The SAFE Network Primer: An Introduction to the SAFE Network

Fantastic work on this doc! Many, many thanks to @JPL and @polpolrene for the hard work and other community members like @fergish and @frabrunelle for helping to fine tune - this is definitely going to be a valuable resource for getting people familiar and interested in the SAFE network. The contrast of the coloured copy and white background in the Keep It Simple and What Does That Mean sections did make readability a problem, on mobile anyway.

Be nice to have it up as a basic website and using github would enable it to be easily updated. Maybe using github pages and having pointing at it. Looks like it was put together in InDesign. Getting all the image assets and turning it into a static html site shouldn’t be a big job. More compelling CTA at the end to join the community forums maybe?

This is such a great resource and should be shared widely. The fact that it situates the project and tech within some political, social and economic context also makes it more accessible. Again thanks for the awesome efforts!


Just putting it out there… I used a file converter to turn the pdf into html and quickly get a web version up.
The resulting html is somewhat of a nightmare but as a simple link to share I think it serves the purpose.
None the less if anybody has time for doing a more thorough job Id close this down or transfer it whichever makes more sense. (With the blessings of the creators) also happy to leave it as is.

1 Like

Yes it’s great to have a web version but the PDF converter results in HTML that’s hard to edit. I can get hold of the InDesign files if anyone wants to create a website from scratch or stick it on GitHub. Point taken on the contrast. It looks fine on my screen but others have mentioned it needs to be beefed up. I’ll ask the designer to tweak it when we produce a new edition further down the line.


Completly off the topic of your document, (it is still downloading) but I love your use of “brought low.” I guess I am a word nerd.

1 Like

Thanks!!! Now I can read it at work tomorrow on my phone.

1 Like

@JPL can you give me the InDesign files, please. I made a translation into Bulgarian and I want to publish it. :slight_smile:


The SAFE Network Primer is now available on the SAFE Network, as it should be, with all the above correx incorporated. Thanks all for the input.


I had fun putting this together (and revisiting long lost and very rusty HTML/CSS ‘skills’ in the process) using @Shane’s fantastic SAFE-CMS. Seriously if you haven’t tried it yet give it a go. The first ‘official’ version will be launched on 11 March but it’s already pretty much there. It’s a super slick app and makes the whole process of creating a blog or safe:// website so much easier. :+1:


Very, very nice! And good looking markup too. :wink: Thanks!


this is the translation into Bulgarian of the “The SAFE Network Primer: An Introduction to the SAFE Network!” Thanks to @jpl and @polpolrene!ръководство-BG.pdf


Really pleased you’ve done this. Thanks a lot @dimitar :slightly_smiling_face:


Oookay - I know this is kind of embarrassing… but since I didn’t find the time sooner to read through the primer I had the pleasure to do it this weekend on the safe network itself! Thanks a lot for this! :heart_eyes:


this is the translation into Spanish of the “The SAFE Network Primer: An Introduction to the SAFE Network!” May I ask someone whose first language is Spanish to check it out?


I am confused by the following mentioned in the primer:
It says a chunk is encrypted using the hash of “another” part of the file, as in, not the hash of itself…
So how is opportunistic caching ever possible? Either it has to be 100% the exact same file, or different common “chunks” are going to need to be exploratorily calculated from what you can already find in the network.

For example, say a bunch of text configuration files all contain about 1mb of common stuff:
(different stuff)
(1mb common stuff)
(different stuff)

For there to be any opportunistic caching, the algorithm to chunk things up will have to align that 1mb chunk differently for different sources and see that chunk is already on the network to take advantage of it.

So any opportunistic caching at all sounds a little far fetched and intensive.

(I think I might be confusing the term opportunistic caching with something else, whatever reduces data redundancy by reusing shared chunks between files?)


Self encryption will compress and then encrypt this (it won’t be similar at all pre encrypt).

The de-duplication (I think you mean) is where the network see’s the chunk is already stored and does not need to store again. As chunks are named with the hash of their content then we can be sure (minus collisions) that the chunk is identical and securely so.

Opportunistic caching is where nodes keep a copy of any chunk they see in a first in first out (FIFO) memory pool. SO if a chunk is requested again, it may exist en-route to the real chunk location and be returned early.


Just to add to @dirvine and perhaps clearup what seems to be a misunderstanding

self-encryption will only produce the same chunk for “1Mb common stuff” if the previous stuff is exactly the same. Self encryption uses the previous stuff to encrypt following parts of the file.

So in order to have the exact same chunk then the file being self-encrypted has to be the exact same upto that data as another file that was stored. Caching is only using the actual chunk (still encrypted when cached) Thus only the hash of that (encrypted) chunk is needed to identify a match.


Have just taken my first cursory read of the SAFE Network Primer. Thank you to the authors for making it a bit less intimidating for this old newcomer. A couple of comments/ questions that may or may not belong in this topic:

  1. The link safe://safeprimer returns " No content found at requested address." in my SAFE browser .

  2. I did not see in the document any reference to performance expectations. When I searched the larger forum, I found this topic Network Performance and Vault Profitability which gets at the question to a degree. I found the comparisons there to the BitTorrent numbers particularly interesting. I guess my naive question is, when using SAFENet with applications for email, social or real-time applications (e.g voice / video calling) will latency / response time and throughput be as good as or better than what is experienced today on the centralized network?

  3. With the regard to anonymity, I am wondering what the mechanisms are to prevent bad actors from trolling, hate speech and criminal activity. I’ve come to this very much as an end-user looking for alternatives to the surveillance-capitalist approach in the mainstream internet but am not sure if the SAFE framework incorporates any kind of protection from malicious actors. Perhaps the kind of protection I am thinking about is not something that is done in the network level but is really one that is addressed at the application level, e.g. a chat or forum application would have the ability to implement what ever level of identity management and moderation methods that its developers deem necessary.

If these questions are way too basic for this forum/topic I would be happy to be redirected to a more appropriate forum or topic. In any case, I will continue to read and learn here in the forum. I’ve managed a couple of small steps on the client side - next up to explore vaults and farming?


Hi @djabril The primer was written with the aim of de-intimidising (if that’s a word) the network so I’m glad it helped a bit at least. It is a bit out of date now though and it’s on my to-do list to update it with all the new developments.

To your questions: the link is now safe://safenetworkprimer (see here for a more up to date list of sites).

Performance expectations is a tricky one as it is very much a work in progress and we didn’t want to speculate at this stage. There are various tradeoffs, eg when you break up a file and spread the chunks around the world it there will be an initial lag in getting those chunks, although after that opportunistic caching should help popular files download faster, but this is as yet untested. Client to client messaging should be very quick, but again there are no tests to use as evidence (as far as I know).

Preventing bad actors doing bad things is beyond the scope of the network itself, which is basically infrastructure, but there will be nothing to stop applications that build on top of that infrastructure from implementing their own rules as they see fit. That one’s an old perennial as you might imagine and there are lots of discussions about it elsewhere on this forum.


Hey, I have just visited the linked Transifex page and it showed the translation was 100% finished but no link – can I get a copy of the Czech Primer somewhere please? Big thanks in advance!

1 Like

I’d say both would be better. A PDF is better for storing on your computer, sharing via a thumbdrive or printing out. But an HTML would be better for sharing as a link you’re right. The problem with an HTML is if you change the size or download them they tend to break apart into multiple files and given this document has pictures with it that would be the html + a folder with the associated picture files. That means a higher chance of something getting lost if someone wants to download the book to share it offline. Not everyone has personal internet you know. And although SAFE is about creating a decentralized internet that doesn’t mean our only audience will be those who use the internet a lot or have regular internet access.

But as I said I do agree with you and think the Primer should be in multiple formats, perhaps HTML, ODT and PDF. That way it wouldn’t be that hard to share, adapt into things like pamphets or posters, and print out straight as a little booklet. Share, adapt and print.

1 Like

Here is a link to google doc. The translation is complete, but formatting is not yet. I also found some translation errors that didn’t make sense to me. I’m still correcting it.