Re: LSR issues

[Sebastiano Vigna]

On Sun, 2005-02-13 at 23:02 +0100, Erik Sandberg wrote:

> I found 2 general issues about the LSR, for your todo:
> 
> - First, someone pointed out that it doesn't work properly with Opera. I 
> verified with W3C's validator (http://validator.w3.org/), that at least the 
> main page http://lsr.dsi.unimi.it/ is _not_ valid html.

Let's put things in perspective.

The mistake you point out is a missing couple of quotes that wasn't
there up to yesterday--my mistake. The entire site validates
flawlessly--I'm pretty maniac about validation and formal specifications
in general. For a comparison, lilypond.org's homepage has no DOCTYPE and
produces 12 errors.

Everything works perfectly with Opera: simply, the snippet database
works just with Opera 8+, because opera is a very flawed browser that
implements all important standards (starting with the DOM) in bits and
pieces, without any particular logic. It does not have a practical
ECMAScript console, moreover, which makes almost impossible to discover
workaround for the problems.

The Opera 8+ thing is new since yesterday, as, following a user request,
I downloaded Opera 8 and checked that they finally implemented a decent
DOM subset. So now it can run ERW, and I set my browser sniffing code so
that it lets Opera 8+ through.

> - We should decide about the licensing. My guess is that we want to make the 
> snippets public domain. And I don't think anything gets public domain unless 
> you specify that explicitly. You can get help on how to write public domain 
> statements at: http://creativecommons.org/

I thought about that, and I would have put up the question myself.

The problem with public domain is that we are going to put in snippets
from the manual, and the manual is not public domain. 

One solution is to state clearly in the entry form that whatever you
write because part of the manual, and that by entering your snippet
you're transferring copyright to "the authors" (that's what's written in
the Copyright statement of the manual: we probably need something more
explicit), which will distribute it under the GFDL.

I don't think anyone will be put down by such a statement, and it would
allow, say, to distribute an automatically generated snippet manual
without legal problems.

On the same line, I started to distribute all the sources of the
infrastructure under the GPL (even if they won't compile presently as
they need a custom MG4J).

-- 
Ciao,

                                        seba