For Approval: Boost Software License - Version 1.0 - August 17th, 2003

[David Abrahams]

on Wed Oct 10 2007, Russ Nelson <nelson-AT-crynwr.com> wrote:

> David Abrahams writes:
>> Russ Nelson wrote:
>>> There was a period of time after switching to the new site when
>>> there were two different approval processes available on the
>>> website.  This is the correct one: http://opensource.org/approval
>>> and it includes directions to send a portion of the approval request
>>> to license-discuss.
>
>  > You mean there are other portions than what's described there?  What
>  > are the other portions and where are they described?
>
> I'm repeatedly totally amazed when computer scientists can't execute
> an algorithm.  Step #1 first, followed by the next step, and the
> next and the next.

Russ, I haven't tried to execute any algorithms yet; I'm just trying
to understand what went wrong in the first place, and to resolve
apparent inconsistencies between what you wrote in your email and the
page of instructions you pointed me at.  Your message indicated that
http://opensource.org/approval discusses sending "a portion" of the
approval request to license-discuss, which leaves open the clear
implication that there are other portions of the approval request that
need to be submitted differently.  That's a bit surprising because the
directions you cite don't seem to mention any other portions or other
submission routes.  The fact that you're ignoring the content of my
questions and responding with your amazement that I'm unable to follow
a step-by-step procedure only leaves me more puzzled.

> Notice the logical flaw in that sentence?  "repeatedly amazed"?  
> Yes, you screwed up by not following instructions.  However since I
> have adequate evidence at this point that the instructions CANNOT be
> followed, it is my bad that the instructions exist in the form they
> do.  My apologies.

OK, thanks, I think.

> Basically, we need to change something because at least half the
> license approval submissions are technically rejectable because they
> didn't follow the instructions.  I mean, Jon Rosenburg from Microsoft
> even asked me "Hey, are we doing this right, because we don't want egg
> on our faces" <------- not a direct quote but that's surely what he
> meant.
>
> So, we don't know what needs to be different, but something needs to
> be different.  

You could start by removing the statement at
www.opensource.org/docs/certification_mark.html that indicates the
process is started by posting to license-approval at opensource.org.

> It's not your fault for doing something wrong when most
> people do it wrongly.  It's just not.

Thanks, but all I did so far was ask my (very careful) lawyer to
follow the OSI's directions.

>  >  > Basically, yes, I saw the approval request, but because the
>  >  > public parts were never sent to license-discuss, I completely
>  >  > forgot about it.
>  >
>  > I get the impression from what you've written here that
>  > license-approval was supposed to be for non-public information
>  > relating to approval requests (what would that be?) and all the
>  > public information is supposed to have been sent to
>  > license-discuss, but AFAICT there's nothing on OSI's website that
>  > makes it clear.
>
> Oh dear.  It's happening again.  I sent you the URL
> (http://opensource.org/approval), I expected you to read and
> comprehend it, and you failed.

How did I fail?  I read the whole thing, and as far as I can tell I
followed it by posting Devin's original request to
license-discuss at opensource.org.  If you interpret people's questions
about apparently implied information in your statements (above, about
the public parts of the request) as a failure to follow directions,
it's no surprise people wonder if they're doing it right.

> If people repeatedly fail to comprehend something, it is the fault of
> the author, not the reader.
>
> But I don't know what to do differently.  The instructions seem pretty
> clearly written.  

I suggest you remove the reference to the license-approval list, and
answer peoples' questions when they ask for clarification about your
statements.

> Danese suggests that better software tools would help.

I don't think computer technology can fix this problem.

>  > One set of self-consistent instructions and a *single* submission
>  > address would be a good start.
>
> http://opensource.org/approval
>
> I tried having a single submission address earlier, and that didn't
> work.  Submissions got lost because they only went to me.  As it is,
> any emails sent to license-approval which don't have a "For Approval"
> in their subject are summarily discarded.  There's WAAAAAAYYYYYYYYYYYY
> too much spam coming to that address.

Well, clearly our original submission did have "For Approval:" in the
subject, but somehow it still got lost.

>  > You might also consider taking a cue from Boost's library approval
>  > process, which seems to work pretty well.  We have a couple people
>  > ("review wizards") who maintain a queue of incoming requests and
>  > schedule dates for open review of each submission, with a "review
>  > manager" appointed for each review.
>
> What software tools do they use?

We don't use anything fancy.  Requests get sent to the boost-devel
mailing list; the review wizards read it and (fairly) promptly deal
with all the requests that come in by assigning review managers and
dates.  For us, it's been a simple matter of clarifying directions and
establishing a straightforward procedure where no one person is solely
responsible at any point after the request is made.

-- 
Dave Abrahams
Boost Consulting
http://www.boost-consulting.com