[Top][All Lists]

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

Re: Integration of new users

From: Anders Breindahl
Subject: Re: Integration of new users
Date: Wed, 10 Aug 2005 02:02:34 +0200
User-agent: KMail/1.7.2

On Wednesday 10 August 2005 00:51, Alfred M. Szmidt wrote:
> [I'm answering to the previous and current poster in this mail.]

Me, both times. I apologize for being so hasty on the follow-up.

> We cannot teach each and every person who asks for help, if you wish
> to hack, you must learn yourself the bits and pieces.  We are to few,
> and we have to little time for such things.  If we would teach each
> person who wishes to learn to program, then we wouldn't have time to
> hack on the Hurd; and that is why we are here.

I believe we -- those who are not (yet ;) ) hacking away at the kernel code -- 
assume that some other work can be done than hacking. That is, in example, to 
simply run the Hurd on a spare machine, or contributing to a wiki -- anything 
that can speed up the development process. I am more or less asking the 
tougher hackers to take action to relieve themselves from easier tasks, if 
such is -- by your estimate -- possible.

>    > The non-native-hacker users are left in the complete dark, as far
>    > as I can see it.
> We can't see it, but since you see it, why don't you help these users
> out of the dark instead of putting the burden on us who can't see it?
> You seem to know what is missing, why don't you write documentation?
> Why don't you update the user guide that is avaiable at
>  Why don't you answer to these posts?

I was, in fact, answering to one, when I realised that it wouldn't be an 
answer. Then I started this thread.

>    =2D The users of the premature GNU system are too much on their own
> I do not understand.

We (the GNUbies) are really willing to do more, after we've installed the 
system. The Debian page [0] one comes over during the installation of 
crosshurd, explains that:
"To start with Hurd development, you should install Debian GNU/Hurd and get 
used to it. Also, join the mailing lists and try to get a feeling for the 
state of the development. Offer your help, and we will tell you what is 
needed to do."

As this page refers to ``lists'' in plural, I assume those lists to be
bug-hurd, help-hurd, hurd-l4, gnu-system-discuss and debian-hurd
Having been on all of these for up to the past few months (I also posted my 
opinion on the naming of the GNU system to gnu-system-discuss), I pick up the 
pace and conclude that those offering their help generally get ignored.
Of course, you can't be held responsible for the text of the Debian site, but 
isn't it a pretty correct description of what you want? (And we only offer 
our help once, as we wouldn't want to repeat ourselves).

>    =2D There seems to be too little flow of information and help
>    downwards
> Downwards where?

From the ``upper'' developers to the ``lower'' users. I won't be after you for 
not helping us with our everyday problems, but helping us get started helping 
ourselves would not (as I see it; still your decision to make) be too large a 
burden, in comparison with its potential?

>    =2D No official documentation exist, nor any central entrypoints
> The offical documentation is distributed with each program, and the
> central entry point is the info program.

Now that information need to be conveyed to the new users. In example, I had 
no idea that the central entry point was `info`. It might be me who haven't 
looked the right places -- but I haven't been directed to the right places, 
so how would I know?
Some central and easily-maintainable site should be developed, IMO. Some 
canonical reference -- and it shouldn't be off-site.

> If any of you would like to make the GNU system more friendly for
> "non-native-hackers", please subscribe to,
> and say what you consider crappy, hard to understand etc.  But note,
> the GNU system isn't even usable for hardcore hackers at this stage,
> but it should soon be usuable (as soon as I get round to building
> everything, which should be soonish I hope).  But please read the
> mailing list before posting. :-)
> Hope this clarifies something.

It did. I appreciate you putting off time replying, but I urge you (the plural 
``you'') to throw some thoughts into how to use the willing users in the 
development process. I suggest an official wiki in the spirit of the one on [1], but with some developer-updated sections:
- What needs testing (when DHCP support becomes a reality, in example)
- What needs to be ported
- What troubles the devs now
- Where the devs keep their tarballs or packages, and what's new in them
- How the devs would suggest us to start porting a piece of software

Also, I opt to maintain some sort of newsletter like Debian Weekly News 
(altough I believe that weekly newsletters would be overshooting the target). 
Would that be worth the effort, and would any of the GNU insiders provide any 
necessary hosting?

Last but not least, let me stress that I really appreciate your efforts. I 
feel happy as a little boy when thinking of exchanging the monolithic kernel 
design with something more advanced -- and you're the guys who are working on 
giving exactly that to the world and me. I really appreciate it, and -- 
needless to say -- I'd hate to be in the way if you think I am.

Regards, Anders Breindahl.


Attachment: pgpvhP507dumt.pgp
Description: PGP signature

reply via email to

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