Gopherfest 2017 | Upspin with Rob Pike

Upspin is an experimental project to build a framework for naming and sharing files and other data securely, uniformly, and globally: a global name system of sorts. It is not a file system, but a set of protocols and reference implementations that can be used to join things like file systems and other storage services to the name space.
Performance is not a primary goal. Uniformity and security are.
Per Rob: Note that this is a project IN Go not ABOUT Go
Gopherfest 2017 Playlist → goo.gle/3Ee0nQh
Subscribe to The Go Channel → goo.gle/go-subscribe
#GoLang #Gopherfest

[@brooklynhousepdx8267]

Even if you don't write gocode, I consider Rob Pike speeches to be engrossing and philosophical. We are a very lucky community.

[@ethangarnier8416]

I am so happy to be a part of the GO community, by far the best. This technology is amazing and can't wait to start using it

[@dplameras]

Upspin is a Plan9 Idea. If you don't know what Plan9 is, its a OS that was going to be a successor to Unix. Another cool concept of Plan9 is that IP sockets are files. Everything is a file :) Where have you heard that before...

[@defnlife1683]

and it also had a lot of namespaces.

[@morrisonbrett]

Awesome idea and thanks for another great presentation from Rob Pike.

[@agsantana]

Really cool idea, it's solves a real problem that many of us have. Hope it sticks.

[@llitfkitfk]

"Don't break the law, but I believe that on the flip side I don't want the law breaking my data either." 24:33

[@Tracks777]

I look forward to more videos

[@john_rambo_27098]

I like his specs...funky

[@fungussa]

Two questions: 1. Possibly a premature question, but how well will Upspin cope when a large number of users have been granted access to a file? 2. And given the data owner's name, will users be able to browse/discover the files that have been made accessible to them by the data owner, rather than the data owner having to message the users about which files have been made accessible to them?

[@Mike-iz9kh]

He mentioned that "list" is one of the permissions that can be granted, which I think answers your question #2.

[@gokukakarot6323]

I just want to say. As much as I don't like that I don't like Golang, I absolutely love Rob Pike and Matz

[@erichstocker4173]

The problem with all this cloud business is that when I want to use a file or service I might not have access to the network or to buy access to a network would cost an arm and a leg. When I have lightroom, its catalog and the raw files I want to work on. I can work whether I have a network connection or not. Also many of the network connections are painfully slow or come and go, etc. I've paid to have access on a flight and midway into the flight the connection was dropped. If my photos or music, etc. were on the cloud I would have no access and the rest of the flight would have been boring. While upspin is a good idea, the current network infrastructure isn't universal, equally fast or dependable.

[@mishasawangwan6652]

Erich Stocker rest of the flight would be boring eh? first world problems .. :P .. snide-ness aside, yeah i agree this cloud stuff and services as services crap is lame and our dependence on constant network connectivity is frightening. it’s like a heroin addiction, but i may argue.. worse; because the average ‘addict’ isn’t even aware of the addiction not even when withdrawals (boredom?) kick in. fortunately, a fix is never far so one may never realize a state of withdrawal neither. maybe not until a catastrophic infrastructure failure or some such would it become apparent. anyway.. apologies for the rant, but i had to do it. it’s an addiction thing. cheers!

[@KirillKovalevskiy]

I think, the elephant in the room is network speed and the fact that you never cache your data. Having you home directory on the cloud is nightmare. With the average internet connection the read/write access will be 100x slower than SSD.

[@rolandteague1778]

This looks awfully similar to AFS that I used back in the 90s at MIT.

[@BraulioCassule]

Slides made in Go

[@bawzzzz]

slides?

[@fennecbesixdouze1794]

So wait, if I'm not mistaken by what he said: anyone listed in the access file will be able to read the rest of the access file and see everyone else that I've shared the data with?

[@gl3nda96]

2:51 does he mean Google? 😂

[@MehulKumar_m3huL]

this is some real shit!

[@laexpearl]

Difference between this and IPFS ?

[@JosephHermens]

kzbin.info/www/bejne/e3-viHicnpWHockm45s

[@vison360]

Looks like worse variation of IPFS with central server and human readable identifiers.

[@yaxiongzhao6640]

Blaze used for global file access. Rob seems like to reapply old ideas in new places...

[@mishasawangwan6652]

Yaxiong Zhao though your comment is two years old i can’t help but remark: there are rarely ‘new’ ideas. i think it’s shortsighted to think there’s little benefit to recycling old ideas (a point i infer from your comment) and one need only look to history to see many examples of what can generally and simply be thought of as ..the fruits of ’iteration’. it’s how we landed on the moon, dawg. oh and.. 9,11 was an inside job. anyway ..just my worthless 2c. cheers!

[@bigdickmusic]

shorty! at 23:32

Sounds like a worse-is-better take on Urbit :) Instead of trying to square the circle by promising a decentralized Internet which would still somehow involve access to popular centralized services like Facebook or Twitter, just in a way that puts the user in power (good luck with that...), focus on doing one thing right that everyone might find useful: resource sharing and access from multiple machines. Of course, the lack of intellectual flirtation with the alt-right movement (unlike Urbit's creator Curtis Yarvin) is also a huge plus.

[@koredeaderele1666]

Got very triggered by 5:18 because I happened to be swiping through Tinder simultaneously 🙃