[Server-devel] let's write a compiler!

[David Woodhouse]

On Mon, 2007-07-09 at 18:24 -0400, Toby Knudsen wrote:
> On 7/3/07, David Woodhouse <dwmw2 at infradead.org> wrote:
> > Hm, the archive of this post at
> > http://lists.laptop.org/pipermail/server-devel/2007-June/000041.html
> > is broken -- most of the text is missing, and the In-Reply-To: header
> > specified in the mailto: link at the top of the page is broken.
> 
> Right.  Maybe your server made an error precisely once...

No, I don't think it happened just once -- I think the server makes the
Message-Id: error every time, and will probably truncate every mail
containing a line starting 'From ' too.

> > On Tue, 2007-06-12 at 14:06 -0400, Toby Knudsen wrote:
> When designing software other libraries may be required than those
> included in Redhat's builds. 

You mean Fedora builds, I believe. And yes, we already do add our own
packages to the Fedora package set where we need to.

>  It's a fine operating system, but many folks hooked on it never dare
> upgrade their libraries and don't know what's broken when they do. 

This is true of many folks who use _any_ coherent and self-consistent
distribution and then start trying to diverge from it. That's why such
divergence is generally a thing to be avoided except where it's
absolutely necessary.

>  I would say to anyone with a project of significant scope or
> requirements beyond those that quite general: manage your own software
> dependencies.

I think you're very much mistaken in that. In the userspace
distribution, we should strive to keep as close to our chosen upstream
as possible -- just as we do with the kernel. There is no point in
reinventing the wheel just for the sake of it, and forking a
distribution entirely of our own.

> I appreciate that you responded to my post.  

Well, the subject of your post doesn't exactly encourage anyone to read
it or reply to it. There are a lot of strange ideas out there, and
writing a new compiler is far from the most sensible of them. I'm not
entirely sure how that relates to the content of your email, in fact.

>  I've not been there for
> some time and (apparently) am not sufficiently resourceful enough to
> figure out how to work with you.  I'm worn out trying to figure out
> how to work on your school server and there's no design document.

Trying to 'figure out how to work on [the] school server' seems like a
rather quixotic task right now, since the school server doesn't exist
yet -- not even in a coherent design form. Be patient -- when we
actually _have_ a design, a 'design document' will be sure to follow.
What part, exactly, were you interested in?

> Many things seem to [happen?] verbally at OLPC and I haven't found that to
> be constructive previously. 

Yes; assuming that I interpreted you correctly, I'm inclined to agree.
That's one of the reasons I've spent so much time in Boston recently. 

>  By my account, I offered skilled labor at no cost and I can't see
> that OLPC is interested or focused enough to make use of my time. 

If you're offering labour on the server alone, then I think it would be
more accurate to say we're not _ready_ to make use of your time. If
you're willing to work on other issues, I'm sure your skills would be
more than welcome.

>  Just trying to figure that much out became tedious and I don't feel
> that I'm welcome at the office.  Do volunteers require an appointment,
> do we walk in?  

Walk in and do what, precisely? Distract someone from their task to
explain to you what our plans for the server are, despite the fact that
we don't actually _have_ a plan which is sufficiently coherent to be
explained yet? Do you think this is a productive use of time? How many
people should do this each day? Just you? 10 people? 100 people?

We're currently exploring the ways in which the server will act as a
'mesh portal' and provide IPv6 routing. Very little of the higher-level
stuff has even been discussed, as far as I'm aware. The most useful
thing you can do right now, if you're desperate for a server-related
task, is either test all the software on the laptop to make sure it
works correctly in an IPv6-only environment, or look at implementing
something akin to NAT-PT which will allow TCP connections to Legacy IP
addresses.

-- 
dwmw2