Re: Temporal Stack

On Sun, 21 Mar 2021 at 13:37, Timothy Holborn <timothy.holborn@gmail.com>
wrote:

> had a think.  thought i'd post it.
>
> IMO there's cause to build into WWW / HTTP a method to support temporal
> lookups, other than simply using archive.org.   i imagine this would
> eventually require ICANN, IETF (etc) support; amongst other implications.
>
> The functional outcome would be an ability to look up a page at a
> particular date.   This may involve differences in who owned the domain
> name at that time (vs. who may own it later on), amongst many other
> implications.  There would have to be a 'format' of 'standards' around how
> to achieve it, for long-term support.
>
> Foundational requirements, prior to more easily engaging CMS providers
> such as Wordpress / automattic, drupal, etc.  would be to define a simple
> concept that could be built upon to do it.  I imagine it may take some
> years to do, and i'm not entirely sure i'm up for it - historically no
> funding for work by civics persons (civilians, working independent of
> contract / employment revenue) for doing W3C works; maybe, with new changes
> that might be reviewed; but regardless,
>
> cost of storage, etc. has been dropping.  I'm not sure what the economic
> model for it would be, but i can think of a variety of ways a solution that
> attends to the economic implications could be forged.  I also think, an
> evaluation may lead to an outcome where it's able to be understood how to
> do it at a lower energy cost than simply employing DHTs / Blockchains
> ("DLTs"), although the file-system layer may be considered independently,
> atm, idk; and don't really want to make the point any more complicated than
> it needs to be for now.
>

A lot depends on the use case

The web as a distributed document system isnt really designed to do time.
Most people I know that have used web standards to work with time have run
into varying degrees of complexity issues

You can rely on a trusted third party like archive.org or try and create a
system where you can ask for older copies of a page, which handles some use
cases

A block chain, which was originally called a time chain, is designed to
store small amounts of data as a time series, and if distributed provides
resilience through replication.  This is generally considered strong enough
for financial grade transactions without relying on a trusted third party.
However many such chains are really a "please buy my token" scheme, pitched
to an unsophisticated user/investor base.

In most cases archive.org or something similar would handle the need to
time travel through many documents.  Time traveling through a financial
ledger (to prove there was no fraud) generally requires stronger
assurances, which is where the security of a block chain comes in

The ability to use a time / block chain to provide assurances about the web
over time, could have some use cases, when the proof of non tampering of
the data is important, or when you might want to reconstruct the history.
Or if one site goes away you might want a permanent record that could be
taken forward by another party

What would be needed for standardization between a DLT and the web would be
a bridge layer between the identifiers on the web (in the strongest sense
URIs), and cryptographic primitives

So it would just be like the regular history, but more trustable.  It would
be helpful to have a use case for this, if you want standardization or
someone to build it


>
> Timothy Holborn
>
>