Thoughts on Google's latest open source project: upspin?
>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.
Other urls found in this thread:
ipfs.io
github.com
github.com
twitter.com
fuck you renee french stop making creepy-ass mascots
fucking lmao
...
I make it a policy to not give a shit about "google's latest open source project" until it's already an industry standard. Because otherwise they're probably going to get abandoned in a month.
Sounds similar to what keybase is trying.
lemme guess, another apache/bsd licensed shit, written in a meme language to be integrated in the google botnet later.
DELET
just looked at it, also looks very interesting
Cute mascot to be honest
>Upspin is written in Go, so the first step is to install the Go tool chain. You will need Go version 1.8 or later.
Oh look, another episode of the botnet trying to take over every aspect of computing.
remember when google was making another OS? Fuchista? anyone?
>Uniformity and security are
But not privacy cause google.
it's completely open source
...
will you be smart enough coder to spot the google jew?
I bet Rob Pike is behind it. That kind of stuff seems right up his alley
yeah, this one looks like it's about to kill itself
>Written in Golang
I'm willing to bet anyone here is smart enough to read it clearly.
reading != understanding
>Written in Golang
I'm willing to bet anyone here is smart enough to understand it clearly.
So, this is IPFS made worse?
> CENTRALIZED keyserver, network fails to work without it.
> needs account
> not an "official" Google project
You should immediately start using this. Heh.
Kek
Chromium (not Chrome) is FOSS and it has been spotted delivering blobs and spying on people by default. Anything that Googles touches turns into shit because Google is the new MS.
If you guys want a true alternative to upshit, use IPFS ipfs.io
I can't wait for any of these distributed content addressed systems to take off, I don't care which, I've wanted distributed everything since I learned about Plan9 which was something like 20 years ago.
IPFS looks like it's actually on track to replace HTTP eventually (not anytime soon) for at least static content. Pairing it with Ethereum or even their own dynamic system makes me excited at the idea we might live long enough to see an actual distributed web. The fact that other projects like Zeronet and this upstart are popping up from time to time gives me hope that we'll reach this goal one way or another. It's been a long time coming.
Just hope it's actual FOSS and not some corporation coming to end it all with their greed to make money.
Also I believe you can already host your own blog with IPFS but I haven't actually tried it.
github.com
I know the craziest dynamic thing I saw on IPFS was a chat program
github.com
And following the development they seem to want to improve dynamic content in the future. Things are looking good and I am more than content with IPFS given that it's at least OSS. Not to mention all the other system I see like this are the same (by that I mean open as well). So the future in general doesn't look too bad, almost anything will be an improvement over our current centralized, name addressed content imo.
is it a watered down lang like python?
It would be hard for me to compare the two, I'd say that Go is more stripped(less syntax) than watered down(additional syntax sugar to do the same tasks).
Obviously you'll have to look at it yourself. I know both and I've had a much better time with Go than with Python, even for small tasks you'd normally use scripting for.
>watered-down lang like python
C cuck detected
>IPFS looks like it's actually on track to replace HTTP eventually (not anytime soon) for at least static content.
Doesn't feel like that will happen.
I mean, it's kinda somewhat capable to do this to some degree, but people won't generally pin enough of the internet and CDN and 'the cloud" will do the job with lower latencies and better bandwidth and more reliability.
It could be a way to share, dunno, wikipedia and archive.org, though.
Python isn't watered down?
Its something that kinda... grew... into something more complex.
I don't think the distributed caching is the biggest feature, imo it's the content addressing, routing, and decentralized nature that are worth adopting over HTTP, especially given that it's a zero effort procedure. Peer hosting is a nice bonus and great for individuals who want to host their own content but I think the big benefit for others comes from basically being a drop in, zero config cdn that can also be replicated and served by the users (keyword also).
The fact that it doesn't rely on dns (but can utalize it) is also a nice bonus.
Even doing a 1:1 migration from http to ipfs(convert your existing web servers and cdn nodes into ipfs nodes) you should still see some advantages without peer assistance, the peer assistance is just icing. The thing you mentioned with wikipedia too makes me think of schools, having students just talk to each other directly and exchange blocks they need over lan is a much nicer approach than setting up and maintaining a central squid server.
If you already have all these systems in place and the resources to maintain them you still seem to benefit, I feel like those who can't maintain those things will also be able to benefit which in turn means users benefit as well. No cdn no problem, no cache server no problem, in theory it should all scale up with activity but worst case scenario it's on par with HTTP in terms of availability but still has extra benefits like content addressing, etc.
Oh fuck I wrote too much.
So what's stopping illegal content from being distributed over something like this?
Lots of cool shit, sure.
But latencies, potential for attacks etc etc. aren't making this all that obvious, if you ask me.
Basically entirely every protocol and codec in widespread use has been sharing illegal content on a large scale, and virtually all have done so from very close to when their adoption started.
Some might say that the things that couldn't be used to spread illegal content were NOT adopted.
I mean, would you like a "copyright, slander safe" Sup Forums sitting atop a locked piece of hardware that monitors all you do?