Gleanings from the GSoC Mentor Summit

[Jonathan "Duke" Leto]

Howdy,

Alvis: I bless you, as Community Manager, as our new Documentation Shepherd.

We take our docs seriously. They *are* just important as code. We
don't need your resume, you got the job :)

Don't be afraid to hurt our feelings. We know our docs are not very
accessible. Be brutally honest about what needs to change.

Duke

On Mon, Oct 24, 2011 at 1:03 PM, Alvis Yardley <ac.yardley at gmail.com> wrote:
> Hello:
>
> I'll offer to "take on" the documentation thing.  The "Constructive Whining"
> document I've been working on, essentially, boils down to one thing:
> Documentation, to include gathering up and setting a torch to (well ..., at
> least properly archiving) all of the incorrect and so very, very misleading
> documentation, floating out on the net right now.  (I've spent hours and
> hours and hours over these last several weeks, following down
> briar-and-bramble-filled trails because of out-of-date documentation, e.g.,
> /parrot/docs; WikiBooks; and
> http://trac.parrot.org/parrot/wiki/HLL%20Resources, to name just a few of
> the examples.)
>
> This said, I am very "new" (by which I mean old and returning after about
> almost nine years of absence from Parrot) and, consequently, may not be the
> best choice to take the lead on something of this magnitude.[*]  I'll leave
> this for others to decide.  Still, I seriously and sincerely want/need
> Parrot's documentation to "make sense."  Because, as of now, it is, to
> parrot (pun intended) cotto, "something of a mess."[**]
>
> Regards,
>
> Alvis
>
> P.S.   If it helps: I was the Editor-n-Chief of my law school's Law Review;
> clerked (meaning I substantially wrote several of the opinions) for a former
> State Supreme Court Justice; wrote and edited many, many legal briefs, legal
> opinions, yada, yada; published a few law review/law related articles; and
> published several op-eds in various smallish newspapers.  I have also
> edited, but not written, several technical manuals for, now, old legacy
> systems.  None of which is impressive.  I mean merely to point out one
> thing: I can write.
>
>        Note: I'm happy to give anyone a run down on my technical skills as
> well (if anyone is
>              actually interested).
>
> P.S.S. Duke, I am still working on the "Constructive Whining" document.
> Fortunately, I discovered something this weekend which "sheds" a great deal
> of light both on Parrot and on it's document set.  Unfortunately, however,
> this has required me to re-write, almost completely, what I had earlier
> written.  (I know these are rather opaque remarks, but, hopefully, they'll
> make sense when I complete the re-write and send it to you.)
>
> ---------------
> [*]   This said, if someone doesn't do it, Parrot may be hard-pressed to
> support the needs of its prospective HLL-developers, who, I believe, want
> their projects to perform superior to, but at least competitive with,
> projects developed on the JVM or the CLI.
>
> [**]  I would like to take this opportunity to make mention of one important
> point: As I am certain everyone on this list is aware, the success of
> getting the documentation set in order will require (1) specific direction
> "from the top," so-to-speak, and (2) a great deal of cooperation from the
> "core" developers.  What I'm trying to say is, in order for a documentation
> project to succeed, the documenting of the project must be seen, by all, as
> almost[***] as important as the code.
>
>       And, truthfully, it is.  If, for no other reason than, project
> membership changes, but the project (and the reasons for its coming into
> being, so-to-speak), hopefully, goes on, and new members are left with
> whatever code and documentation the former members left behind.  In short,
> it is not just a project-maintainability-type thing, but a
> project-sustainability-type thing.
>
> [***] Fwiw, I never have been one to subscribe to the notion that the
> documentation is "just as important as the code."  Why?  Two reasons: (1)
> The code is the deliverable and (2) in the end, you have the code.  And,
> although it may prove unreasonably and unnecessarily painful to do so, if
> one person wrote it -- given sufficient desire and time -- another can
> figure it out.  It's just that, often, the latter are lacking.)
>
>      These latter remarks are just my $0.02 and, perhaps, all they're worth.
> :-)
>
>
> On 10/24/2011 12:45 PM, Jonathan "Duke" Leto wrote:
>
> Howdy,
>
> I am at the Git Together 2011 right now, so I am time constrained.
>
> Documentation Shepherd:
> https://twitter.com/#!/parrotvm/status/128518846997999618
>
> To clarify, I have been fiddling around with Buildbot, but I don't care too
> much about whether we use Jenkins of Buildbot. But we absolutely, positively
> need to have distributed testing. This will solve *so many problems*.
>
> I talked to Alexander Graf from qemu and he showed me how I can
> generate binary rpms for Parrot for many different platforms via the
> build.opensuse.org build server. I will send another email with all
> the gory details.
>
> I talked with RTEMS people and they are still very excited about getting
> Parrot
> on RTEMS. They are very close to getting dlfunc support in RTEMS, which is
> huge, and will lay the foundation for dynamic languages on real-time
> systems.
> No concrete plans yet, but wheels are turning.
>
> +1 to setting up a community metrics website. Perhaps metrics.parrot.org ?
>
> Duke
>
> On Mon, Oct 24, 2011 at 1:20 AM, Christoph Otto <christoph at mksig.org> wrote:
>
> Howdy,
>
> This weekend, dukeleto++ and I had the privilege of attending the GSoC
> Mentor Summit.  The summit is a participant-run unconference for mentors
> of GSoC projects, by the hackers for the hackers, at Google's offices
> and on their dime.  Both of duke and I picked up a treasure trove of
> ideas and made some good connections with OSS hackers who we wouldn't
> usually get to hang out with.  I tried to take notes at all the sessions
> I attended and have boiled them down to a delicious list of action
> items.  There's some great stuff in the list that I hope will help make
> Parrot a better project and community, but it's considerably more than I
> can handle all by my lonesome.  Below is a summary list of the tasks I
> came up with.
>
> * Jenkins -  it's Java, but it's also awesome.  We need better CI and
>  many projects are using Jenkins to some very cool (and very lazy)
>  automated testing.  This would also help us make better use of the GCC
>  compile farm and OSUOSL's Supercell.
> * pdf.js apparently has a pretty cool bot.   Someone outght to see if
>  there's anything worth stealing and report back to parrot-dev.
> * dddash.ep.io is a spiffy and simple graphical display of developer
>  metrics.  This might be fun to put on a subdomain of parrot.org.
> * docs owner - our docs have always been something of a mess.  If
>  someone's interested, we could benefit from having a hacker who owns
>  our documentation and makes it his/her business to make our docs
>  shine.  This is a substantial commitment.
> * gsoc/gci analysis - We need to know how gsoc and gci are helping
>  Parrot, what our goals are and if we're meeting them.  We need
>  someone to lead this effort and to do some reporting about how gsoc
>  has gone in the past and where we can improve.  Note that Google
>  likes orgs that do this because it shows that we're concerned about
>  long-term results.
>
> I'll be going through them in more detail at #parrotsketch this Tuesday.
> If anything on the list looks interesting to any of you, please drop by
> and we'll see about putting you to work.  Alternately, just jump right
> in!  It's almost always better to ask for forgiveness than permission
> when hacking on Parrot.*  dukeleto++ has already started seriously
> looking at Jenkins and I've got a couple of blog posts to write and toys
> to experiment with.
>
> All this may be more than we can handle during one #ps, but I hope we
> can find people to work on a couple of these items.  Be thinking about
> what you'd like to do and I'll see you at #ps,
>
> Christoph
>
>
>
> *Except where it's not.
> _______________________________________________
> http://lists.parrot.org/mailman/listinfo/parrot-dev
>
>
>
>
>
> _______________________________________________
> http://lists.parrot.org/mailman/listinfo/parrot-dev
>
>



-- 
Jonathan "Duke" Leto <jonathan at leto.net>
Leto Labs LLC
209.691.DUKE // http://labs.leto.net
NOTE: Personal email is only checked twice a day at 10am/2pm PST,
please call/text for time-sensitive matters.