-19:
Hello, when will Filecoin's economic model be released?
----------------------------
你好，Filecoin的經濟模型什麼時候發佈?
回答-19：
Cryptoeconomics will be announced once we have finalized all mechanisms.
————————————————
一旦我們完成了所有的機制，將會公佈密碼經濟學。
問題-20:
There are four filecoin implementations in process. All of them have to be compliant with the spec. We had defined the mainnet launching milestones. The question is when is the spec finalization date? We will expect between the date and mainnet launch, the spec will keep no change.
————————————————————
有四個filecoin實現正在進行中。所有這些都必須符合規範。我們已經定義了mainnet啟動的里程碑。問題是規格定稿日期是什麼時候?我們將期待在日期和主要網推出之間，規範將保持不變。
回答-20：
The protocol spec has stabilized -- though the document itself needs many more improvements to be very clear and readable. We will improve it over the coming weeks, and we will base the Filecoin Improvement Proposals (FIPs) process on top of it. -- juan
——————————————
協議規範已經穩定下來——儘管文檔本身需要更多的改進才能非常清晰和可讀。我們將在未來幾周內改進它，並在此基礎上建立Filecoin改進建議(FIPs)流程。——胡安？
回答-21：
我們很期待疫情過去之後來中國和大家見面！
問答-22:
Will the mining hardware support AMD's GPU before the mainnet launch, such as rx580, rx480, etc.?
————————————————————————
礦機硬件是否會支持AMD的GPU在mainnet發佈前，如rx580, rx480等?
回答-22：
We're focussing on NVIDIA for now, AMD might not be happening before mainnet.
——————————————
我們現在關注的是NVIDIA, AMD可能不會發生在mainnet之前。
問題-23:
How many developers are currently working for the mainnet launch?
——————————————
目前有多少開發人員正在為啟動mainnet而工作?
回答-23：
The Filecoin project itself has around 25 engineers working full-time on the project, but that's just a small fraction of people working in the ecosystem. Hundreds of engineers and researchers have contributed code, research and pull requests to all four implementations, Filecoin proofs, and dozens of projects building on Filecoin.
——————————————————
Filecoin項目本身約有25名工程師全職從事該項目，但這只是在生態系統中工作的一小部分人。數以百計的工程師和研究人員為所有四個實現，Filecoin證明以及基於Filecoin的數十個項目貢獻了代碼，進行研究並提出請求。
問題-24:
Is it a hard condition for the 256GB of memory in the future?
-----------------------------
對於未來256GB的內存來說，這是一個有挑戰的條件嗎?
回答-24：
Great question.
Our implementation makes some particular tradeoffs, for example, it prefers using more RAM over doing more computation in order to have a "fast" sealing, however this does not have to be the true for all miners. We expect contributors will optimize this code for what it works best for them.
There is a lot of space for tradeoffs, just to give you some example:
- Most of the RAM is currently filled with the DRG graph, however, this edges of these graphs could be locally computed. This means that you could free up large amount of memory at the cost of having some of your CPU regenerating the edges of the DRG graph.
- RAM is very fast for computation, if you are willing to have a slower sealing, you could use other storage devices (e.g NVMe) which could help increase throughput of your CPU.
I cannot guarantee that the above optimizations can lead to useful optimizations for your mining setup.
We will also try to explore different tradeoffs with our main implementation.
——————————————
好問題。
我們的實現做了一些特別的權衡，例如，它更喜歡使用更多的RAM而不是做更多的計算，以獲得一個“快速”的密封，但是這並不一定適用於所有的礦工。我們希望貢獻者能夠優化這段代碼，使其最適合自己。
這裡有很多權衡的空間，只是給你一些例子:
-目前大部分RAM都被DRG圖填滿，然而，這些圖的這些邊可以在本地計算。這意味著您可以釋放大量內存，代價是您的一些CPU重新生成DRG圖的邊緣。
內存是非常快的計算，如果你願意有一個較慢的密封，你可以使用其他存儲設備(e。可以幫助增加您的CPU的吞吐量。
我不能保證上面的優化可以為您的挖掘設置帶來有用的優化。
我們還將嘗試在主要實現中探索不同的權衡。
問題-25:
Does mining consume a lot of electricity? What kind of resources are required
————————————
挖礦會消耗很多電嗎?需要什麼樣的資源？
回答-25：
Computer use a lot of electricity! But, in general, much of the electricity powering Filecoin mining operations is contributed to providing storage. It's different from pure proof-of-work cryptocurrencies in that a large portion of the energy used goes towards _both_ providing a useful real-world service and securing the blockchain.
——————————————————
電腦很耗電！但是，通常情況下，Filecoin挖礦的大部分電力都用於提供存儲。它與純工作證明加密貨幣的不同之處在於，它將大量的精力用於提供有用的真實世界服務和保護區塊鏈。
問題-26:
Will there more q&a sessions like these before mainnet ? I think these are great. Once per month ahead of launch would be ideal.
————————————————
在mainnet之前會有更多類似的問答環節嗎?我覺得這些很棒。最好每個月提前一次。
回答-26：
We are keeping our heads down racing towards mainnet launch but we will try our best to have more Q&A sessions in the future!
——————————————————
我們正在努力推進mainnet，但我們將盡我們最大的努力，有更多的問答環節在未來!
問題-27:
Does each node need an independent public IP after the mainnet goes online？
-------------------------------
網絡上線後，每個節點是否需要一個獨立的公共IP ?
回27:沒需要
問題-28:
1. The international epidemic is very serious, the impact of the epidemic on the United States, whether there is still an impact on the launch of the main network?
2. Can the first stage be successfully completed? Are there any new changes?
------------------
1. 國際上的疫情很嚴重，疫情對美國的影響，是否還有對主網上線的影響?
2. 第一階段能順利完成嗎?有什麼新的變化嗎?
回答-28：
The epidemic is very serious. We have made our plans to work through the epidemic, and don't anticipate any impact. However it is possible that if the situation gets drastically worse either our development teams or the community of miners may be severely affected. We will see how things develop. We hope everyone will be safe and not get sick! After launch, we won't be able to do much -- the network will go on with whoever is able to participate. Hopefully this epidemic will slow down soon! Stay safe! -- juan
————————————————————
疫情非常嚴重。我們已經制定了應對疫情的計劃，預計不會有任何影響。然而，如果情況急劇惡化，我們的開發隊或礦工社區可能受到嚴重影響。我們將看到事情如何發展。我們希望每個人都是安全的，不要生病!在發佈之後，我們將不能做太多——網絡將與任何有能力參與的人一起繼續運行。希望這種流行病能很快減緩下來!保持安全!——胡安
問題-29:
Hi @Mr. Juan Benet, now the barrier for common people to participate in is really high. What will you do to lower the barrier and how will you balance the main net safety and popularity?
——————————————————
嗨@Mr。貝內特(Juan Benet)，現在普通人參與的門檻真的很高。你將採取什麼措施來降低這一障礙?你將如何平衡主要網絡的安全性和受歡迎程度?
回答-29：
Three answers
*Storage Mining*: We expect storage miners of many sizes. We expect small storage mining to be possible -- in the ranges of 100s of TB. We expect it to get easier as Proofs-of-Replication get cheaper to do. That said, most data archival wants to be in safe, well-maintained storage racks.
*Retrieval Mining*: smaller miners that are well located will be able to do retrieval mining -- store data and serve it. This is less developed right now, since we want to get everything out sooner, but you'll see some of this working in a more fleshed out way soon. You can already pay for retrieval with payment channels, now anybody can store things and serve them. We expect very small miners to be able to participate in this way.
*Other roles:* we expect other mining roles to develop over time -- many kinds of computation or activity in the network can develop over time. For example things like Repair, or caching, or aggregation can become mining roles for smaller miners.
——————————————————————
三個答案
*存儲挖掘*:我們需要各種尺寸的存儲挖掘工具。我們預計小型存儲挖掘將成為可能——在100兆TB的範圍內。我們期望它會變得更容易，因為複製的校對變得更便宜。也就是說，大多數數據檔案都希望放在安全、維護良好的存儲架上。
*檢索挖掘*:位於良好位置的小型礦商將能夠進行檢索挖掘——存儲數據併為其提供服務。現在這個還不太完善，因為我們想讓所有的東西都儘快出來，但是你很快就會看到其中的一些更充實的工作方式。你已經可以支付檢索與支付渠道，現在任何人都可以存儲的東西和服務他們。我們期待非常小的礦商能夠以這種方式參與。
*其他角色:*我們期望其他挖掘角色隨著時間的推移而發展——網絡中的許多計算或活動都可以隨著時間的推移而發展。例如，諸如修復、緩存或聚合之類的東西可以成為小型採礦者的挖掘角色。
問題-30:
Is there any research to avoid SNARK? I believe it's too heavy for the hardware requirement.
————————————————————
有什麼研究可以避免SNARK嗎?我認為它對硬件要求來說太重了。
回答-30：
This question although quick and short, it's a very deep and important research question. I took some time to reply and I will split it in two questions:
Q1: Can we remove SNARKs from our current proofs?
We use SNARKs to "compress" the proofs our Merkle tree proofs and checking the correctness of some SHA. If we were not to use SNARKs, our proofs would be in the order of hundreds of kilobytes instead of a hundreds of bytes.
This is because our vector commitment is a Merkle tree and their inclusion proofs, although logarithmic are still very large (in our proofs a tree has depth 30, which would mean that a single Merkle tree inclusion proof is about ~1kb - we need hundreds of those!)
Replacing Merkle trees with better vector commitment could help us avoid using SNARKs. We have actually active researchers and recently released an update on ePrint (https://eprint.iacr.org/2020/149) on better RSA vector commitment. Unfortunately, they are far away from practical :(
————————————————————————
這個問題雖然簡短，但卻是一個非常深刻和重要的研究問題。我花了一些時間來回復，我會把它分成兩個問題:
問題1:我們能從目前的校樣中去掉這些問題嗎?
我們使用SNARKs來“壓縮”Merkle樹的證明，並檢查一些SHA的正確性。如果我們不使用snark，我們的證明將會是幾百kb而不是幾百字節。
這是因為我們的向量承諾是一個Merkle樹和它們的包含證明，儘管對數仍然是非常大的(在我們的證明中，一棵樹的深度是30，這意味著一個Merkle樹的包含證明大約是~1kb——我們需要數百個這樣的證明!)
用更好的矢量承諾來代替Merkle樹可以幫助我們避免使用snark。我們有活躍的研究人員，最近發佈了一個更新的ePrint (https://eprint.iacr.org/2020/149)更好的RSA矢量承諾。不幸的是，它們離實際還很遠
閱讀更多 TREE東森 的文章
