We're up to at least six now. Surprise, surprise, the search function doesn't go back to the first two in October 1998--the archives have been wiped. Those two were on or just before 19 October, when both CmdrTaco and someone else posted the exact same link. Then there was one on 27 January, one on 8 February, one on 3 March, and this one on 12 March. Possibly as many as two of them have posted anything new...
If you really want to distribute digital audio and video streams throughout the house, IEEE-1394 (and HAVi) was designed for it. Support is still pretty sparse even in the consumer-electronics arena, but it's coming. Linux support is farther off still. But at 400-1600 Mbps, it's in the same bandwidth ballpark as Ultra2/Ultra3 SCSI.
Is there any particular advantage to making PNG an ISO/IEC standard? These standards are sold for profit, and so the standard will not be freely available if it goes ISO/IEC. So, why do it?
The PNG spec, like the VRML 2.0 spec before it, will remain freely available. The ISO will have the right to sell the version with their boilerplate and logo, but the technical content will be identical, and there will be no restrictions on the distribution of the W3C and PNG Development Group versions.
As for benefits: believe it or not, there are companies (and especially governments) that require the use of an international standard wherever possible. In the case of the US, it is often the case that companies wishing to do business under government contract must support the relevant standards in their products--and to the extent that this support is not seen as a complete waste of time and code, it generally ends up in the versions of the product that you buy, too.
ISO/IEC standardization is also seen as somehow better/safer than standardization promulgated by Internet upstarts like the W3C, IETF or PDG. Again, that may result in more and/or better support in your favorite products.
Of course, these things take time, and it looks like there will be another FCD and four-month voting period before PNG even makes it to the Draft International Standard stage, which itself lasts a good number of months. So it's pretty safe to say that there are no immediate advantages to all of this.
By the way, it's conceivable (though not definite) that zlib/deflate will be next.:-)
Does anyone else find it ironic that the PNG logo on the PNG website is itself a GIF?
Correction: it's a GIF to you, presumably because you have a stupid (i.e., not conformant with HTML 4.0) browser. To anyone with a clueful browser, it's a PNG. Here's the source (which, of course, you could have looked at yourself):
The only general-purpose approach that guarantees a PNG image to most of the browsers capable of viewing them natively is content negotiation, something that requires direct access to the server (and also something that even Apache doesn't implement very elegantly). Did I mention that all current versions of Navigator have a bug in their implementation of client-side negotiation, a tiny bug that nevertheless means IIS servers will never send an embedded (that is, IMG-tagged) PNG image to them? Check out puzzlemaker.com, one of the better-known casualties.
But I will do my best to make sure Mozilla/Gecko has excellent PNG support. I have just one little detail to get out of the way first...
So shouldn't we be adding it to GIMP, Mozilla, Imlib and others? Can you provide a URL?
As others have noted, MNG is pretty much fully specified (and note that it includes a JPEG/PNG hybrid called JNG that provides transparency and gamma/color-correction to JPEG), but there is basically zero in the way of sample source code for it--and it is much more complex than PNG. On the plus side, Gerard Juyn has offered to make his MNGeye viewer code available as the starting point for a "libmng" implementation, but no one has had time to do anything with it yet. So those who are truly keen on sprite-based animation a la MNG can always get started and make life nicer for all of us.
In fairness, I should also note that Flash is in many respects competitive with MNG, and it includes some PNG-like support as well. So despite the fact that MNG includes some features not present in Flash, it may well turn out that Flash "wins" simply by virtue of being out there, being "open enough," and having several implementations (well, at least two, anyway).
Even if Flash becomes the de facto next-generation animation standard, I suspect that JNG will find favor, too. If you're already linking with libjpeg and libpng, adding support for JNG requires almost no additional code.
The CFA's full report is still in MIME quoted-printable format, except with the MIME headers stripped off. This leaves lots of "=" signs lying about and completely destroys several of the tables that were created with massive bursts of non-breaking spaces. Alas, if only the zippies had heard of munpack...
Slashdot Mirror
We're up to at least six now. Surprise, surprise, the search function doesn't go back to the first two in October 1998--the archives have been wiped. Those two were on or just before 19 October, when both CmdrTaco and someone else posted the exact same link. Then there was one on 27 January, one on 8 February, one on 3 March, and this one on 12 March. Possibly as many as two of them have posted anything new...
If you really want to distribute digital audio and video streams throughout the house, IEEE-1394 (and HAVi) was designed for it. Support is still pretty sparse even in the consumer-electronics arena, but it's coming. Linux support is farther off still. But at 400-1600 Mbps, it's in the same bandwidth ballpark as Ultra2/Ultra3 SCSI.
The PNG spec, like the VRML 2.0 spec before it, will remain freely available. The ISO will have the right to sell the version with their boilerplate and logo, but the technical content will be identical, and there will be no restrictions on the distribution of the W3C and PNG Development Group versions.
As for benefits: believe it or not, there are companies (and especially governments) that require the use of an international standard wherever possible. In the case of the US, it is often the case that companies wishing to do business under government contract must support the relevant standards in their products--and to the extent that this support is not seen as a complete waste of time and code, it generally ends up in the versions of the product that you buy, too.
ISO/IEC standardization is also seen as somehow better/safer than standardization promulgated by Internet upstarts like the W3C, IETF or PDG. Again, that may result in more and/or better support in your favorite products.
Of course, these things take time, and it looks like there will be another FCD and four-month voting period before PNG even makes it to the Draft International Standard stage, which itself lasts a good number of months. So it's pretty safe to say that there are no immediate advantages to all of this.
By the way, it's conceivable (though not definite) that zlib/deflate will be next. :-)
Correction: it's a GIF to you, presumably because you have a stupid (i.e., not conformant with HTML 4.0) browser. To anyone with a clueful browser, it's a PNG. Here's the source (which, of course, you could have looked at yourself):
<H5 ALIGN="center">
<A HREF="img_png/pnglogo-blk.jpg">
<OBJECT WIDTH="256" HEIGHT="192" TYPE="image/png"
DATA="img_png/pnglogo-blk-sml1.png">
<IMG WIDTH="256" HEIGHT="192" SRC="img_png/pnglogo-blk-sml1.gif"
ALT="[PNG: 256x192 colored-balls logo]"></OBJECT></A>
</H5>
The only general-purpose approach that guarantees a PNG image to most of the browsers capable of viewing them natively is content negotiation, something that requires direct access to the server (and also something that even Apache doesn't implement very elegantly). Did I mention that all current versions of Navigator have a bug in their implementation of client-side negotiation, a tiny bug that nevertheless means IIS servers will never send an embedded (that is, IMG-tagged) PNG image to them? Check out puzzlemaker.com, one of the better-known casualties.
But I will do my best to make sure Mozilla/Gecko has excellent PNG support. I have just one little detail to get out of the way first...
As others have noted, MNG is pretty much fully specified (and note that it includes a JPEG/PNG hybrid called JNG that provides transparency and gamma/color-correction to JPEG), but there is basically zero in the way of sample source code for it--and it is much more complex than PNG. On the plus side, Gerard Juyn has offered to make his MNGeye viewer code available as the starting point for a "libmng" implementation, but no one has had time to do anything with it yet. So those who are truly keen on sprite-based animation a la MNG can always get started and make life nicer for all of us.
In fairness, I should also note that Flash is in many respects competitive with MNG, and it includes some PNG-like support as well. So despite the fact that MNG includes some features not present in Flash, it may well turn out that Flash "wins" simply by virtue of being out there, being "open enough," and having several implementations (well, at least two, anyway).
Even if Flash becomes the de facto next-generation animation standard, I suspect that JNG will find favor, too. If you're already linking with libjpeg and libpng, adding support for JNG requires almost no additional code.
The CFA's full report is still in MIME quoted-printable format, except with the MIME headers stripped off. This leaves lots of "=" signs lying about and completely destroys several of the tables that were created with massive bursts of non-breaking spaces. Alas, if only the zippies had heard of munpack...