Am 07.03.21 um 14:16 schrieb yuv:
On Sun, 2021-03-07 at 09:22 +0000, Bruno Postle wrote:
currently there is a single person doing code maintenance,
collecting translations, _and_ doing the releases - these could
easily be different roles.
This! What made Hugin great and successful in years past was the co-
operation and co-ordination of the different teams / authors to the
point of trusting and sharing full access to their respective code
repositories; being able to inter-changeably assume the different roles
for the different packages; reaching mutual agreement on repositories
and other tools to use to simplify interchangeability of roles; co-
ordinate releases, bug fixes, etc.
I felt it was difficult to keep my foot in the door. Sort of
disheartened. Like, I got stern admonishments rather than friendly
encouragement. I prefer to run 'my own show' now, where I don't step on
anyone's toes.
That co-ordination seems to have gone missing, and re-building it will
require sacrifices and compromises on all ends.
I'm not sure if there is the will to rebuild stuff, much less sacrifice
anything. I have tried to help keep my python interface afloat, but
apart from that really I wouldn't want to touch any of hugin's code
anymore. I'd even let that go and not be too sad about it, if someone
came up with a better free stitcher, but I'm used to it and it does it's
job, more or less. And I can tweak the sources locally to bend them to
my will. And fuse my stacks separately, when hugin insists that my
assignment of stacks is irrelevant to it's heuristic approach. When it
comes to python, what I now use is cppyy
(https://cppyy.readthedocs.io/en/latest/index.html), that's much cooler
than swig.
I have just taken a half-hour stroll down memory lane. The main
comment I have: Why have no admins been added to Hugin / Panotools /
Enblend in ages? My memory fades, but all the names I see are people
that were either there before me, or were added on my initiative. Who
has taken care of the team after I moved on because life?
* have any new contributors been invited?
* if yes, why have they not accepted the invitation?
* if no, why?
* how many other contributors has the ecosystem missed? not just admins
(the highest level of access/trust).
Look at the hugin website: 'hugin is now stable'. I suppose that's the
idea. Yes, there have been great innovations in the past few years - the
control point tool, entering names of raw files to be converted to TIFF
by external helper programs, a dynamic range compression button, and I
discovered one can even use the mouse wheel to change the hfov in the
openGL preview! And that's only mentioning the stuff that immediately
springs to my mind!
A couple of weeks ago I even managed to unlock the positions of images
in a stack! It was easy! I only had to find the right submenu and click
on a bit of empty white space to get the dialog offering me to do so.
Only took me half an hour to figure it out, with the help of looking
through a few postings on hugin-ptx, which I would quote now if I could
find them again with the search tool, I suppose I need other keywords
than 'stack' and 'unlock'.
Below follow a longish (and incomplete) list of signs of what is in my
view organizational rot. Those signs point to extra work. I am not
asking the exisisting and dwindling team to take on that extra work. I
am saying that the organizational rot is the consequences of the
failure to welcome and embrace new contributors. Who wants to join a
dwindling team who does not welcome change?
Good question.
adding team members requires two thrusts:
(a) individuals inspired to become team members, to bring new energy,
new ideas, new code, and eventually change to the projects;
(b) an existing team that is welcoming new contributors and accepts the
change that they bring.
where are those thrusts?
And now for the longish list.
* why are the projects still using Mercurial? IIRC back when the
switch to a unified VCS was made, Mercurial won because it was the one
with most even support amongst operating system (Windows!). Meanwhile,
the landscape has shifted, git is the tool. That would require some
flexibility from all individuals, including Kay (bitbucket? exotic! I
would not adopt it at this time).
I've been on bitbucket with all my projects for many years - for
upstream git repos, online presence and issue tracking bitbucket is just
fine, and I like that they are independent. They've been good to me,
nice and helpful, and I won't turn my back on them just because they are
'exotic'.
Downstream - that would be hugin, when it comes to pv - can be anything
doing git. The debian science team, where I maintain vspline, has moved
to gitlab, and I was happy about it, the old solution was a bit awkward.
gitlab is clean and capable, I think it's a good choice. But it's
definitely git now wherever it's hosted, I agree.
* why are the projects still on Sourceforge? History is past, and they
may have corrected they blunders. But there is no going back and today
Gitlab is the place to be (which would also replace Launchpad, that I
introduced ages ago).
I never liked sourceforge, and I did not create an account back when I
contributed to hugin, because I did not want to sign the terms of
service. I wanted nothing to do with it, that's why I sent in patches
instead. I'd still like to see hugin move.
* it pains me to see F*book promoted on the Hugin's website. Really?
it surely explain less traffic to this mailing list, which means less
potential contributors coming its way. Don't expect an
influencing/advertising tool that has an interest to isolate users into
echo chambers to be a positive to this community. if I was around the
time that decision came up for discussion, I would have vehemently
opposed it.
Gosh, I did not even see that. Yeah, posting to hugin-ptx has not been
rewarding for me recently. I started several attempts to get *anyone*
interested in my program. I did not get very far - a couple of
enterprising linuxers who built it and never said much more about it. Do
they use it? Do they like it? I have no idea. Some thanked me and even
gave a bit of praise.
When I offered ready-made Windows binary packages 'for evaluation'
recently, I had *two* takers, who never wrote back when I sent them the
download links, either to me or to the list. I had something else in
mind with 'evaluation'. And with a couple of linux people building pv, I
had thought I'd get at least ten Windows users. Not so. Funny that now
all of the sudden when I announce I can even stitch and fuse with pv,
there is some echo. Hardly anyone seemed to be interested in 'pv, the
multiplatform FOSS panorama and image viewer'. Maybe fb has an
integrated viewer.
* the Enblend/Enfuse website is still... mine! What is the new
generation of contributors waiting to leave their mark? Last news from
five years ago? OK, the tool has matured and no need for new releases,
but someone with a little bit of marketing flair could point out to the
continued flow of Hugin releases, and now to Multiblend, a very
interesting development?
I don't think people at enfuse/enblend are interested anymore. I still
saw the same flaws (primary-coloured sparkle which I assume is a failure
to do saturation arithmetics) after years, so now I'm happy I have my
own variant of the algorithm. I just haven't managed seam optimization -
yet.
* the Panorama Tools website list the last new to almost eight years
ago! the latest release tarball is two years ago.
Panorama tools... That was not moving even years ago when I was
contributing to hugin. It's *stable*.
I could probably find a bunch more, but the point is: encourage new
contributors to take over and replace the current generation and take
this community to its future with as little interference as possible
from us oldies. Else, the community will die when we inevitably do.
So please, please, please, make this a more welcoming place to new
contributions. Kay and Harrie are no newbies, so imagine how difficult
this is for someone who has not been around for that long. The
alternative is to fade into irrelevance, which is why I asked the
thought-provoking question of whether pv is already a replacement to
Hugin.
Since you're playing devil's advocate (hah) - I can wrap all panotools
transformations in vspline, most of them are even easy to port to proper
vector code. I could throw in a few calls to helper programs like
cpfind, celeste, PTOptimzer - and that would yield a 'core' program
which can do the most important bits, one could take it from there. But,
as I said, hugin is good for image registration, and as Harry and Bruno
said, it can still do quite a few things which pv doesn't, and wasn't
meant to do. I still favor two separate programs, but I'd be happy to
see hugin evolve. Maybe some of the tricks pv can do can serve as an
inspiration. And, with pv, I like to 'move fast and break things'. Come
for the ride!
Kay
--
A list of frequently asked questions is available at:
http://wiki.panotools.org/Hugin_FAQ
---
You received this message because you are subscribed to the Google Groups "hugin and other free panoramic software" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to hugin-ptx+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/hugin-ptx/1cd2ad63-550c-c86e-f05a-f58147e21ae6%40yahoo.com.