Will IPFS be in jeopardy in 2020?

星际视界IPFSNEWS

特邀专栏作者

2020-03-27 16:49 (UTC+8)

This article is about 1987 words, reading the full article takes about 3 minutes

The IPFS future is within reach.

AI Summary

Expand

The IPFS future is within reach.

Previously, silent-observer wrote an article "The Unstable State of IPFS in 2020". The article stated that IPFS will be difficult to develop in 2020, and pointed out the current difficulties faced by IPFS and the shortcomings of the development team.

On March 25, Molly Mackinlay, the head of the IPFS project, published an article in response, which was recognized by many IPFS community partners, including the handsome co-founder of Textile Andrew Hill. (The official Filecoin team had the first dialogue with the co-founder of Textile）

The original text of the response is as follows:

1. IPFS is stable and growth is the core

In short, I can't agree more with your statement that "IPFS is in an unstable state".

I think it's obvious that IPFS has tremendous momentum this year. Whether it's the core developers or the entire ecosystem. Both IPFS and libp2p have seen huge performance gains.

First of all, Testground is a necessary and huge step forward for us to test and simulate network performance (not only for IPFS, but also for all p2p networks) and make necessary performance improvements to content routing and DHT.
There is a new documentation site to make it easier for new IPFS users to learn about IPFS; making sure the core working group brings important features into the community (roadmap here).

Second, there are a lot of exciting advancements in the ecosystem.
For example a new rust-ipfs implementation is being developed, Terminal.co has a pretty good netlify equivalent on ipfs; UnstoppableDomains has ipfs native browser and wordpress templates, and Audius, Qri, Texitle, 3box, Pinata, fission, cypress Tiny, Anytype, Temporal, and a ton of other groups all made exciting announcements in the first three months of the year.
I think the team and the community are growing faster than ever. This is the furthest thing from an "unstable situation" - if we're on a high growth trajectory!
Finally, as the project lead, I have spent most of my time working on the IPFS project over the past year. Frankly, this has been the strongest year for the team and the community during this time.
I think your warning: any decision "could make IPFS a thing of the past" is anachronistic, alarmist and wrong. I think you're trying to help, but writing anonymous, sensationalized articles doesn't really help me with your problem, nor does it help the team focus on releasing the biggest go-ipfs release (alpha) since 0.2.3.

2. Developer activity is (relatively) organized

I think your misunderstanding of how the IPFS working group coordinates comes partly from perceptions of working groups, some corrections here:

We don't have the same engineers on ResNetLab, Testground, and Hydra (the new enhanced nodes required for the 0.5 network upgrade) - each has dedicated owners, although some particularly prolific engineers help with multiple projects (especially helping push priority projects).

Although the Testground team did a retrospective at the end of Q4, a more detailed product plan could help us narrow down exactly what's in/out of scope much faster - we definitely have a development roadmap and a very Detailed specifications.
Plans do not always stand the test of implementation. So the pain points pointed out were re-scoped and narrowed down to the most critical MVPs. The reason teams do retrospectives (which we do publicly) is to learn and iterate, and to summarize the lessons we've learned for other/future teams.

The IPFS Working Group has a very active core set of contributors across many different organizations, which has not been depleted by focusing on other projects. As you can probably see, this year we're focusing more on energy (and thus pausing work on some less-than-perfect projects), but it means we have more people than ever working on the things that matter most Improve.

3. The price of IPFS is reasonable

As far as "IPFS is expensive", there are plenty of places where you can save on IPFS.

Incentivized networks such as Filecoin/Storj that allow more users to contribute unused storage space can help reduce the unit price of storage, but there are other savings - people storing large, highly overlapping datasets (such as those in OpenSuse) package managers of the class) can benefit from deduplication savings from snapshots of the data they store, while other groups (such as Netflix) can see performance gains (and bandwidth savings) from direct p2p transfers between the nodes they run on !).

The cost per byte of an "IPFS service provider" is not a good way to judge the overall cost of IPFS, and I think it can also be significantly lower than the costs we see today.

4. The usability of IPFS is improving

IPFS is a rapidly growing open source project, and I agree that we haven't entered an area of "easy to use" yet, especially for non-developers. Much of the long-term vision is still a work in progress and is being refined every day, and there are plenty of opportunities to get involved to help.

A quick look at the forums I have revealed is more exciting than having people frustratingly trying to debug bugs - people getting tips on how to try new things and move the ecosystem forward.

We're not great here yet, but we've come a long way, and there are tons of groups (like "Texitle", "3box", "Terminal.co", etc.) creating projects that bring IPFS into more areas tool.