On 23/07/2020 05:15, Bald Eagle wrote:
> 3.8 is in alpha.  End users _should_ know what that means, and 3.7 is still
> available for those that want to use it.  I and others have been using 3.8 for
> long enough that maybe it should be a beta?   I don't use M$ anymore - is there
> a 3.8 windows executable?

A beta can be made available, yes, but we need to work out what's in it
first as the features should not change between betas and final release.
See a little later on in this message regarding why we need to work this
out.

> If anything, it would look very nice on the website to have recent update, to
> show that there's still active development, and provide a handy link to GitHub
> or whatever the official channel would be...

Agreed. I will post a message to the front page.

> There's not - but 3.8 has a lot of features that Christoph worked his ass off to
> implement, and it would be great for people to use those, appreciate the
> differences and improvements, and find any little lingering problems that might
> be hidden in there.

Well here's the problem. Yes Christoph worked hard on it. But we simply
do not know if he considers the work complete or not. He just vanished.
We don't know why (and nor does he need to tell us; he's a volunteer and
can do as he pleases) but it does - for the first time in the project's
history - put us in a position where we're looking at doing a release
without the assistance of the person who made most of the changes going
into it.

I was concerned he may have been unwell as I noticed that his activity
on github (across all projects he contributes to) ceased at the same
time (I think roughly mid last year). A few of us have sent him hello's
via email but we've had no response. As of about two or so months ago I
see he has started to do stuff on github again so we know at least he's
there. Personally I wish him well and hope all is OK; I'm not going to
bother him since clearly if he wanted to communicate with us he would.

When discussing how to move forward on our dev mail list the point was
raised that we probably need to inspect some of his more recent changes
as we simply don't know if they were intended to be complete or a
work-in-progress. If in doing so we find or suspect the latter to be the
case then we need to work out what to do: try to finish it or remove it.

> I sincerely doubt that I have the skills or round tuits required to do that, but
> it would be helpful to anyone considering it to know what "managing the release"
> typically entails.

I'll answer that in a separate followup.

> AFAIK, it's in CT, and D.B. has physical access.

If by D.B. you mean Dik Balaska, no, he doesn't have physical access.
It's still in CT but was moved into a secured datacenter a number of
years ago.

> I've asked before how much is stored on it, and what it would take to simply

I don't recall seeing that (or if I answered it), but I think the
current backup is about 200GB in total (though the core content such as
databases and website content is much less - 200GB includes stuff like
log files, OS binaries and everything).

> _copy_ the content, lest a HDD crash wipe out nearly 25 years of images, scene
> files, includes, etc.

Fully understand your concern and it's one of my worries as well. New
system has RAID, the old one is a single spinning disk now (used to
be RAID but that controller died a long time ago). However the data is
well protected: there's an external SSD plugged into it that gets a
nightly backup via a cron job, plus I run a nightly rsync from my office
which keeps a full copy of the server contents here in Australia. If the
disk died the most that would be lost would be maybe a day's worth of data.

> There's a POV-Team, a dev team, a TAG team ...   what else?

Well the dev team is the same thing as the POV team basically. Maybe
once in the past when there were like 20 or 30 regular contributors we
may have made a distinction otherwise.

The TAG doesn't exist any more (they all moved on). All coordination is
done via a single mailing list.

> How many different computer systems are at play?

One.

>
http://news.povray.org/povray.off-topic/message/%3Cweb.5f0077dd17b7b05ffb0b41570%40news.povray.org%3E/#%3Cweb.5f0077dd17b7b05ffb0b41570%40news.povray.org%3E

Your comment about posting historical info is a good one and I'd like to
one day give a brain dump of what I can remember.

FWIW I also at some point will make the full revision history from the
time we started using perforce until we shifted to git public. I tried
migrating it to git (there is a tool for it) but it really didn't work
particularly well at the time (I think it was still in beta). Ideally
I'd like it all in git but it just may not be possible to do it cleanly
as there are some fundamental differences in the way the two systems
work. Worst case I'll put up a web interface direct to the perforce
server (they have a tool for that).

Having as much revision history available as possible is important for
long-term maintenance as it tells a story of what was done when and why.
I think I have changes going back to at least 1997 or something - I was
running a local perforce server to manage the windows code before the
team moved to it as a whole when I started managing the project.

(Prior to that we managed changes via email using diff and patch so
there's no real history to be gleaned prior to the above).

> POV-Ray is ... a company? At least for the purposes of being on paper?

Back when we changed to the AGPL3 we were assisted by the SFLC (Software
Freedom Law Center, which is part of the FSF). As part of this process
they advised (paraphrasing, I don't recall the exact words) that it
would be wise to have a legal entity to hold the intellectual property.
Hence the company.

Also FWIW having it in a single legal entity has other advantages in
that it makes it a bit simpler to transfer should that be required; e.g.
if I was incapacitated or passed away unexpectedly I have made it known
I wish the company to be passed to a suitable organization such as the
FSF (who can dissolve it and take over the IP if they wish, or just keep
it registered).

> Money gets spent - does money come in?  Is there a balance?

If you refer to Persistence of Vision Raytracer Pty. Ltd., then no and
no. It has never spent anything and has never received funds. Its only
outgoings are the annual accountant + corporate registration fees, which
I pay on its behalf.

> Are there corporate or academic interests that help keep POV-Ray afloat and in
> active development?

No.

> "There are some issues in play that most people would not expect."  That still
> seems to be a bit of a cliffhanger.

I suspect the main reason I said that was in relation to the situation
with Christoph. We wanted to have one last reach-out to him before
saying anything more.

> So, my puzzlement and curiosity are due to the fact that there is even a "behind
> the scenes" at all.  It seems odd to me that there is a lot of time and effort
> and money spent on a piece of software that is --- for what? --- if not the end
> users?  I'm not saying that in an entitled way - I'm raising it in all
> seriousness.

Fair question. The answer is that our 'end users' are more than those
who participate here or download POV-Ray for personal use. POV-Ray is
used by people who have code generated by programs or scripts (e.g. in
molecular modeling, academia in general). How many of these still remain
I don't know as they aren't required to tell us about it but my
understanding is it's not uncommon. These users don't deal with the
program in the same way you do - many aren't familiar with SDL and use
POV simply as part of a tool chain.

While I consider you guys (by which I mean anyone who uses it for fun)
as our primary user base I still need to care about other classes of
users. While you guys here may be able to adapt to a breaking change in
the way SDL works (either in syntax or in the way an image renders) in
many cases those who use POV from generated SDL won't have the knowledge
or means to fix it. They should be able to download the latest official
version from our website and just have it work as expected if this is at
all possible.

> all the personal cost and aggravation for something that is publicly available
> for free?  The sensible, though naive answer would be "so that people can use
> it".

Yes, it's so people can use it. ALL our users matter to me, including
those who simply use POV as a tool. This is why I'm so cautious about a
formal release that may break stuff. Even before I joined the project it
had a principle of preserving backwards compatibility where possible.
It's just how we did things.

Stability and compatibility is a promise I took on when I was elected
team leader and I'm keeping it. While I'm prepared to allow 4.0 (if it
ever happens) to break things entirely, any 3.x release needs to stay
stable - both in terms of output and in terms of not crashing when
someone does a multi-day render (memory leaks can be a bitch).

We had hoped the move to github along with the AGPL would encourage
people to fork POV and provide nice stuff for users who like to
experiment & have fun with and take pressure off the need to be seen to
have regular releases.

I know you're happy with the 3.8 alpha and it may be it's perfectly
suitable as a release, but there's a process that we need to go through
first which I will outline in another message.

> Perhaps there is not "secrecy" - but there seems to be an abundance of
> ....privacy?  ...silence? with regard to "Persistence of Vision Raytracer Pty.
> Ltd."  which gives rise to confusion, baseless but nevertheless natural

I think you're the first person who has ever asked me about the company
(or if not then it was a long time ago). There's nothing secret about it
or the team really. We're just a bunch of folks who (nowadays)
communicate via a single email list which typically gets little traffic.

> As I've stated before - POV-Ray could use some exposure, advertising, social media
presence, and fundraising.

About fundraising: maybe there's a place for it, maybe not. One thing I
will say is that back in the very early days of POV (before I even
joined) the issue of funds was discussed by the original developers and
a decision was made that the project would neither solicit or accept
cash donations to the project as then we'd need to work out how to split
it up, and money sometimes does funny things to people.

That's not to say we don't accept donations at all: for example over the
years several of the hard drives on the server have been donated by
users, and several of us have had programming-related books we wanted
given. But we have never solicited cash and at least for now I'll stick
to that.

If in the future we need funding for a developer then I don't personally
have a problem with someone paying them direct or some such arrangement
(e.g. crowdfunding) but it would not be officially run by us and would
likely be treated no differently than if someone were writing it as a
volunteer.

Long-term if we wanted or needed a more formal funding arrangement I'd
have to talk to the SFLC about it and probably do it through the FSF.

> the enthusiasm and energy and talent that is needed to make that 3.8 happen.
> And that's an observation, not an assignment of blame.

I don't disagree. We perhaps diverge on how to get to that point.

Honestly at this stage I think the best thing I personally could do to
spice things up is get Moray out, even if it's a binary only release. It
means diving into code I didn't write and don't understand but it's at
least a solid target, and who knows, I may even enjoy it.

But as far as 3.8 goes, yes it will happen but when I can't say as we
have to solve the manpower problem. See my followup regarding your
question about what a release manager needs to do (may be a day or two
before I post that, let's see how much free time I have later today).

-- Chris

Post a reply to this message