Cool stuff! Although I don’t think I’ll be using www on Safenet. Way to confusing for a lot of people. safe:polpolrene should be fine Will there be a reset before launch, where all the DNS names are removed again? Or is there a possibility that these names will be alive in the 1.- network?
Yep, it should be UTF-8 compatible
I think these sort of UI improvements can certainly be made in future iterations of the plugin. Worth noting that it’s not just for an actual url entry into the browser directly by the user that the plugin can react to, it can also respond to a url request to fetch data from say an image link in another page on the traditional internet. As long as the browser backend receives a request to fetch data from a url that has the protocol
safe:, this plugin should be able to intercept and provide that data.
You dont have to login to the network to view public sites/ retrieve public data from the network.
For now it’s very much just via MaidSafe, but once it’s deemed stable and feature complete it should make it’s way to the Mozilla hosts.
There should be. Prolly worth considering this for now as an elaborate POC.
The main goal in this week with the Plugin/Desktop App is to show App-Devs how they can start building/thinking-about apps for the SAFE network and give them a hint of the different port options we are hoping to be providing in the coming weeks(Do note launcher isnt still added to this process, it should be in the next sprint hopefully ;)). Just thought this was a nice example where we had user interaction from both traditional desktop apps and also via a browser plugin
Everything is utf8 compliant. So whatever’s within ambit of utf8 correctness, should work
Not sure, but is this just related to parsing of the url you mean? In that case, it’s just the url parser that needs to deal with it.
… NSA … Just by the amount of money spent on defense and intelligence, i would assume they are a pretty capable team. Nah but i don’t know the answer to this - i’ll leave it to others probably
My RFC pull request gives some flavour of the kind of thing I have in mind, and why it would be useful. To progress in this area we need to decide what approach to take and of course who will do the plugin work. I think it would be best to have one standard plugin and client side API, so part of the “official” plugin. (Details here: SAFE browser plugin URL handling / dynamic HTML on client side -> click “Files changed” tab to read).
Like @polpolrene I’m not sure of the value of “www” in the DNS. Can you elaborate on what you mean by services in this context - what services do you anticipate and why are they needed? They are largely redundant on the current web as far as I understand, and just serve to confuse users, search engines and website control panels!
And I want to jump up and down too This is great work everyone. It must be good to feel things are getting so close to putting your ideas and apps out there on a network that anyone can install.
Also, within any directory listings, there will need to be a clear indication of the access allowed on files. Will a user be conscious of what they are sharing?.. Perhaps use coloured indicators on file listings [Green:Amber:Red] for [Private::PrivateShared::Public]. I wonder Windows users etc might not be familiar with file permissions and that aspect is critical.
I’m expecting necessarily that whatever a plugin can do, will be available outside of a browser too.
Off topic thought#1:
People widely are not very original and will rarely understand the need for obscure choices of account detail, unless prompted. If there are a large number of people over a long period, choosing their own account name:password:pin, there is a high chance of then clashing with an existing account or of a bot brute forcing those accounts. I’ve seen there is an ID provided, so perhaps that usurps the account name now?.. otherwise something like a generated 12 x random word account name, is easy to understand… and perhaps more user friendly than a string of alphanumeric?.. perhaps allow brainwallet like credentials.
Off topic thought#2:
If most blocks are 1MB, will transfer of those be an indicator of network usage? Would network mostly just work, if an arbitrary 1MB +/-10% was used for any one file… or is 1MB necessarily hard in the code?
No MaidSafe keeps the tradition of being brutally decentralised high
So this first test release will be very flawed, but there is no centralised (in-house) data storage; it will be your own hard drives that store and retrieve the data. So from these tests we will need to learn “how to learn from tests run in the wild”, but there is no mock-routing or “controlled” part in this very early release. It will be flawed, but not faked
This is for app devs to recognise global service names like the /etc/services in linux, so you know port 80 is www *usually) and 25 is telnet etc. So these services will build so you know if you retrieve a www service then parse as html pages, blog as a blog type connector etc. It’s an idea to allow globally agreed services really, if that helps. So you can have
and apps will know to treat these as such. Some may be built in to the network as fixed, but like service not fixed in any way they cannot be ignored or broken by clients parsing www as email fro instance, but fixed as in feels globally recognise.
It will ti in to hopefully some standards across apps to allow interoperability I hope.
Oh we definitely aren’t planning on stopping with just url interception. we just are for this example right now thats getting worked on this week. Current focus is to first setup the ports and get the plugin/app “functional” with a minimal scope and once thats done, extending it should be simple enough(hopefully )
For example, if we look closer at the screenshot @brian_s posted for the firefox plugin logs, you can see log messages indicating directories/files getting created. We only need “Gets” from the browser this week for our minimal requirement, but to confirm the process the guys also exposed the put functions. So expanding this scope to expose the full API to the browser environment shouldn’t be a hard thing
Hey that’s great! There wasn’t as much clarity on this subject so thank you for that. I was trying make my best assumption with the posts I had found on the topic.
@Viv: I’m sure you know that already, but just to be sure. With firefox Version 42, mozilla only allows signed extensions. Although they just released v40, my guess is that the release of v42 and the SAFE extension probably happen on the same day
The very fact that we are talking about what is arguably a parallel Internet is just extraordinary. This occurs to me on just about every comment or thread, it’s always in the back of my mind.
I remember years ago, reading what was a hillarious comment in a subreddit (probably R/technology), “I found that out on the internets”. I’ve chuckled about this perspective ever since. Of course, with a parallel Internet being developed a la MaidSAFE, the concept of the Internets is no longer funny, but a vital form of online existence and communication.
As we all now see, holding on to 1 Internet that corrupt organizations control is a recipe for disaster already in the works. I wanted to end this comment on a positive note, so I’m typing this on the internets!!
Remember the “the internet is a bunch of tubes” Even our politicians in AU said this, not just a US politician.
Just download the plugin and start accessing safe sites? That is brilliant! A great , simple way to start experiencing safe net .
Perhaps www could be implied if no service is specified ?
Yes www is default for now.
So, I could have a safe site called safe:mysite and it would automatically load mysite’s html page?
Yes, it would But for this sprint, am not sure whether it would be available (Just because of API restrictions as of now).
No worries, I’m thinking long term
What? I still get:
/home/phil/.cargo/registry/src/github.com-0a35038f75765ae4/sodiumoxide-0.0.6/src/crypto/stream/chacha20.rs:3:11: 3:37 help: run
rustc --explain E0432 to see a detailed explanation
error: aborting due to 2 previous errors
Could not compile
To learn more, run the command again with --verbose.