Re: [XBoard-devel] Installer

[h.g. muller]

> I think a link to a relevant page instead of a download.  But yes, 
> provided it comes with the disclaimer that the winboard project doesn't 
> endorse any particular add-ons or tools.

Indeed, a link would go to 
http://www.open-aurec.com/wbforum/viewtopic.php?f=19&t=49439 ,
(which would of course be updated to offer WinBoard 4.4.0 rather than 4.3.15)
and we could describe it in the section "Downloading XBoard / WinBoard" as:

"
Current policy of the XBoard project is to offer only files produced in the
project for download, and not bundle them with any third-party software.
Such downloads for XBoard can be found at <GNU download page>,
and for WinBoard at <GNU download page/winboard>
Installing the executable downloads from there will only provide you
with a winboard.exe and the associated help files, but will contain
a short tutorial manual for how to install engines and other essential
support programs.
More elaborate installs, that also contain some pre-installed demo
engines, protocol adapters, graphics or tournament manager are hosted
by third parties, e.g. at <WB forum downoad page>, but GNU does not
carry any responsibility for what they contain or how they are maintained.
"

> > I see it more as a mutually benificial arrangement. [...]
>
> Beneficial to some, a select few who for no particular reason got the boon 
> when others are equally deserving.  Now don't take this the wrong way, I'm 
> not actually accusing anyone of anything-- it's just hypothetical.   But 
> how do we know that nobody gets a kickback (monetary or otherwise) from 
> inclusion in a bundle or from causing the winboard project to plug some 
> specific software.  Far fetched?  Maybe, but weirder things have happened 
> in this world.  It's easily avoided too.  I think that's one reason among 
> many for sticking with freely distributable software or not bundling at all.

I agree that this is a danger. It would not worry me too much, though if 
someone got
a kickback out of it. Suppose for the sake of argument that Pulsar would go 
commercial,
but still allowed us to distribute this version for free. If it would 
benifit our users to supply
it, I still would see no reason to shun Pulsar. The aim is not to minimize 
other people's
income, but our maximize our own benifits. IMO it would only be 
reproachable if we
started including stuff that is really sub-optimal, because we open 
ourselves to bribary.

> If a 3rd party wants to deal with attaining relevant permissions, it's of 
> no concern to this project.
>
> > MicroSoft is criticised for not revealing the interface specs, (to make 
> > competition impossible)
> > and commercial anti-trust violations (pricing some products at a loss, 
> > funding them by profits on other, unrelated products, to bankrupt competition).
> > We don't do any of that.
>
> Microsoft is criticized for a lot of things.  But there are versions of 
> windows that don't include IE  because of bundling.  Not because of 
> interface specs.

I thought this fell under the commercial anti-trust: they supplied the 
browser for free,
(= no extra charge) funding it with proceeds from the OS.

> > It was not my idea to provide WinBoard and engines in one bundle.
> I'm sure it wasn't.  This whole issue never entered my mind until you 
> mentioned possible licensing issues hosting the install files on git.
> Since then, I've done lots of thinking about it.  I'm convinced that 
> bundling hurts some people, and could lead to licensing problems down the 
> road.  Permission can be revoked, unless it's in writing and labeled as 
> permanent.  And really, we should be able to do what we need to with files 
> in the project at any time.  Inability to do so because of permission 
> issues means to me that we are dealing with files that shouldn't be in a 
> GNU project at all.

Revoking of permissions would indeed be a serious problem. But there is a 
lot of third-party
software that can never suffer from this. For instance stuff released under 
a GPL or stuff
explicitly licensed as freeware. I took care to select mostly stuff like 
that: Polyglot, Fruit,
Elephant Eye, UCCI2WB are all GPL, Fairy-Max is open-source freeware, PSWBTM
is open source with a clear license that allows re-distribution. Joker, 
HaQiKi D and the
fonts are freeware.

Your latter conclusion is not obvious to me. We should be able to do what 
we need to
acheive the end result, but if there are several ways to accomplish that, 
some allowed,
but some forbidden, why not do things the way they are allowed? For me it 
is the result
that conts, not the method by which it is acheived. (Of course if it would 
be too
cumbersome, we could drop it because of practical considerations, but I 
would not
drop it out of principle.) E.g. putting the installer.exe in git seems a 
very workable method
to me. Even if it would contain stuff we have no other rights than to 
distribute them as part of
the installer, it would be just what we were doing. (And if we don't even 
have these rights, we
obviously have no reason to touch the stuff at all). People wanting to work 
on the installer
would simply pull the latest install.exe, have it self-extract in a 
temporary place, modifies
the files there, and then build a new one (from a RePackage folder with 
install script that
was pulled from git separately), and commits the new version. I don't see 
any problems
with that.

> Rather than bundle, I would prefer to see some solid howtos that are 
> maintained.  They could contain links and precise setup instructions-- 
> even scripts to set up software so the user wouldn't have to do anything 
> complex.  Of course with disclaimers that the winboard project doesn't 
> endorse any particular external software.  Along those lines, we might 
> consider making it easy for other software to determine the winboard 
> install location.

Yes, this would be a good way to do it. Installing over the network by 
directly pulling
the third-party stuff from the original source would be the ultimate in 
this respect,
and it is actually what I would prefer. But I don't know how to make a 
network installer.
Do you? Would you want to make the setup manual and / or scripts?

> Is that really your goal, dictating protocol?  If so, it really flies in 
> the face of the previous sentence.

It is not a goal, but it would sure make our work easier if efforts to 
extend the protocol would not
be frustrated by an enormous amount of inertia in the community. I would 
like to see a situation
where other GUI developers and engine authors are eagerly watching our 
website waiting for
protocol extensions that would allow them to implement what their customers 
want.

Mind you, I want open competition between GUIs and engines using WB 
protocol. I don't wat
people to design competing protocols and create incompatibity. That is in 
no ones interest!

> I'm not sure what the point is here.  None of that has anything to do with 
> the legal and ethical implications of the winboard project advocating 
> select external software.

We mut be miscommunicating, then. I thought you main objection on conflict 
of interest was
that the a lot of the "third-party" software was actually mine.

> I certainly applaud your efforts in the winboard project.  I probably 
> never made that clear.  I simply do not have the time to do the amount of 
> work you have on the project.  So please don't take my criticism here as a 
> jab at your obvious enthusiasm for the project, nor at the vast amount of 
> hard work you have contributed.
>
> Perhaps I'm just being pedantic about this issue.  Nobody else seems to 
> care-- at least I haven't seen anyone else who's concerned about it.

Well, I do think it would be better to leave such concerns to they higher 
echelons of GNU.
They must have policies for what you can put on their download page as 
bundles, and
we have done it before with 4.2.7 containing timestamp, timeseal and 
Crafty. Either we
were breaking the rules then, or the rules are more liberal than you 
imagine. I suppose
that GNU would not make these policies without involving people more 
qualified than we
in legal matters...