From mboxrd@z Thu Jan  1 00:00:00 1970
From: =?utf-8?Q?Ludovic_Court=C3=A8s?= <ludo@gnu.org>
Subject: Re: Guix beyond 1.0: =?utf-8?Q?let=E2=80=99s?= have a roadmap!
Date: Mon, 01 Jul 2019 11:48:20 +0200
Message-ID: <87r27arspn.fsf@gnu.org>
References: <877e97vws8.fsf@gnu.org>
 <3C897E90-CDE7-4656-B431-F511C5FCC740@lepiller.eu>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Return-path: <guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org>
Received: from eggs.gnu.org ([2001:470:142:3::10]:53556)
 by lists.gnu.org with esmtp (Exim 4.86_2)
 (envelope-from <ludo@gnu.org>) id 1hhsvA-0008WI-U8
 for guix-devel@gnu.org; Mon, 01 Jul 2019 05:48:25 -0400
In-Reply-To: <3C897E90-CDE7-4656-B431-F511C5FCC740@lepiller.eu> (Julien
 Lepiller's message of "Fri, 28 Jun 2019 00:24:03 +0200")
List-Id: "Development of GNU Guix and the GNU System distribution."
 <guix-devel.gnu.org>
List-Unsubscribe: <https://lists.gnu.org/mailman/options/guix-devel>,
 <mailto:guix-devel-request@gnu.org?subject=unsubscribe>
List-Archive: <https://lists.gnu.org/archive/html/guix-devel>
List-Post: <mailto:guix-devel@gnu.org>
List-Help: <mailto:guix-devel-request@gnu.org?subject=help>
List-Subscribe: <https://lists.gnu.org/mailman/listinfo/guix-devel>,
 <mailto:guix-devel-request@gnu.org?subject=subscribe>
Errors-To: guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org
Sender: "Guix-devel" <guix-devel-bounces+gcggd-guix-devel=m.gmane.org@gnu.org>
To: Julien Lepiller <julien@lepiller.eu>
Cc: guix-devel@gnu.org

Julien Lepiller <julien@lepiller.eu> skribis:

> * accessible installer and installed system

+1

> * use the intensional model described in Eelco's thesis (should reduce ou=
r computing needs)

I think we should discuss what we expect from that model.

For example, =E2=80=9Ccontent-addressability=E2=80=9D might not be a goal i=
n itself: we
already have the content hash of all the store items, we already have
deduplication, and with IPFS we=E2=80=99d get the benefits of
content-addressability for data transfers=E2=80=94all without moving to the
intensional model.

Reducing computing needs might be the most compelling argument to me: in
some cases changing an input deep down in the graph wouldn=E2=80=99t entail=
 any
rebuilds.  We could change (guix build utils) without triggering any
rebuild, for instance.  It wouldn=E2=80=99t help in other cases such as sec=
urity
updates, though.

There are also practical issues: what to do with non-deterministic build
processes, how to migrate, etc.

> * channel browser (along with package and services)

Good idea!

Ludo=E2=80=99.