This brings up a bigger point, one that has certainly been raised before: Are all these OSS licenses really necessary, or productive?... I think that a massive consolidation of OSI-approved licenses is in order.
That is of course what the proponents of AFL and OSL have in mind, that everyone will relicense BSD-like projects to AFL and, and relicense GPL and other copyleft projects to OSL, and abandon all of the others. Good luck.
How do you stop non-techies from going "Oh, somebody loves me! I'll just read this message... OHNOS MY HARDDRIVE!"? - Peter
How is this a troll? It's early and short, but looks on-topic to me.
Every morning I possess about 20 KLEZ worms. After a few months of KLEZ mailbombing I got "POP3 Scan Mailbox" and set it to schedule anything over 50kbytes for removal, so at least they don't wear out my modem any more.
[My main problem] has been that when i install it it takes over all the images on my computer and makes mozilla the default loader for them..jpegs,.gifs, everything, as far as i can tell. any ideas on how to disable this?
Edit->Preferences->Advanced->System
If you uncheck some types, Mozilla kindly gives them back to the application that had them before.
SECTION 410. NO IMPLIED WARRANTIES FOR FREE COMPUTER PROGRAM. The warranties under Sections 401 and 403 apply to a computer program only if the licensor intends to make a profit from the distribution of the copy of the program or acts generally for commercial gain derived from controlling use of the program or making, modifying, or redistributing copies of the program.
"Also, for each author's protection and ours, we want to make certain that everyone understands that there is no warranty for this free software."
(emphasis mine)
AFAIK, UCITA has no power to change that.
It does change that, though. It says that warranty against patent infringment must be specifically disclaimed, or you provide a warranty despite your "no warranty" clause.
We added the UCITA-specific language to the libpng license (the real one used by libpng, not the one posted at OSI) several years ago.
Yes, they would hold for JNG's image compression method 8 (JPEG) which is the only one defined right now, and alpha compression method 8 (JPEG) which is one alternative (method 0, PNG IDAT, being the other). If push comes to shove we could define a different method, but someone would have to invent that different method and make it available royalty-free.
I seem to recall that it's some sort of personality code, "I" being "introverted", and so forth. ISTR that gov't officials are mostly the ISTJ type while software geeks such as myself are usually INTP.
A number of posters have noted that most people have little of importance on their hard drives.
The real problem is that most people have gigabytes of trash or stuff that is easily recoverable, like downloads, object files and executables, and a few hundred kilobytes of crucially important data like their own original program sources. Automatic backup systems are too stupid to tell which is which, and manual backups just don't get done.
So did I. It took a while yesterday to figure out how to get rid of MSIE entirely. One thing I like is that Moz isn't subject to the Trolls' "slashdot widening" trick. Downside is a (bug?) that locks up the mousewheel when I select a choice in the moderation menu -- but found out that the wheel is activated again simply by clicking anywhere on the page.
JNG isn't what I had in mind though - it's just JPEG compression underneath. I meant create an ordinary (zlib-compressed) PNG file which is almost, but not quite identical to the original image and so can compress better.
That wouldn't require a new format spec, though, just some precompression processing of the image, like the "smooth" option that's available in the IJG JPEG software, or simple color-reduction that people do already. The result would still be a regular PNG file, decodable by current PNG decoders.
That's JNG (JPEG Network Graphics) which is JPEG wrapped in PNG-style chunks along with an optional alpha (transparency) channel. Mozilla will display them and IrfanView will process them.
The JNG spec is available somewhere on the PNG web site, http://www.libpng.org/pub/png
The frames of a MNG datastream can be in the JPEG (JNG) subformat as well as in the PNG format. Read the JNG spec at the PNG web site, http://www.libpng.org/pub/mng/
JPEG allows arbitrary headers; one such header is the EXIF [exif.org] header which most digital cameras will include. This includes stuff like date and time taken, focal length, etc. The problem is that since it's an extension of JPEG rather than a mandatory part of the standard, any software is free to ignore the EXIF header, and neglect to preserve it when modifying the image. For example, take a JPEG from a digital camera, with date and time helpfully included in the EXIF header... run ImageMagick "mogrify" on it; perhaps to resize it or to change the JPEG compression ratio -- EXIF header disappears
Recent versions of ImageMagick preserve the EXIF header. This is a FAQ on the ImageMagick mailing lists: why is my thumbnail 13kbytes instead of 3kbytes like the file that other applications produce? A: You have to remove the EXIF profile (which is actually encoded as a JPEG APPN1 tag) explicitly. The same goes for IPTC (newswire info) and ICC color profiles.
I must be confusing people tonight. Did you not read this from the post to which you replied:
Images using the indexed pixel type must resort to "simple" transparency to emulate a "real" alpha channel, however.
Sorry -- I interpreted your "simple" to mean "binary".
Simple transparency is effected through PNG's tRNS chunk, and yes, it does work just as you describe for indexed PNGs. Sufficient for Web work, but: (1) still not a true 8-bit alpha channel, though effectively interpreted as such; and (2) potentially far narrower in range than alpha channel support for truecolor/grayscale PNGs (8-bit v. 8-bit/16-bit).
That's all true, but nevertheless you can do a variety of interesting things with it such as a 1-color PNG with 256 levels of transparency, 16-color PNG with 16 levels of transparency, down to GIF-style 255 colors with binary transparency using one index for a "transparent" color (except that one color could be partially transparent if you like).
I think it is a crying shame that the most popular browser fails to support PNG transparency properly even though PNG has been around now for seven years and approved by W3C as the standard image format for the web 5-1/2 years ago.
If you want lossy compression of photographic images with a PNG-style alpha transparency, you can use JNG (from the PNG group) which is simply JPEG wrapped in PNG-style chunks. The spec has been frozen for over a year and it's implemented in Mozilla.
And yes, alpha transparency is supported *only* for truecolor and grayscale types, not indexed types.
Alpha transparency is supported in indexed PNG images via the "tRNS" chunk which in effect gives you a 256-entry palette of RGBA, where R,G,G, and A all range from 0 to 255. The "toucan" image at the PNG web site is such a file.
Several people including Chris Lilley of W3C and myself have suggested on the comments list that if W3C publishes a Recommendation with patent claims, it should do more than simply listing the claimants, as was done in the SVG spec. There should be at least some commentary that identifies exactly what patents are involved and exactly what part of the Recommendation might infringe on them. The SVG Recommendation lists two patent numbers, and ten other companies who simply say that they have intellectual property that might be infringed by the SVG Recommendation.
Do you agree that W3C should include this type of extra information in any Recommendation that depends on RAND patents (I prefer to call them UFAD for unfair and discriminatory, though)?
Fits on a 5.25 inch disk and runs well in 128kb of memory! Fast and easy to use!
My favorite was the unnamed OS that came with the AIM-65 (6502-chip-based) machine. It fit in a 4k EPROM and ran well in 1k of memory and an additional 4k of EPROM for application code such as the assembler. Problem was it completely filled the EPROM so when I wanted to upgrade the assembler to handle new 6502-C instructions, first I had to rewrite some of the existing code to take fewer bytes, to make room for the new stuff.
What happens when someone takes their GPL copy and makes a bugfix or an enhancement? Does that force a code fork, or is there a mechanism to get the contribution into the other-licensed versions?
how come nobody ever asks why people watch tv for five hours a day when nobody's paying them too?
Mod this up. It's a viewpoint that the Congressman might understand. Five or ten hours a day of coding is more challenging and intellectually rewarding than five hours of watching TV.
The Enoch Pratt Free Library in Baltimore, MD, offered low-cost shell accounts ($36.00 per year) for several years (ca. 1996-98) until they were forced by some commercial ISP to stop the "unfair competition".
Here's my guesstimate. Assuming an ISO9660, written with names averaging 40 bytes long (double-byte encoding), you could fit something like 17 million names on a single disk.
They'll probably alphabetize and gzip the list, which should get better than 100 million names per disk. BTW, Mark Adler of NASA wrote gzip's "inflate" function, which would be needed to decode the list on Mars.
There's a retired Cray-2 (formerly bob.brl.mil) in the reception area at the Army Research Laboratory, Aberdeen Proving Ground, now serving as a coffee table.
Slashdot Mirror
This brings up a bigger point, one that has certainly been raised before: Are all these OSS licenses really necessary, or productive? ... I think that a massive consolidation of OSI-approved licenses is in order.
That is of course what the proponents of AFL and OSL have in mind, that everyone will relicense BSD-like projects to AFL and, and relicense GPL and other copyleft projects to OSL, and abandon all of the others. Good luck.
The AFL, unlike MIT and BSD, does not require that the copyright
be maintained in derivative works, but only in distributions of
the Original Work.
How do you stop non-techies from going "Oh, somebody loves me! I'll just read this message... OHNOS MY HARDDRIVE!"?
- Peter
How is this a troll? It's early and short, but looks on-topic to me.
Every morning I possess about 20 KLEZ worms. After a few months
of KLEZ mailbombing I got "POP3 Scan Mailbox" and set it to schedule
anything over 50kbytes for removal, so at least they don't wear out
my modem any more.
[My main problem] has been that when i install it it takes over all the images on my computer and makes mozilla the default loader for them. .jpegs, .gifs, everything, as far as i can tell. any ideas on how to disable this?
Edit->Preferences->Advanced->System
If you uncheck some types, Mozilla kindly gives them back to the application that had them before.
Here's a good thing:
11. Amendment # 11: Section 410 (new)
Add the following new section:
SECTION 410. NO IMPLIED WARRANTIES FOR FREE COMPUTER PROGRAM. The warranties under Sections 401 and 403 apply to a computer program only if the licensor intends to make a profit from the distribution of the copy of the program or acts generally for commercial gain derived from controlling use of the program or making, modifying, or redistributing copies of the program.
Wait a minute ... the GPL says:
"Also, for each author's protection and ours, we want to make certain that everyone understands that there is no warranty for this free software."
(emphasis mine)
AFAIK, UCITA has no power to change that.
It does change that, though. It says that warranty against patent
infringment must be specifically disclaimed, or you provide a warranty
despite your "no warranty" clause.
We added the UCITA-specific language to the libpng license (the real
one used by libpng, not the one posted at OSI) several years ago.
Glenn
would those same patents hold for the JNG format?
Yes, they would hold for JNG's image compression method 8 (JPEG)
which is the only one defined right now, and alpha compression
method 8 (JPEG) which is one alternative (method 0, PNG IDAT, being
the other). If push comes to shove we could define a different
method, but someone would have to invent that different method
and make it available royalty-free.
Glenn
I seem to recall that it's some sort of personality code, "I" being
"introverted", and so forth. ISTR that gov't officials are mostly
the ISTJ type while software geeks such as myself are usually INTP.
A number of posters have noted that most people have little of importance on their hard drives.
The real problem is that most people have gigabytes of trash or stuff that is easily recoverable, like downloads, object files and executables, and a few hundred kilobytes of crucially important data like their own original program sources. Automatic backup systems are too stupid to tell which is which, and manual backups just don't get done.
So did I. It took a while yesterday to figure out how to get rid of MSIE entirely. One thing I like is that Moz isn't subject to the Trolls' "slashdot widening" trick. Downside is a (bug?) that locks up the mousewheel when I select a choice in the moderation menu -- but found out that the wheel is activated again simply by clicking anywhere on the page.
Glenn
JNG isn't what I had in mind though - it's just JPEG compression underneath. I meant create an ordinary (zlib-compressed) PNG file which is almost, but not quite identical to the original image and so can compress better.
That wouldn't require a new format spec, though, just some precompression processing of the image, like the "smooth" option that's available in the IJG JPEG software, or simple color-reduction that people do already. The result would still be a regular PNG file, decodable by current PNG decoders.
I'd like to see some kind of lossy PNG
That's JNG (JPEG Network Graphics) which is JPEG wrapped in PNG-style chunks along with an optional alpha (transparency) channel. Mozilla will display them and IrfanView will process them.
The JNG spec is available somewhere on the PNG web site, http://www.libpng.org/pub/png
The frames of a MNG datastream can be in the JPEG (JNG) subformat as well as in the PNG format. Read the JNG spec at the PNG web site, http://www.libpng.org/pub/mng/
JPEG allows arbitrary headers; one such header is the EXIF [exif.org] header which most digital cameras will include. This includes stuff like date and time taken, focal length, etc. The problem is that since it's an extension of JPEG rather than a mandatory part of the standard, any software is free to ignore the EXIF header, and neglect to preserve it when modifying the image. For example, take a JPEG from a digital camera, with date and time helpfully included in the EXIF header... run ImageMagick "mogrify" on it; perhaps to resize it or to change the JPEG compression ratio -- EXIF header disappears
Recent versions of ImageMagick preserve the EXIF header. This is a FAQ on the ImageMagick mailing lists: why is my thumbnail 13kbytes instead of 3kbytes like the file that other applications produce? A: You have to remove the EXIF profile (which is actually encoded as a JPEG APPN1 tag) explicitly. The same goes for IPTC (newswire info) and ICC color profiles.
Glenn
I must be confusing people tonight. Did you not read this from the post to which you replied:
Images using the indexed pixel type must resort to "simple" transparency to emulate a "real" alpha channel, however.
Sorry -- I interpreted your "simple" to mean "binary".
Simple transparency is effected through PNG's tRNS chunk, and yes, it does work just as you describe for indexed PNGs. Sufficient for Web work, but: (1) still not a true 8-bit alpha channel, though effectively interpreted as such; and (2) potentially far narrower in range than alpha channel support for truecolor/grayscale PNGs (8-bit v. 8-bit/16-bit).
That's all true, but nevertheless you can do a variety of interesting things with it such as a 1-color PNG with 256 levels of transparency, 16-color PNG with 16 levels of transparency, down to GIF-style 255 colors with binary transparency using one index for a "transparent" color (except that one color could be partially transparent if you like).
I think it is a crying shame that the most popular browser fails to support PNG transparency properly even though PNG has been around now for seven years and approved by W3C as the standard image format for the web 5-1/2 years ago.
Glenn
If you want lossy compression of photographic images with a PNG-style alpha transparency, you can use JNG (from the PNG group) which is simply JPEG wrapped in PNG-style chunks. The spec has been frozen for over a year and it's implemented in Mozilla.
Glenn
And yes, alpha transparency is supported *only* for truecolor and grayscale types, not indexed types.
Alpha transparency is supported in indexed PNG images via the "tRNS" chunk which in effect gives you a 256-entry palette of RGBA, where R,G,G, and A all range from 0 to 255. The "toucan" image at the PNG web site is such a file.
Glenn
They should supply an uninstaller. @home didn't.
Several people including Chris Lilley of W3C and myself have suggested on the comments list that if W3C publishes a Recommendation with patent claims, it should do more than simply listing the claimants, as was done in the SVG spec. There should be at least some commentary that identifies exactly what patents are involved and exactly what part of the Recommendation might infringe on them. The SVG Recommendation lists two patent numbers, and ten other companies who simply say that they have intellectual property that might be infringed by the SVG Recommendation.
Do you agree that W3C should include this type of extra information in any Recommendation that depends on RAND patents (I prefer to call them UFAD for unfair and discriminatory, though)?
Fits on a 5.25 inch disk and runs well in 128kb of memory! Fast and easy to use!
My favorite was the unnamed OS that came with the AIM-65 (6502-chip-based) machine. It fit in a 4k EPROM and ran well in 1k of memory and an additional 4k of EPROM for application code such as the assembler. Problem was it completely filled the EPROM so when I wanted to upgrade the assembler to handle new 6502-C instructions, first I had to rewrite some of the existing code to take fewer bytes, to make room for the new stuff.
What happens when someone takes their GPL copy and makes a bugfix or an enhancement? Does that force a code fork, or is there a mechanism to get the contribution into the other-licensed versions?
how come nobody ever asks why people watch tv for five hours a day when nobody's paying them too?
Mod this up. It's a viewpoint that the Congressman might understand. Five or ten hours a day of coding is more challenging and intellectually rewarding than five hours of watching TV.
Glenn R-P (libpng)
The Enoch Pratt Free Library in Baltimore, MD, offered low-cost shell accounts ($36.00 per year) for several years (ca. 1996-98) until they were forced by some commercial ISP to stop the "unfair competition".
Here's my guesstimate. Assuming an ISO9660, written with names averaging 40 bytes long (double-byte encoding), you could fit something like 17 million names on a single disk.
They'll probably alphabetize and gzip the list, which should get better than 100 million names per disk. BTW, Mark Adler of NASA wrote gzip's "inflate" function, which would be needed to decode the list on Mars.
There's a retired Cray-2 (formerly bob.brl.mil) in the reception area at the Army Research Laboratory, Aberdeen Proving Ground, now serving as a coffee table.