Is there a safe search engine?

hi I downloaded the safenet browser and are now connected to the save network, if I want to search some webpages now with the browser is there a save search engine ?

5 Likes

As far as I know there isn’t a search engine yet (although already discussed on the forum).
But you have a list of sites in this topic: Outdated: Alpha 2 Community Websites, Apps & SAFE Messaging IDs.
There is also safe://listy.

5 Likes

Also, SAFE + Solid = significant implications for the future of applications and “search” by the looks of things @texxgoro . :sunglasses:

7 Likes

Will a search engine for the Safe Network become available anytime soon? Is there a plan to create a search engine?

1 Like

Hi, @SantMat. Once there is sufficient public data on the network then there will be demand for a search engine, and that demand will be met with a solution.

3 Likes

Hi, everybody.
I have to add -
How is It planning implementation of search engine if all data are encripted in the SAFE net?

Lots of data will be published too (all websites). So even though all the chunks are encrypted, files that are published can still be read, so those can be searched.

1 Like

What do You meen “to be published”?

Data that’s on a SAFE website is public and can be read by anyone with the right software (eg the SAFE browser). That data can potentially be indexed and searched, even though its encrypted. Private data is not searchable however.

1 Like

he says the files can be personal or public. Public files are available to everyone

1 Like

Indexable though i think … You can’t probably make out what it is but you can certainly flag a location as “interesting”. Which is why i would guess it’s of paramount importance that users don’t use passwords (unless they are sure what they are doing) to hide stuff as that stuff’s going to linger around forever (for analysis and what-not) esp. if all data in the SAFE Network is made immutable (so either ImmutData or Mutable/Append-onlyData which is append-only but content already put in is not mutable) - like an Internet Archive thingy. Much safer to let the libs (safe-client-libs/safe-crypto/self-encrypt) do that for you using strong random cryptographic keys and only password being your credentials to log into the Network to subsequently fetch all those keys.

6 Likes

I’ll just chime in with my constant revisiting (however short it is every time) of the storage structure question.

What do we need to create a scalable searchable structure?

One first step would be a simple pattern to get a limitless key-value store (I.e. more than ~1k entries of an MD).

I am not fully happy with my persisted dictionary implementation. But it’s a move in a direction. It’s nice when it comes to size, not exactly limitless, but it can be made big enough to suit the needs of most mortals. What I am less happy about is:

A) Since it is based on creating multiple levels of indirection (MD entry => MD address), where each level raises the db capacity to the power of 3, placing values always starts at the very furthest leaves (well, there will not be any closer leaves once you set the size - i.e. number of levels).
B) Multithreaded access as to scan the structure (which you don’t want to do with any non-trivial sized dB), trips the client. Could be the c# client, could be deeper - I don’t know. (Symptoms are very similar to a previous bug that I found though).
C) It’s not written in a low level language.

Something like it could be used to store indices of the data stored in the same kind of structure. We could also try relational dB over key value store á la CockroachDb (even though bigger problems await around that corner).

With that said I have given it a rudimentary try - with my very limited time - (and will continue with whatever small fragments of time that I have) but I hope others are thinking about concrete implementations, that can leverage the MDs for scalable storage that is also efficiently queried. It is needed for any app beyond personal organisation of files - and most certainly for searching.

5 Likes