The link to āFFI calling conventionsā in the Client Libs introduction pages redirects to the home page.
I agree; I think the libraries should separate concerns. It feels like RDF is another layer in the OSI stack, so that makes it sensible to be available as a separate module. I also think NFS should be a separate module. These kind of āstandardised conventionsā could also include DNS (PNS). I would imagine a lot more conventions will come up that might have no place in the same library as the core one.
From my experience that makes maintenance and testing easier, and also explaining the usage of libraries. Seeing it as a stack of dependencies makes it clear what relies on what. It eases documentation too.
Thanks all for the feedback - all very welcome! Like Sarah said, weāre not professional videographers or interviewers and this series is designed to be very informal and casual. But we do take your feedback on board. Weāre planning to film again this afternoon so iāll try keep from nodding too much
Will keep this in mind, we want this introduction to be approachable for everybody, but sometimes we developers may not even realize we are speaking jargon Are there any specific examples of things you didnāt understand?
We have yet to upload a lot of the pages, and there will be some broken links for the time being. The page you mentioned will go up in the next week or so.
The ideal form for introduction to anyone, would be to assume no acronyms or contractionsā¦ expand first instance of each. Even Client Libs could.should be noted is Client Libraries?.. if I knew nothing, would it all parse?.. so, first instance of Rust is Rust Programming Language, with a link out perhaps to what is Rust wiki somewhere nice and easy. If itās all soft and squishy then itāll be like quicksand and suck noobs inā¦ exciting them that they understand it all! If you want to overdo it, there are tools to measure reading age, which for the current perhaps would tempt moderate techy awareness requiredā¦ depends the audience how much you need to spell out. For me itās fairly good as is.
Yes, weāre also strongly in favour of this option. Even just from the sheer complexity standpoint. E.g., while querying encrypted data in Vaults is theoretically possible, it would require mind boggling tricks and considerable effort to achieve that.
Yeah, it definitely is, in a sense that you would surely be able to use primitive data types (ID/MD/AD or whatever) and store arbitrary binary data on the network. But Iād say RDF should be strongly favoured for structuring apps data, albeit this approach might seem opinionated. Itās important to make RDF/Linked Data as ubiquitous as HTML though, because implied possibilities are huge: S in SAFE would stand not only for Secure but also for Semantic. Thatās why weāre considering basing some of the core Client Libs abstractions (like NFS files/containers) on RDF too.
excited to read the updated PARSEC paper (although wonāt understand it ).
Good call to go peer-review IMO, this is cutting-edge computer science after all.
Are not you putting all the eggs in the same basket by doing that, taking into account that quic protocol can be āeasily blockedā by ISPās / governments?
It shouldnāt be better add quic to the network protocol stack and integrate it in dynamic connections implementation as a one more protocol available?
Or you could drink some wiskey at the distillery at the foot of the mountain before making a video, like I did a couple of years ago (the drinking part) Btw the people at the distillery also donāt like to make videos, it seems. They showed me this dated intro video, older than any wiskey you could buy there.
Thanks to the team for keeping the fires burning. Iām certain the videoās will get better as time goes on - it took me nearly a year of making one every week to get my audio decent! [ embarrassed ]. I appreciate the informal approach of the interviews and discussion, so IMO and as others have pointed out, itās just going to be a process of developing the A/V techniques to make great media and being careful to dumb it down so that the grandparents can understand at least the general ideas of it.
Just like any brand new radio show, TV series, even for professional broadcasters and actors it takes a little time to āwear in your shoesā. Iām sure we have all noticed it.
That embarrassed feeling, just natural nerves - just stage fright. All listeners and viewers are looking out for is that you are having a good time up on stage.
I look at my early android tutorial videos and I see myself trying to do what I thought a video tutorial should sound like - acting. Nah - just stick up a camera, and talk as yourself, with background noise if you have it. A lot of popular YouTube stars do just that.
I still remember your excellent intro to SAFE video. It was top class, and essentially just you talking while showing in a clear straightforward way.
So I think your advice is good and your own output backs that up. And I hope youāll do some more SAFE videos yourself one day!
I hate doing that stuff, but will have a go again at some point if I have to (hint). For my DevCon presentation (which was a pre recorded screencast) I did several retakes which makes it time consuming for me - that was partly to get within the time limit, and because I had to record the audio separately though.
Soon Iāll have something that will make a great screencast demo I think - short and simple, but powerful.
Iām waiting on a couple of changes to SAFE Browser and some bug fixing in my code, but I think it will impress and reach people we are keen to have join us at this stage.
So @Zoki, if you or anyone else fancies playing with this and making a screencast for it PM me for details.
The next HTTP version, HTTP/3, will be based on QUIC. So when it goes mainstream, SAFEnet traffic will seam undistinguishable from HTTP. If ISPs were to block QUIC, they would be blocking HTTP as well, which probably is very unlikely.