[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Future Direction of GNU Hurd?

From: Richard Braun
Subject: Re: Future Direction of GNU Hurd?
Date: Mon, 14 Sep 2020 23:04:54 +0200
User-agent: Mutt/1.10.1 (2018-07-13)

On Mon, Sep 14, 2020 at 12:41:42PM -0700, Jonathan S. Shapiro wrote:
> I'm reluctant to say this, because I know several of the people who have
> put a lot of their heart and soul into the Hurd over many, many years.
> The fundamental problem with the Hurd is the same as it has always been: it
> is a solution looking for a problem. Hurd advocates have not been able to
> clearly articulate what problem is being solved and why it is a problem
> that users should care about or be concerned about. This has been the state
> of the Hurd *for 30 years*. I am reluctant to say something so
> discouraging, but when a project has not moved forward substantially in 30
> years it is probably time to devote your very real and talented energy to
> something that will actually have an impact in the world.
> We went through this at Xanadu, and it definitely wasn't an easy thing to
> admit to ourselves or to act on. It helped, in our case, that we were able
> to sell off the intellectual property assets that we had developed.
> Nonetheless, the people who had put so much work and so much commitment
> into the project were both sad and angry to see it end. Some of the old
> guard went and formed a new, open source project to try to carry the work
> forward (which was a surprise, because Ted isn't exactly a fan of open
> source). Though the participants remain enthusiastic, that project has made
> no substantive progress since the 1980s.
> At one time, Xanadu and the Dynabook were referred to as "Platinum
> Vaporware." Both ended up having significant impact. The Dynabook by
> inspiring the current generation of tablet devices, and Xanadu through Tim
> Berners-Lee, whose decision to drip transclusion and charging/payment
> structures resulted in what we now know as the World Wide Web. I think it
> isn't clear yet what the Hurd alumni will build from their experience, but
> some of them have certainly done interesting things on other projects.
> I'm not trying to make any comment here about technical merit. I'm only
> trying to suggest that it may be time to step back, take a deep breath, and
> ask yourself what the best way is for you to have an impact on the world.
> If it is Hurd, great. If it is not, figure out what it is and go do that.

For me, the problem the Hurd attempts to solve is how to turn the file
system into a generic problem-solving tool. Unix started the trend with
its partially executed "everything is a file" idea. Its successors pushed
that idea even further. The Hurd makes it accessible to any user with
unprivileged translators. I have seen many situations professionally
where I had to develop new kinds of interfaces and re-implement standard
operations on those interfaces when I could just have plugged my
application into the file system like a translator, and have all the
power of the classic Unix text/file processing toolkit available, at
the very least during development and for debugging, and why not in
production for most non performance-critical operations.

Beyond that, I don't find your message particularly discouraging
considering the current state of things. It's honest and fair, and
perfectly conveys what I had in mind when I wrote that "I don't see
why the Hurd should absolutely "succeed". It's but one possibility
path among many others". Too many people tend to dismiss those paths
as "failures" but I tend to think they sometimes greatly contribute
to the "successful" projects, which are really just those which
turn out to get more used and continue to evolve, and not always for
the "right" reasons, very much like animal species.

I have learnt a lot contributing to the Hurd and it has broaden my
imagination and my mental skillset in general, in a way similar to
how e.g. functional programming ideas can greatly help imperative
or object-oriented codebases. It can at least still serve as a reference
on how to (or not to) do certain things, and for that reason alone
remains an interesting path to explore.

Richard Braun

reply via email to

[Prev in Thread] Current Thread [Next in Thread]