A wholehearted thanks to everyone who has donated to libjpeg-turbo in
the past few days. Because of your generous monthly and yearly
sponsorships through GitHub, we will have a 30% higher monthly budget
going forward. Because of your generous one-time donations through
GitHub and PayPal, in addition to the new monthly and yearly
sponsorships, the General Fund for 2023 is now over 60% larger.
Combined with an expected payment from another source, we will now catch
up to the deficit in about four months instead of 15!
Thank you thank you thank you!!!
You can see the current list of public GitHub Sponsors (those who chose
not to be anonymous) at the bottom of this page:
https://github.com/sponsors/libjpeg-turbo
All public sponsors also have a libjpeg-turbo sponsor badge on their
GitHub profiles.
DRC
On 7/3/23 5:56 PM, DRC wrote:
Binaries and source tarball are here:
https://sourceforge.net/projects/libjpeg-turbo/files/3.0.0/
Change log is here:
https://github.com/libjpeg-turbo/libjpeg-turbo/releases/tag/3.0.0
Sorry for the delay, the reasons for which were expressed in previous
emails
(https://groups.google.com/g/libjpeg-turbo-announce/c/zXCK0qw1pIE/m/yMLo__OuAAAJ).
SignPath is still having issues with their upstream CA, so the
Windows installers for libjpeg-turbo 3.0.0 are temporarily unsigned
until those issues can be resolved. Ultimately, however, the delay
produced a more stable 3.0.0 release, since most of the bug reports
against 3.0 beta were filed after the beta cycle was supposed to have
ended.
On that subject, I want to remind everyone that, while I do my best to
maintain this project with the highest possible quality standards and
using industry best practices, I cannot test for every possible usage
scenario. The three-month beta testing phase is intended to allow the
community ample time to thoroughly test a new release series.
*Please* use that time to test the new release series rather than
waiting until after the .0 release!
Why there may never be a libjpeg-turbo 3.1
------------------------------------------
I will continue to fix bugs in libjpeg-turbo and issue bug-fix
releases in the 3.0.x release series, but there will not be a
libjpeg-turbo 3.1 release series *unless* this project can secure more
general funding. As it stands, libjpeg-turbo only has general funding
for about 8-10 hours of labor per month. Finishing the 3.0 beta
release required borrowing against all expected general funding for
2023, and fixing all of the 3.0 post-beta bugs required borrowing
against all expected general funding through September of 2024. If
things continue apace, then libjpeg-turbo is effectively in
"maintenance mode". That means that no new features (even minor ones)
can be considered, and tech support will be limited, for at least the
next 15 months.
I maintain three open source projects for a living. Those projects
have saved companies many millions of dollars in labor and IT costs,
but I take home less pay than that of a typical starting teacher.
(For those outside of the U.S., a starting teacher's salary is a
rhetorical benchmark for very low pay. We criminally underfund
education here just like we criminally underfund open source
development.) What I earn through independent open source development
is about 20-25% of what my skills would be worth to a corporate
employer, and I have turned down numerous offers from such employers
over the years in order to continue working on these OSS projects
full-time. I don't expect to get rich from open source development,
but it is also unfair for organizations to profit handsomely from my
work while I am expected to do that work for free.
Because the majority of my income is from VirtualGL and TurboVNC,
uncompensated labor on libjpeg-turbo forces me to steal time away from
those more lucrative projects. Thus, I simply can't eat labor cost on
libjpeg-turbo anymore. (I ate hundreds of thousands of dollars' worth
of it from 2010-2018, but when the libjpeg-turbo 2.0 release caused me
to go into debt in 2018, I had to stop.) However, major releases still
require a lot of general labor. Because the project is so high-profile
(used by literally billions of people every day through major web
browsers, operating systems, and image viewers/editors) and is an
ISO/ITU-T reference implementation, it undergoes intense scrutiny.
Unfortunately, that scrutiny typically lags the development of new
features by enough time that specific funding is no longer available
to fix unanticipated bugs in the new features. Thus, it has become
increasingly difficult to estimate the labor costs of new features
upfront. In order to adequately cushion against unforeseen bugs, I
would have to grossly overestimate every new feature to the point that
no one would be willing to fund the feature. Instead, I have had to
accept the risk of unforeseen bugs in order to keep libjpeg-turbo
moving forward, and that has forced me to perform many hours of
general labor on the run-up to every new release series. The
companies that have partially (but unfortunately not fully) funded
each new release series have pressured me to put out the .0 releases
in a timely manner, so that has left me with no choice but to borrow
against expected future funding. Specific to 3.0, I had to spend more
than 70 uncompensated hours fixing bugs discovered in the beta
release. Most of those hours were spent fixing bugs in the lossless
JPEG feature, but I also spent a significant amount of time fixing
bugs in the partial image decompression and RGB565 features (both of
which were integrated to support Android and both of which have
required me to eat the cost of bug fixes or beg for money a handful of
times since they were integrated.)
This project has only been kept alive because of the libjpeg-turbo
General Fund, which is funded by Cendio AB and Crimson Vista and your
ad hoc donations, as well as a few large ad hoc donations from Google.
However, the current General Fund is only enough to pay for general
project maintenance and support. To keep libjpeg-turbo moving forward
in any meaningful way, including providing community support,
integrating or implementing minor features, supporting popular new
platforms, etc. requires more general funding. To keep libjpeg-turbo
moving forward in a more proactive/progressive way, including
expanding the coverage of the JPEG spec (3.0 essentially expands our
coverage of the JPEG spec by a factor of five), expanding SIMD
coverage to new algorithms or instruction sets or CPU architectures,
supporting less popular new platforms, improving the APIs, hardening
security, improving fuzzer coverage, enhancing the build system,
improving automation, etc. requires even more general funding.
Essentially I have been expected to keep libjpeg-turbo moving forward
in a proactive/progressive way, but the amount of general funding only
allows for maintenance mode. In the best case, that puts me in the
position of begging for funding on a regular basis, which distracts
from meaningful open source development. In the worst case (where we
are now), that puts me in the position of stalling the project for
potentially more than a year until funding can catch up to the labor
I've already contributed. It just isn't working, and I don't know how
to fix it. It's ironic that the most popular/ubiquitous open source
project I maintain (by at least four or five orders of magnitude) is
much more difficult to fund than the others. Maybe a JPEG codec isn't
"sexy" enough? Maybe people take it for granted because the complexity
of it is hidden behind higher-level interfaces and applications?
Still, though, the organizations that develop those higher-level
interfaces and applications should not be taking libjpeg-turbo for
granted, particularly when most of the recent issues I was expected to
fix for free were security-related.
I have expressed in the past that the business model necessitated by
independent open source development may not be the best fit for a
critical infrastructure project such as libjpeg-turbo. Thus, I am
open to acquisition by a larger organization, if that organization is
both O/S-agnostic and CPU-agnostic and has a vested interest in both
the survival and the openness of libjpeg-turbo. However, I also don't
expect such an offer to ever materialize. I have built significant
value in libjpeg-turbo-- including its reputation, its ubiquity, and
its user community-- over the past 13 years, but it is obviously not a
cash cow. Still, though, unless I can secure enough general funding to
move the project forward in a meaningful way, the only other options
are acquisition (however unlikely) or putting the project into
maintenance mode for the foreseeable future. The latter is where we
are now.
How can you help? If every individual developer who used
libjpeg-turbo on a regular basis donated just $5-10/month to the
project through GitHub Sponsors
(https://github.com/sponsors/libjpeg-turbo), we'd have a healthy
amount of general funding. But that requires everyone to take
responsibility for the health of this project, not to rely on the few
corporate benefactors that have kept libjpeg-turbo on life support for
the past five years.
DRC
--
You received this message because you are subscribed to the Google Groups
"libjpeg-turbo User Discussion/Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to libjpeg-turbo-users+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/libjpeg-turbo-users/dd3c51c6-4eb1-a56f-cc46-bc687a7b9200%40virtualgl.org.
